U.S. Department of Health and Human ServicesU.S. Department of Health and Human Services
PDF Version: http://aspe.hhs.gov/daltcp/reports/5yrpenn.pdf (267 PDF pages)
This report was prepared under contracts #130-81-0021 and #130-81-00022 between the U.S. Department of Health and Human Services (HHS), Office of Social Services Policy (now the Office of Disability, Aging and Long-Term Care Policy) and Temple University. For additional information about the study, you may visit the DALTCP home page at http://aspe.hhs.gov/daltcp/home.htm or contact the office at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, SW, Washington, DC 20201. The e-mail address is: webmaster.DALTCP@hhs.gov. The DALTCP Project Officer was Ron Conley.
This project was supported by the U.S. Department of Health and Human Services under two contracts: #130-81-0022 and #130-81-0021. The views expressed herein are those of the authors and do not necessarily reflect the official position of the Office of Human Development Services, the Office of the Assistant Secretary for Planning and Evaluation or the U.S. Department of Health and Human Services.
James W. Conroy, Principal Investigator
Celia S. Feinstein, Director
of Case Studies, Tracking, and Environmental Assessment
James A. Lemanowicz,
Senior Data Analyst
Robert R. Walsh, Director of Client Satisfaction Study
Valerie J. Bradley, Principal Investigator
John W. Ashbaugh, Director
of Cost Study
Mary Ann Allard, Senior Analyst
The Recommended Citation for this publication is:
Conroy, J.W., & Bradley, V.J. (1985). The Pennhurst Longitudinal Study: A report of five years of research and analysis. Philadelphia: Temple University Developmental Disabilities Center. Boston: Human Services Research Institute.
Additional copies of this publication may be obtained postpaid at $15 per copy, check payable to Temple University, from:
Temple University
Developmental Disabilities Center
9th Floor, Ritter Hall Annex
Philadelphia, PA 19122
Attn: Kathy Ann Baus
The Temple and Human Services Research Institute research team would like to recognize the enormous contributions of the National Advisory Council for this study and of the Pennhurst Study Work Group. The five years of commitment from the members of these groups helped immeasurably to enhance and maintain the integrity of the study.
The members of the National Advisory Council were Dr. David Braddock, Dr. Robert Bruininks, Dr. William Datel, Ms. Irene Dixon, Mr. Robert Gettings Dr. Jack Noble, and Dr. Wayne Smith.
The members of the Pennhurst Study Work Group were Mr. Michael Bennett (Pennhurst Implementation Team), Mr. Michael Lottman (Hearing Master), Ms. Carla Morgan (Special Master), Dr. Russell Rice (Regional Commissioner for Mental Retardation, Southeast Region), Ms. Margaret Rubin (Pennsylvania Developmental Disabilities Planning Council), and Ms. Vicki Stillman-Toomey (Director of Pennhurst Implementation Team).
The federal officials in the Department of Health & Human Services who conceived of this project and followed its progress for five years deserve a large part of the credit for its success. Among them were Mr. John McCann, Dr. Ronald W. Conley, Dr. Jack Noble, Mr. William Crunk, Mr. John DeWilde, Mr. Ed Marcus, and in the final years of the study, Ms. Eileen Bradley, Mr. Edward Vreeswyk, and Ms. Ann Schoonmaker,
Several state officials also provided indispensable support, cooperation and assistance: Mr. Gregory Pirmann (Pennhurst Center), Mr. George Kopchik (Pennhurst Center), and Dr. Jennifer Howse (Office of Mental Retardation).
This report is dedicated to the 1154 people who lived at Pennhurst Center on March 17, 1978.
 |
DEPARTMENT OF HEALTH AND HUMAN SERVICES | Office of Human Development Services, Region III |
| 3535 Market Street Philadelphia, Pennsylvania 19104 |
March 1, 1985
The Pennhurst Longitudinal Study was a five-year, in-depth review of the effects of the court-ordered deinstitutionalization of Pennhurst residents. Its aim was to provide federal and state officials and others with information to make better policy decisions regarding the processes related to the deinstitutionalization which is underway in many parts of the country. Federal planners, recognizing the significance for clients, families, communities, and states of Judge Broderick's and other similar actions, decided in 1979 to launch this five-year study.
This project was unique in several respects; perhaps most importantly as a joint endeavor among the Region III Office of Human Development Services, the Assistant Secretary for Human Development Services, the Assistant Secretary for Planning and Evaluation, and the Commissioner of the Administration on Developmental Disabilities, as well as the Deputy Secretary for Mental Retardation in the Commonwealth of Pennsylvania.
The study has yielded considerable information which is not only valuable to other researchers, but also instructive to legislators, judges, and federal, state and local program administrators and policy makers. We commend the foresight of the initiators of the project and the dedication of those who completed it. In addition, we acknowledge the thoughtful contributions of the Pennhurst Study Advisory Committee and Work Group.
Eileen Bradley
Regional Administrator
When United States District Court Judge Raymond Broderick issued his opinion in the Halderman v. Pennhurst State School and Hospital case in 1977, it was considered the most far-reaching legal event in the field of mental disabilities to date. Unlike other federal judges who had primarily focused their attention on the improvement of institutional settings, Broderick ruled that Pennhurst State School and Hospital was incapable of providing constitutionally appropriate care and habilitation. This finding led him to conclude that the residents of Pennhurst, those on the waiting list to the institution, and any other mentally retarded person in the community "at risk" of institutionalization at Pennhurst should be provided services in less restrictive settings in the community.
Following Broderick's ruling and the issuance of his decree in March, 1978, plaintiffs in 20 other states began the process of seeking similar relief. Recognizing the potential national significance of the Pennhurst case, leadership in the Department of Health and Human Services (DHHS) -- specifically in the Office of the Assistant Secretary for Planning and Evaluation (ASPE), and the Region III (Philadelphia) office of Human Development Services (OHDS) -- decided to support a five year longitudinal study which would:
The main value of the study has been its utility in providing DHHS, state, and court officials with information on which vital short and long term policy decisions can be made. From the initial conception of the project, the Pennhurst Longitudinal Study has been a partnership involving the Commonwealth of Pennsylvania, the court, ASPE, and OHDS Region III. Further, because of the information needs of the Office of the Special Master and the Pennsylvania Office of Mental Retardation, their representatives also participated in the original design and in the ongoing oversight of the project. In addition to direction provided by the DHHS project officers, the Pennhurst Study Work Group was also established to ensure the study's continued relevance. Member of this group include representatives of the Special Master, the Hearing Master, the Office of Mental Retardation, the Pennsylvania Developmental Disabilities Council, and the Region III Developmental Disabilities Office.
Further, in order to ensure the national relevance of materials emerging from the Pennhurst Longitudinal Study, an Advisory Committee was appointed. The Committee is comprised of national experts in the field of mental retardation, a representative of the national organization of state mental retardation commissioners, and others interested in the general area of litigation and deinstitutionalization..
The Longitudinal Study, which is being conducted as a collaborative effort by the Temple University Developmental Disabilities Center in Philadelphia and Human Services Research Institute in Boston, is divided into three major parts:
Impact on Clients and Communities (Temple University)
This portion of the study involved monitoring trip developmental progress of the study population, the services they received, the quality of their living environments, and the level of their satisfaction -- both at Pennhurst and after relocation to the community. This segment also included an assessment of the impact of deinstitutionalization on families of clients, both in anticipation of the action to be taken under the decree and following the actual relocation, and the attitudes of others in clients' local communities both before and after deinstitutionalization. Included in this study component were case studies of several Pennhurst residents which provide a more in-depth exploration of the impact of the case on particular individuals.
Briefly, the study population had the following characteristics:
Impact on Costs (Human Services Research Institute)
The results of this portion of the study include an assessment of the costs and cost configurations of services provided both at Pennhurst and in community settings. For as many service categories as possible, average cost per unit of service at Pennhurst and in the community were calculated. These service unit costs were applied to the reported units of service received by individual clients. From this, the staff derived estimates of total costs for each relocated client, as a function of how much service the client actually received.
History and Implementation Analyses (Human Services Research Institute)
This study area included a continually updated historical account of the implementation of the Pennhurst decree and the events surrounding the litigation for the first three years of the study. In the course of these assessments, the actions and intentions of policy makers were highlighted. Further, the interrelationships among events and key system actors are chronicled and the implications for state and federal policy were explored. In addition, four aspects of implementation were singled out during the course of the study for extensive investigation and analysis.
This final report of the results of the Longitudinal Study integrates qualitative, quantitative, and cost findings into one comprehensive report in order to facilitate a review of the varied strands of evidence generated by both Temple University and Human Services Research Institute. The material is organized as follows:
Chapter II -- History of the Case
This chapter provides a summary of the six Historical Overviews prepared during the first three years of the Longitudinal Study. It is organized chronologically, with special sections on legal developments, Pennhurst-related developments, and general system developments at each historical stage. The chapter concludes with a brief summary of events for the last two years.
Chapter III -- Implementation Issues
This chapter summarizes three of the implementation issues that have been addressed in depth during the course of the project. The first topic is the Special Master mechanism employed by the federal court to monitor and enforce the Pennhurst decree. The second area focuses on the actions and reactions of the state defendants in the case as contracted with those of state defendants in other suits around the country. The third analysis is a multi-state assessment of limitations and constraints to the implementation of court decrees.
Chapter IV -- Growth and Development
This chapter describes the results of the systematic assessment of client growth and development among the study population both at Pennhurst State Center and in the community. Chapter IV through Chapter VIII represent the findings from the quantitative studies. Four of the five chapters are introduced by a digest from one of the project's case studies.
Chapter V -- Consumer Satisfaction
This chapter presents the results of surveys of clients in the study population to determine their level of satisfaction with their surroundings both in the institution and in the community.
Chapter VI -- Quality of Environments
This chapter reports the findings of surveys of client environments both at Pennhurst and in the community.
Chapter VII -- Family Attitudes
This chapter concentrates on the responses of families to the process of deinstitutionalization and focuses on changes in their attitudes over time.
Chapter VIII -- Neighbor Attitudes
This chapter discusses the results of surveys of neighbors of the clients in the study population both before and after community living arrangements were developed.
Chapter IX -- Comparative Cost Analysis
This chapter describes the comparative costs of the provision of services at Pennhurst Center and in the community.
Chapter X -- Impact of Court-Ordered Change
This chapter, which is also the fourth and final implementation analysis, explores ten questions regarding the impact of the Pennhurst case on the mental retardation system in Pennsylvania. It draws together quantitative, qualitative, and cost findings in order to shed light on the issues.
Chapter XI -- Policy Implications
This final chapter summarizes the larger policy questions that have been uncovered by the study and offers suggestions for future planning and policy development.
Where appropriate, instruments used to collect information are included in the Appendix to the report.
As part of the longitudinal evaluation of the Halderman v. Pennhurst case, a series of Historical Overview reports was prepared in order to chronicle key events surrounding the implementation of the court decree and to analyze the roles of various actors in the implementation process, Since 1980, a total of six Overview reports were prepared -- one every six to eight months (in the last two years of the study, the Overviews were replaced by brief updates). The historical analyses describe the major activities surrounding the implementation of the decree, and assess the constraints and limitations on such actions. The reports also characterize the influence that the litigation had on the general service system and, conversely, how general system factors affected the requirements of the decree.
In preparing for each Overview report a number of data gathering activities were conduted. In addition to reviewing relevant legal documents, state reports and regulations, and other materials, staff interviewed a cross section of key actors in the state including county, provider, state, legal and consumer representatives.
Each Overview report corresponded to a specific time period during which certain key events concerning the decree transpired. The first Overview was somewhat different from subsequent reports because it set the stage for ensuing analyses. As such, it served two major functions: (1) to describe the context in which the litigation was brought; and (2) to highlight those activities that took place immediately after Judge Broderick arrived at his decision in December 1977. All other phases in the historical analysis of the Pennhurst case are presented in three parts: legal activities, Pennhurstspecific activities, and general system developments.
Some of the key background elements reviewed in the first historical account included the following:
Overview of the study area -- The history and characteristics of the Southeast Region of Pennsylvania (Philadelphia, Bucks, Delaware, Montgomery and Chester Counties) were described and the history of Pennhurst Center was presented. As map of the Department of Public Welfare regions is included in Exhibit 2-1, and the population trends at Pennhurst State Center are described in Exhibit 2-2. A detailed discussion of the social and economic characteristics of the state, the region, and the five counties is included in the appendix.
Legal history of the litigation -- Similar litigation in the field of mental disabilities was reviewed and compared and contrasted with the Pennhurst case; the use of public law litigation as a tool of social reform was also analyzed.
Events leading up to the suit -- A brief account of the rationale for the Pennhurst suit was summarized including the early expose of conditions at the state center, the pivotal role of the Pennsylvania Association of Retarded Citizens, attempts by the state to upgrade Pennhurst and create alternatives in the community, and finally the filing of the suit by David Ferleger on May 30, 1974 on behalf of Terri Lee Halderman for both injunctive relief and money damages. The complaint was later amended when PARC intervened in the suit; money damages were dropped from the remedy and the five Southeast Pennsylvania counties as well as the state were named as defendants in the suit.
Legal arguments made and the remedies sought -- The plaintiffs argued that both constitutional and statutory law guarantees mentally retarded persons a right to habilitation and a right to receive services in the least restrictive setting. Moreover, the plaintiffs maintained that Pennhurst was incapable of ensuring the rights of mentally retarded persons because of deplorable conditions. The remedy sought was the ultimate closure of the facility and the movement of residents to less restrictive community-based services. The state did not significantly contest the facts presented by the plaintiffs but asserted that the proposed remedy went beyond the powers of the courts. The Judge's attempt to get all parties to agree to a form of relief failed and he proceeded to fashion one of the most complex decrees in the field of mental disabilities litigation.
State mental retardation system -- Certain general characteristics of Pennsylvania's mental retardation system were described including such factors as the state/county partnership arrangement, the influence of 1966 MH/MR Act, and the growth of community living arrangements. The influence of these factors on the implementation of the court ordered reforms was assessed.
Specific litigation activities within the mental retardation framework -- In addition to the organizational aspects of Pennsylvania's mental retardation system, the court's requirements were reviewed. The responsibilities of the Office of the Special Master (OSM) -- the court's primary enforcement mechanism -- were described, including the way in which the court orders would be monitored, and planning and resource development would be conducted. (A summary of the major orders that make up the decree is included in Exhibit 2-3).
| EXHIBIT 2-1. Department of Public Welfare Regions |
 |
| EXHIBIT 2-2. Population Trends at
Pennhurst Center (1966-1984) |
||||||
|---|---|---|---|---|---|---|
| Date | Clients
on Books |
Clients
in Residence |
Admissions | To
“Normalized” Environments* |
Transfers | Deaths |
| 1966 | 3071 | 2864 | 51 | 57 | 22 | 53 |
| 1967 | 2979 | 2778 | 45 | 43 | 22 | 56 |
| 1968 | 2761 | 2300 | 48 | 77 | 338 | 54 |
| 1969 | 2653 | 2029 | 94 | 184 | 158 | 61 |
| 1970 | 2534 | 1893 | 75 | 237 | 12 | 39 |
| 1971 | 2414 | 1780** | 92 | 169 | 49 | 40 |
| 1972 | 2217 | 1704 | 90 | 63 | 79 | 21*** |
| 1973 | 2047 | 1584 | 89 | 116 | 142 | 28 |
| 1974 | 1718 | 1488 | 75 | 97 | 84 | 23 |
| 1975 | 1619 | 1424 | 81 | 49 | 85 | 20 |
| 1976 | 1545 | 1399 | 46 | 35 | 22 | 17 |
| 1977 | 1448 | 1322 | 20 | 61 | 17 | 16 |
| 1978 | 1257 | 1211 | 1 | 76 | 1 | 26 |
| 1979 | 1170 | 1145 | -- | 43 | 1 | 18 |
| 1980 | 964 | 927 | -- | 38 | 155**** | 17 |
| 1981 | 912 | 857 | -- | 85 | 1 | 12 |
| 1982 | 817 | 669 | -- | 177 | -- | 11 |
| 1983 | 676 | 592 | -- | 69 | -- | 8 |
| 1984 | 576 | 399 | 1 | 180 | 5 | 9 |
| SOURCE: J. Gregory, Pirmann, Pennhurst Center, 1985. * Normalized environments include group homes/apartments, family care program, return to family, independent living, etc. ** This figure, from 1971 to 1980, represents clients counted in residence, but who were actually living at Pinehill Rehabilitation Center. *** This figure, from 1972 to 1980, includes deaths occurring at Pinehill. **** This figure represents the 155 clients discharged to Pinehill when that facility became a free standing ICF/MR. |
||||||
| EXHIBIT 2-3. Major Court Orders | |||
|---|---|---|---|
| “Original Order” (March 17, 1978) Judge Broderick’s initial order included the following requirements:
|
|||
| Order for
the Interim Operation of Pennhurst (March 5, 1979) This order includes the following requirements:
|
|||
| “Employee Order” (April 1979) This portion of the decree established an Office of Employee Services as part of the Office of the Special Master. The Office was created in order to provide counseling and guidance to those employees of Pennhurst State Center who lost jobs because of court-mandated deinstitutionalization. The order also included a schedule for the ultimate closure of the institution. This order was nullified by the circuit court on December 13, 1980. |
|||
| “Children’s Order” (June 8, 1979) This order requires the following:
|
|||
| “Hearing Master Order” (April 24, 1980) This order, which was necessitated by the ruling by the Third Circuit, mandated the following:
|
|||
| “Implementation Order” (March 2, 1982) This order included the following directives:
|
|||
| “Consolidated Order” (August 26, 1983) This order consolidated and updated the previous orders and added the following provisions:
|
|||
| “Final Settlement in Halderman v. Pennhurst” (July 12,
1984) The following are the major points included in the agreement reached between the plaintiffs and the defendants:
|
|||
What emerged in this analysis is a picture of an already complex system beset with the usual array of structural and political problems, and faced with meeting the very specific and immediate mandates of a complicated and far-reaching court decree. The weight of the discussion provided a view of both the limitations of judicial intervention in the area of mental retardation and the ways in which litigation can shape the course of reform in this field. It further highlighted the unique position of the Office of the Special Master in an ongoing state statutory, regulatory, and political environment.
In assessing the progress of implementation of the decree during this phase (which concluded in early December 1979), it is safe to say that the major constraint to compliance was the defendants' unwillingness to accept the results of the district court decree. Such resistance was manifest in continue appeals and a hope that Judge Broderick's decision would ultimately be overturned. This posture made it extremely difficult to secure the planning and funding commitments necessary to begin the movement of resident out of Pennhurst in the numbers envisioned in the original order. This singular fact, unlike aspects of litigation in other states where consent decrees have been signed, made the case and its implementation during this period, unique.
Other factors that influenced implementation were primarily derivative of larger system problems that would have constrained any major deinstitutionalization activity. They include the following:
Additionally, there were constraints that were peculiar to the litigation:
Given legal doctrine in the field, it is difficult if not impossible to force a state legislature to appropriate funds to implement the decree.
Current structural, organizational and political problems surrounding the relationships between the Office of the Special Master and the defendants constrained an easy and mutually trusting relationship.
There were no officially recognized county plans to guide the implementation of the decree in the Southeast Region.
The nature of the individual planning process on behalf of Pennhurst residents was long and tedious and resulted in numerous delays that purportedly discouraged the participation of some local providers.
The addition of Pennhurst case managers at the county level, while accepted in some counties, caused consternation and resistance in other counties.
The implementation of the Employee Order was constrained by the inability ot OSM to secure job placements and training resources.
The role of OSM with respect to planning caused duplication and confusion and removed accountability from the state and the counties.
Legal Developments. On December 13, 1979, the United States Court of Appeals for the Third Circuit issued its en banc opinion in the Pennhurst case. In a six to three decision, the court affirmed certain aspects of Broderick's decree and negated others. The three areas that were eliminated included: 1) the requirement that alternative employment be found for all Pennhurst employees; 2) the presumption that Pennhurst would eventually close; and 3) the portion of the decree banning all admissions to Pennhurst. The core of the decision was affirmed, however, since the court supported the community presumption.
With the appeals court decision in hand, there was an opportunity for the parties to negotiate a settlement. The appointment of Dr. Jennifer Howse, former director of the Willowbrook Review Panel in New York, generated optimism among the plaintiffs that agreement could be reached. Given her past position, Dr. Howse was seen as an aggressive and articulate spokesperson for the interests of mentally retarded persons.
Negotiations began early in 1980 and continued for several months. The major area of disagreement among the parties was the ultimate role of Pennhurst. Specifically, discussions focussed on how many persons should be considered part of the class and, therefore, eligible for community placement. After several proposals from both sides, there was still no consensus on the magnitude of placement.
OMR staff maintained that the sticking point was a disagreement regarding the speed of CLA development and client movement, while the plaintiffs asserted that discussions broke down both over the placement schedule and system improvement. Given the high hopes of each side, the inability to reach agreement left the parties with a great deal of bitterness and ill feeling. Each side felt that its position had not been respected and blamed the other for the ultimate failure of the discussions.
During the negotiations, Judge Broderick revised his original order to conform with the changes made by the court of appeals. The changes narrowed the scope of his original order and placed additional emphasis on the individual rather than the collective aspects of the remedy. Included among the changes in his order was the termination of the Office of Employee Services (OES) -- a unit that had been established at Pennhurst to ensure the protection of institutional employees. Upon termination of the OES, the Office of the Special Master (OSM) issued a special report outlining the multiple problems involved in finding alternative employment for Pennhurst employees.
Broderick also created the Hearing Master. The Hearing Master was directed to conduct individual reviews involving contested institutional discharges and where instititutional admissions were being sought on behalf of a class member. In filling this critical position Judge Broderick selected Michael Lottman, an attorney with broad background in the field of mental disabilities.
Finally, Judge Broderick allowed the Parent-Staff Association -- a group of anti-deinstitutionalization parents and Pennhurst employees -- to intervene in the litigation. This move further fragmented the case by introducing a group that was neither supportive of the plaintiffs nor entirely supportive of the defendants. Moreover, the Parent-Staff Association was receiving (and continues to receive) financial support from the American Federation of State, County and Municipal Employees (AFSCME), a powerful union with similar interests in maintaining institutions.
Following the court of appeals decision and the collapse of the negotiations, the defendants, and the Parent-Staff Association had to decide whether or not to pursue the case in the U.S Supreme Court. Each group had different motives for seeking certiorari, and some were reluctant about taking such an important test case to the Supreme Court. In the end, however, all parties sought review.
On June 9, 1980, the Supreme Court agreed to take the Pennhurst case. In granting certiorari, the Court agreed to hear arguments in four areas: the ability to enforce a private right of action either under the Developmental Disabilities (DD) Act or under general or federal civil rights provisions; the ability of the DD Act to support the breadth of the remedy in the Pennhurst case; the ability of the state Mental Health and Mental Retardation Act of 1966 to support the Pennhurst case; and the extent to which federal courts can intervene in state affairs. In addition to granting certiorari, the Court granted a partial stay prohibiting "involuntary" discharges from Pennhurst. Although there were several different interpretations of the Court's partial stay, in the end, Judge Broderick issued an order directing the Hearing Master to schedule a hearing for each Pennhurst resident for whom a community living arrangement had been prepared. The purpose of the hearing was to determine whether or not the transfer of Pennhurst clients to the community was in fact voluntary"
Pennhurst-Related Developments. Shortly after her arrival in Pennsylvania, Dr. Howse created the Pennhurst Implementation Team (PIT). The major responsibilities of the PIT included: serving as a liaison to OSM and other key actors; providing continuity between the policies developed for Pennhurst and the Southeast Region, and statewide policies; and ensuring that any positive benefits growing out of the litigation were expanded statewide. One of the first tasks taken on by the PIT team was the preparation of a staffing study at Pennhurst Center that recommended the addition of 107 direct care staff at a cost of $11 million. The relationship of the PIT to the Master's Office was complicated by the negative feelings surrounding the failed negotiations, and a growing polarization between the state and OSM staff regarding implementation of the court decrees.
One of the key concerns of the PIT and other OMR staff during this time was resource development for the Southeast Region. Though the Commonwealth's proposed implementation order showed 150 community placements for Pennhurst residents and another 100 slots for class members in the community by 1981, only 65 CLA slots were actually projected by OMR for "Phase I" of annual placement activity for Pennhurst residents. "Phase 2" included an unspecified number of placements as a result of the initiation of a community-based ICF/MR program.
Judge Broderick was not satisfied with the projected placements and circulated an implementation plan of his own in the form of a proposed order. The proposed order included movement of 150 Pennhurst residents to the community, the creation of 150 CLA beds for community class members and the movement of 150 out-of-region Pennhurst residents. State defendants criticized the order because it was unrealistic given the existing system capacity.
During this period the attorney for the original plaintiffs, David Ferleger, raised serious questions regarding suspicious deaths at Pennhurst State Center. The Commonwealth responded by commissioning a study by outside consultants regarding medical practices at the institution. The issues of adequacy of medical practices and resident abuse and neglect at Pennhurst were among the key areas to be monitored by OSM staff. Despite a rocky start, OSM monitoring activities and subsequent reports were accepted by Pennhurst staff and ultimately led to certain reforms and policy changes at the institution.
General System Developments. In addition to bringing on new staff to OMR, Dr. Howse also developed a reorganization plan for the office immediately after her arrival. One of the major organizational changes was the creation of a unit to oversee the development of community-based intermediate care facilities for mentally retarded persons (ICF/MRs). The new unit prepared a proposal for the development of small ICF/MRs as part of Pennsylvania's Title XIX plan of compliance. Regional Health Care Financing Administration (HCFA) staff had numerous reservations about the small ICF/MR proposal. Despite the reservations of the Federal Regional Office, OMR staff began to develop a plan to operationalize the ICF/MR program. In the initial planning stages, there was no special focus on the Southeast Region and the community placement requirements of the Pennhurst case.
Legal Developments. Up to the Summer of 1980, the Halderman v. Pennhurst case had been characterized by numerous appeals, stalemates and continuing confrontation among the parties. The ensuing period was much the same. The U.S. Supreme Court heard oral argument on the circuit court decision in Pennhurst, Judge Broderick signed an implementation order covering movement of class members into community living arrangements for the ensuing two fiscal years, and two related cases -- Romeo v. Youngberg and In Re Joseph Schmidt -- were decided.
As noted in the previous section, Judge Broderick interpreted the Supreme Court's stay to mean that no one could be moved from Pennhurst unless the transfer was "voluntary." The Parent/Staff Association, the group that originally requested the stay, disagreed with the Judge's interpretation and on November 4, 1980 went again to the Supreme Court to renew its request for stay and to ask that Judge Broderick suspend implementation of the decree. The state but not the counties also joined in the request. The Supreme Court denied the request immediately prior to the oral argument on the Pennhurst case.
On December 8, 1980, the Supreme Court heard arguments on the Pennhurst case. The primary focus of the oral argument on both sides focussed on whether the Developmental Disabilities Act supported the comprehensive remedy ordered by Judge Broderick. The defendants (petitioners) maintained that the Developmental Disabilities Act rested solely on the general spending power granted in Article I, Section 8, of the Constitution. Though the petitioners differed among themselves regarding the extent of federal enforcement authority under the Act, they all maintained that no substantive rights had been conferred by Congress. The plaintiffs (respondents) argued that Congress specified substantive due process and equal protection rights under the Fourteenth Amendment.
A week before the Supreme Court argument, the U.S. Court of Appeals for the Third Circuit rendered an en banc decision in the Romeo v. Youngberg case. The suit, which was originally filed in the federal district court in 1976, involved a profoundly retarded resident of Pennhurst State Center. While confined at Pennhurst, Romeo was injured on over 70 occasions either by injuries that were self-inflicted or the result of attack by other residents. The action was brought on behalf of Romeo by his mother who sought compensatory and punitive damages from the defendants because of violations of the resident's Constitutional rights under the Fourteenth and Eighth Amendments.
The plaintiffs lost in the federal district court but appealed the decision citing irregularities in the trial and in the Judge's instructions to the jury. In its ruling, the circuit court remanded the case back for a new trial noting that the district court, in an effort to distinguish the suit from a malpractice case, adopted a standard that was too rigorous in the context of a civil action. The circuit court proposed alternative jury instructions and requested the lower court to reconsider its earlier exclusion of expert medical testimony. Although Romeo won a favorable judgment in the circuit court, other legal hurdles remained before money damages could be awarded. In the meantime, the state defendants petitioned the U.S. Supreme Court for a writ of certiorari.
Another significant case, In Re Joseph Schmidt, was decided by the Pennsylvania Supreme Court. The case arose when the Allegheny County mental health and mental retardation administrator requested that Joseph Schmidt be committed to Western State Center. The Commonwealth intervened asserting that the Center was not an appropriate residential arrangement as required by the Mental Health and Mental Retardation Act of 1966. The lower court ruled that the responsibility for finding an appropriate placement for Schmidt was at the county level. The county appealed to the state supreme court asking for a clarification of which unit of government, the county or the Commonwealth has the responsibility to develop long term residential care. The state supreme court found that the 1966 Act created a right to care in the least restrictive environment and that the responsibility for such care rested with the Commonwealth.
On March 2, 1981. Judge Broderick signed an implementation order setting a placement schedule for Pennhurst class members. Many of those interviewed for the project were perplexed that Broderick chose this period to issue the order given the imminence of the Supreme Court decision. Several observers speculated that the Judge had become increasingly frustrated by the pace of movement of individuals out of Pennhurst and was concerned that resources that had been allocated for placement would revert to the state general fund.
The Judge ordered that, from March 2, 1981 to June 30, 1981, the Commonwealth develop 61 community living arrangements for Pennhurst residents (not covered by the children's order) and 9 similar arrangements for retarded class members in the Southeast Region. These figures coincided with the state's allocation letters to the counties for 1980-81. For the second year, the Judge ordered 150 Pennhurst and 100 community class members placed -- the identical targets presented by the Commonwealth to the court in May 1980. Broderick justified an additional 100 out-of-region placements by noting that since OKR had placed several hundred persons out of state centers in other regions, the state could therefore find community living arrangements for out-of-region Pennhurst residents.
Pennhurst-Specific Developments. Fiscal concerns preoccupied state an community staff during this time period. Several counties in the Southeast Region used part of their allocation for FY 1980-81 to cover reported short-falls in existing court-ordered placements and other unanticipated fiscal constraints. Certain counties attributed some of the deficit to the court requirements. On the other hand, OSM staff contended that some counties were interpreting certain IHP requirements too literally and providing certain services (nursing, etc.) at greatly increased costs. In order to rectify the budgeting problems, OMR staff prepared both short term and long term solutions: first, they covered the existing deficits through a modification of the 1980-81 allocation; and second, they developed special procedures for monitoring and controlling the use of expansion funds for Pennhurst class members.
At Pennhurst, an $800,000 contract was awarded to the Northeast Emergency Medical Association (NEEMA) to provide medical care for residents. The use of a contract, which included nine physicians and a medical director, was a response to concerns regarding deaths and other medical care issues at the institution.
Other significant developments during this period included the removal of OSM's appropriation from the overall Pennhurst budget and its inclusion as a separate line item in the Governor's 1981-82 proposed budget. As could have been predicted, this action drew the legislature's attention to OSM"s almost $1 million budget.
The activities of the Hearing Master were praised by most observers -even those who did not necessarily agree with his decisions. His approach was viewed as fair and his opinions literate and comprehensive. Up to this point, five of the Hearing Master's rulings had been appealed to Judge Broderick -four regarding community placement from Pennhurst and one regarding an admission to Woodhaven Center. In three of the five cases, the Judge upheld the Hearing Master; the other two required pre-placement visits before a "voluntariness" hearing could be held. The "pre-placement" decisions did not address any of the complex issues raised by the appealing parties -- they merely deferred a decision for a later time.
General System Developments. One of the major events during this period was the release of the Governor's proposed budget which provided $10.2 million in new program funds for OMR -- $2.3 million of which was targeted for new CLAs and $6.05 million for community ICF/MRs. Though the overall budget for the Department of Public Welfare was lean, mental retardation services continued to receive favorable funding increases.
The ICF/MR program -- with a proposed development strategy of 500 beds for FY 81-82 -- continued to encounter resistance by Regional Health Care Financing Administration staff. In order to achieve a resolution, a meeting was held in Washington D.C. with Central HCFA staff. At the meeting, it was suggested that the state prepare a waiver under Section 1115 of the Social Security Act to accomplish its objectives. Although a waiver was developed, OMR staff never submitted the request given the change in administration in January 1981. OMR staff continued to develop plans for the small ICF/MR program including clarifying agency roles and responsibilities, issuing program memoranda that listed the criteria for DPW approval of ICF/MR proposals, and preparing an implementation plan.
OMR staff were determined to use the ICF/MR program to promote small, community-based living arrangements. In a December 1980 memorandum, the size of new ICF/MRs was limited to a maximum of eight beds on non-contiguous sites. The proposed implementation plan called for converting large CLAs to ICF/MRs and developing new facilities to serve only "self-preserving" clients.
Legal Developments. During this period, the legal theories and theoretical legitimacy of the Pennhurst case, were challenged, the enforcement prerogatives of the Judge were tested, and the responsibilities of the defendants to comply with various aspects of the decree were reinforced.
The major legal event during this period was the decision by the U.S Supreme Court to reverse the Third Circuit Court of Appeals decision in Pennhurst and to remand the case to the lower court for consideration or reconsideration of the remaining Constitutional and state and federal statutory issues (i.e., Section 504. the Eighth and Fourteenth Amendments to the Constitution, the Pennsylvania Mental Retardation Act of 1966, and other sections of the Developmental Disabilities Act). On April 20, 1981, the Supreme Court ruled, in a six to three decision, that Section 6010 of the Developmental Disabilities Assistance and Bill or Rights Act (DD Act) does not create any substantive rights to "appropriate treatment" in the "least restrictive" environment.
Because the Court did not address itself to any of the legal issues considered by Judge Broderick, it provided only partial guidance to the lower courts regarding the future course of the litigation. However, though the Court's decision did not automatically vacate Judge Broderick's decree, it did alter the tone and momentum of the litigation. The defendants, in order to test the implications of the ruling, sought a stay of the decree from Judge Broderick pending the Third Circuit review. As he had on three other occasions, Judge Broderick denied the request.
In spite of the Supreme Court's decision, the Judge responded strongly to the state's withdrawal of funding from the Office of the Special Master. As mentioned in the previous section, the 1981-82 appropriation for the Master was placed in a separate line item of $900,000. Some observers saw this move as an attempt to prod the legislature to cut OSM's funds while Commonwealth representatives maintained that the shift to a line item was intended to clarify the issues for the legislature and to avoid misrepresenting the level of resources for Pennhurst Center.
In its final action on the budget in June the Legislature cut the Master's Office appropriation to $35.000. Following this action. the Commonwealth indicated to the court that it could not pay the court's monthly payment orders. In August 1981. Judge Broderick found the the Department of Public Welfare and Secretary Helen O'Bannon in contempt and assessed a $10,00 per day fine to run each day after September 2, 1981 that the payment orders were not obeyed. Several requests for stays by the Commonwealth were denied and the Commonwealth elected to pay the fines instead of OSM. In the meantime OSM staff "volunteered" their services without pay for a period of three months.
In another assertive action, the Judge issued an order to show cause why the Commonwealth and four counties (Chester County was in compliance) should not be held in civil contempt for failing to obey his March 2nd, 1981 "implementation order." The contempt hearings raised a number of key issues. For example. defendants and plaintiffs were using different definitions of who was placed and who was not. By July 31, 1981, OSM showed that of 90 persons slated for community residences, only 15 had been officially placed, while the Commonwealth's figures showed a total of 81 placed. During the course of the proceedings, the Judge also became concerned about the lack of county attention to the IHP process spelled out in the OSM guidelines.
The final legal event during this phase was the consummation of the first consent agreement since the Pennhurst case was decided in 1977. The plaintiffs and the City and County of Philadelphia agreed to settle placement issues raised in the civil contempt proceedings described above. By signing the agreement, Philadelphia did not admit contempt of the March 2nd Order, but agreed to make its required placements by September 30, 1981. Moreover, the Philadelphia defendants agreed to pay $15,000 in attorney's fees to plaintiffs' counsel for costs incurred during the contempt proceedings and to provide a performance fund as an expression of "good faith."
Pennhurst Specific Developments. For the most part, placements of Pennhurst class members in 1980-81 went more smoothly than in the previous year. By July 31, 1981, almost all residential and day programs had been developed. A number of constraints, however, were cited by the counties including delays in site identification, zoning obstacles, community resistance and client crisis situations.
Escalating costs of programs for class members became an issue during this period. Many of the per diems, according to county staff, fell in the $70.00 to $100.00 range. Some county staff maintained that the per diems were sometimes high because of the complexity of the clients' residential and day program needs. Such costs were not necessarily questioned by county commissioners since most of the programs are 100% state funded. State resources, however, were becoming more limited, especially in light of the 1981-82 budget for the Southeast Region. Since $8 million in new program funds were cut by the legislature, OMR staff had to adjust its funding commitments to the Southeast Region. Although the Commonwealth indicated to county staff that funding would be available to cover its court-ordered requirements, some counties were concerned about future funding for the court orders.
Two other important developments affected the on-going implementation of the Pennhurst Decree. As part of their response to Judge Broderick's March 2, 1981 order, OMR staff proposed to establish a "special management unit" in the Southeast Region with responsibilities for reviewing all TIHPs and IHPs for Pennhurst class members. The unit, to be based at Pennhurst, would include two staff persons -- one of whom was the former case management supervisor for Chester County. OMR staff anticipated that initially the unit would prepare revised IHP guidelines to "streamline" the procedures set out by OSM and would eventually take over responsibility for IHP monitoring from OSM. In addition, OMR staff proposed to use the data collection strategies developed by Temple University as part of the Longitudinal Study, to aid in monitoring individual clients. OSM staff were somewhat skeptical about the proposal since Temple's data was analyzed only on an aggregate basis.
Developments at Pennhurst State Center continued to focus on medical services including the medication reviews conducted by an outside medical consultant. According to Pennhurst staff, the presence of Dr. Ziring increased the level of interest and knowledge among direct care staff regarding medication issues. Meanwhile, a new medical director was hired as part of the NEEMA contract and plans for improved medical services for Pennhurst residents and for those residents making the transition to the community were initiated.
General System Developments. In July 1981, the Pennsylvania Legislature completed work on the state budget and approved a $22 million increase for the Office of Mental Retardation. This increase was $8 million lower than the Office had requested and as such, cut into plans for new programs. To make up the loss, OMR planned to use carry-over funds and dollars freed-up from conversions of existing CLAs to ICF/MRs.
OMR's proposal to develop small ICF/MRs was also completed during this period. Although OMR staff had never intended that the program become a major component in the implementation of the Pennhurst remedy, 112 community ICF/MR beds were included in the FY 1981-82 projections to meet the court-ordered requirements for the Southeast Region. Because of the budget cuts in new programs, OMR staff had to revise their original estimates of the number of ICF/MRs beds that would be developed throughout the state. A total of 225 beds as opposed to 317 beds would be converted to ICF/MRs and 200 new ICF/MR beds instead of 504 beds would be developed statewide. Further because of a ban on new construction, ICF/MRs would be limited to existing housing. As a result, only "self-preserving" clients would be served.
Other general system developments included a statement by OMR Deputy Secretary that Marcy State Center in Western Pennsylvania would be closed by 1982. Based on a feasibility study conducted the year before, OMR staff determined that Marcy residents were among the most appropriate candidates for community placement in the entire state system and that a significant savings would occur if all residents were placed in alternative living arrangements.
Legal Developments. In the second court of appeals decision in Pennhurst, Broderick was upheld on the basis of provisions of the state's Mental Health and Mental Retardation Act of 1966. The majority opinion distinctly avoided the more thorny Constitutional issues of and federal statutory questions remanded by the Supreme Court. Instead, the opinion stated that the entire superstructure of Broderick's decree could be supported by state law. The court of appeals ruling, though a victory for plaintiffs, left the remedy somewhat at the mercy of the state legislature and opened up the possibility of significant changes in the 1966 Act.
In the meantime, Judge Broderick congratulated OMR staff for their diligence in carrying out those tasks necessary to ensure rapid compliance with his implementation order. He did not find, however, that the county defendants had pursued implementation with the same zeal. As a result, the Judge found some of the counties in contempt but did not assess fines since compliance had been virtually achieved. The Commonwealth's performance may have influenced the Judge's receptivity to a reduction in the Master's Office and to the transfer of some compliance responsibilities to the Commonwealth.
Though the Commonwealth won kudos from the district court, compliance for fiscal year 1981-82 was somewhat clouded by the lack of state action on out-of-region placements. This issue, coupled with the possibility that in-region placements would not be completed by June 30, 1982, raised the possibility of additional proceedings during the summer.
On the O'Bannon contempt issue, Judge Broderick relieved the Secretary and the Commonwealth of the responsibility to pay the $10,000 daily fines. By early January 1982, the fund had swelled to approximately $1,200,000. The Judge agreed with the state that any further collection of fees would be inequitable since the funds paid by the Commonwealth were already in excess of the amount needed to operate the Master's offices. The Judge used the funds to reimburse OSM personnel for back pay and left the remainder in interest bearing accounts.
The Friend Advocacy program, the future role of the OSM and the transfer of monitoring responsibilities to the Commonwealth were the major issues that surfaced in another round of negotiations among lawyers for the defendants and the plaintiffs. Aside from discussions surrounding the Philadelphia County consent, this was only the second time that some or all of the parties had entered into serious negotiations. There were two major sticking points according to those interviewed: (1) the frequency of Commonwealth monitoring of class members living in the community; and (2) the timing of any transfer of monies from the contempt fines back to the Commonwealth. Although the parties were close to agreement on the first point, there was substantial difference of opinion on the return of fines to the Commonwealth. In the end, the negotiations collapsed. Ironically, shortly after the negotiations broke down, OMR program staff and OSM were able to reach an agreement of the transfer of monitoring.
The final legal event during this time period was the Judge's decision to vacate his order of July 14, 1980 which directed the Hearing Master to hold hearings on every individual being placed out of Pennhurst and to determine whether the placements were "voluntary." With the elimination of voluntariness hearings, the Hearing Master would be limited to cases where "beneficiality" of a placement was challenged by the client or his parents or guardians.
By the close of this period, there were several legal issues left hanging fire. First, the outcome of the Romeo v. Youngberg litigation was not known. Though Romeo differed in character from the Pennhurst suit, it offered the Supreme Court its first opportunity to define the Constitutional rights of institutionalized mentally retarded persons. The decision by the Court would certainly affect any future rulings on the Pennhurst litigation. Second, and somewhat related, it was not clear whether the Supreme Court would grant certiorari in Pennhurst for a second time thereby opening the issue of federal court jurisdiction in the enforcement of state laws.
Also, it was not clear what Judge Broderick would do to structure compliance beyond the end of fiscal year 1981-82. It was thought that the Supreme Court's action in both Romeo and Pennhurst would influence whether he would keep the pressure on the Commonwealth through an implementation order covering future placement.
Pennhurst-Specific Events. As discussed above, OSM and OMR staff were successful in forging agreements regarding the transfer of some compliance responsibilities including the approval of IHPs, monitoring of community facilities housing Pennhurst class members, and case manager training. The agreement included time lines for activities and a stipulation that OSM would remain involved for some period of time and would conduct joint reviews of several TIHPs and IHPs, as well as joint monitoring visits. County staff reaction to the shift was generally positive.
Funding for the remaining placements to be made under the March 2nd Order and possible future orders became increasingly more difficult to obtain during this period. As a result, funding for FY 81-82 placements was to some extent “boot-legged" from other sources. Because of an underestimate in the amount of federal funding accruing to the mental retardation program, a one time only surplus was generated. These funds, which were not part of the community services appropriation, were channeled to the counties through so-called "blue-black" contracts with the state. These agreements included augmented funding for the expansion of community programs under the order. County staff interviewed during this phase were concerned about developing new programs because of potential shortfalls in annualization funds for FY 82-83. Philadelphia and Bucks counties were told by OMR staff that adjustments would have to be made in their Pennhurst placement schedule because of the Governor's new "no growth" budget. Other counties saw that the only way to expand programs for the upcoming fiscal year was to increase the size of facilities.
The size of ICF/MR programs continued to be a point of contestion among certain counties. Delaware County decided against developing ICF/MRs since the state would not accept its proposals for two, 15 bed facilities. Other county staff were concerned about the costs of ICF/MRs given the additional federal requirements. In its December 1981 ICF/MR status report, the state reported a total of 20 sites, with 91 beds in the Southeast Region that would serve Pennhurst class members.
The implementation of out-of-region placements was stalled because the state continued to maintain that placing Pennhurst residents in other parts of the Commonwealth would jeopardize Title XIX "run-down" requirements in state centers and, therefore, federal reimbursements.
General System Developments. Events in the state legislature events occurred during this period signalled a potential change in the Commonwealth's orientation to programs for mentally retarded persons. House Bill 1824, introduced in the General Assembly in September 1981, called for significant changes in the Mental Health and Mental Retardation Act of 1966. The bill grew out of concerns about the spate of individual litigation in county courts of common pleas around the state, the decision in the Schmidt case, and the Third Circuit's action in Pennhurst. Specifically, the legislation was intended to make it clear that the 1966 Act was not an entitlement statute and that there was no presumption in favor of least restrictive settings. As a result of pressure from the Pennsylvania Association for Retarded Citizens and other interested groups, the bill was returned to committee by a vote of 23 to 25.
In other legislative actions, pressure from a variety of disaffected groups including parents of institutionalized persons, providers and some county staff resulted in the passage of Senate Resolution 63 -- a measure calling for an investigation of the Office of Mental Retardation and, in particular, the community programs it funds and supervises. The sponsor of the resolution had publicly criticized the Deputy Secretary of Mental Retardation for her lack of sensitivity to parents of mentally retarded persons.
In the Fall of 1981, an investigator was hired to staff the effort. The investigator, a former county district attorney with no background in mental retardation, spent several months conducting site visits and obtaining information about the mental retardation system. Prior to hearings on the results of the investigation, a preliminary report was prepared. Some of the concerns cited in the report included high turnover among CLA staff, failure to consult parents regarding placement of family members, and limitations on the size of ICF/MRs.
The role of parents of mentally retarded persons in decisions affecting the placement of their adult or minor child was a primary issue in the investigation. In partial anticipation of legislative action on the problem, the Department of Public Welfare issued a policy memorandum regarding parent participation in late 1981. Although the policy provided parents with access to the court of common pleas as a last resort, some parents felt that this option was not sufficient. As a result, a "Parents Rights Bill" was introduced in order to place the burden on the State to prove that the recommended placement was the correct one for the adult client and to force the Commonwealth to pay parents' legal expenses no matter what the outcome of the appeal.
During this time period, state licensing standards for CLA programs were finally published. The standards applied to all community residences 24 hour care was provided to one or more mentally retarded persons. OMR staff anticipated that all CLAs and PLFs would be licensed within one year. In those cases where facilities were found to be deficient, it was the state's intent to provide six month provisional licenses. Some county provider staff were concerned that the new standards might result in significant increases in per diems for those facilities not in compliance.
Legal Deyelopments. During the last phase, developments in the litigation revolved primarily around two actions of the Supreme Court -- one that directly affected the case and one that could indirectly affect the course of the lawsuit. Specifically, the Supreme Court granted certiorari in the Pennhurst case for the second time and also issued an opinion in the Romeo v. Youngberg case. Further, Judge Broderick issued a far-reaching order that will probably result in the disappearance of the Special Master. The remaining legal events mirror those of past periods and include ongoing appeals of almost every facet of the Judge's decree and another intense but failed set of discussions regarding the possible agreement in the case.
The Supreme Court decision in the Romeo case was relevant to the Pennhurst litigation because it was the first time that the Supreme Court considered the substantive constitutional rights of involuntarily committed mentally retarded persons. As such, the opinion suggested some of the reasoning that might be applied by the Court in its second hearing of Pennhurst case. In reviewing the Third Circuit's opinion in Romeo, a majority of the Supreme Court found that involuntarily detained mentally retarded persons have the following constitutionally protected rights: reasonably safe conditions of confinement, freedom from unreasonable bodily restraints, and minimally adequate training as reasonably may be required by these interests.
With respect to a "right to treatment" the Court defined the term narrowly to mean habilitation that would diminish Romeo's dangerous behavior and therefore avoid unconstitutional infringement of his safety and freedom of movement rights. This interpretation was far different from the court of appeals finding that such persons have a right to treatment in the least restrictive fashion and according to accepted medical practice. The Supreme Court also noted that in determining whether an individual's constitutional rights had been violated, his liberty interests must be balanced against relevant state interests. Finally, the Court stated that in ascertaining liability, the Constitution only requires that courts make certain that professional judgment is exercised and that judges should not take sides regarding which of several professionally acceptable choices should have been made.
In June 1982, the Supreme Court once again granted certiorari in Pennhurst. The major foundation of the defendants' request for review was the Eleventh Amendment and the multiple ways in which it shields states from inappropriate instrusion by the federal courts. The nub of the defendants' position was summed up in the following quote from their brief: "Unless it [the Third Circuit] is reversed, the decision will give federal courts a free hand in the management of state programs despite the absence of any federal interest at all."
There was much speculation regarding the reasons why the Court granted certiorari including the possibility that the Court may want to address a much larger issue -- whether federal courts should avoid ruling on Constitutional questions if a state law claim is available. If the Court is interested in clarifying issues regarding OSM, it will have to be in retrospect since, on August 12, 1982, Judge Broderick issued an order requesting that the Special Master develop a plan for phase-down of her operations by the end of the calendar year.
The timing of the Commonwealth's petition for certiorari caught a number of key observers by surprise, including the Secretary of Public Welfare. In mid-May, a series of intense discussions had begun between the Secretary and the President of the Pennsylvania Association of Retarded Citizens (PARC) in an effort to find some common ground that might eventually lead to agreement between the two parties. Several days after the discussions were initiated by the Secretary of DPW, PARC's President presented a proposal specifying seven major steps including foregoing a request for Supreme Court review of the Third Circuit ruling. Soon after the working document was submitted, it was learned that the Commonwealth had already filed the petition. As a result, discussions were terminated and once again the parties were left feeling frustrated and disillusioned with the process.
Despite the continuing inability of the parties in the Pennhurst case to reach a consent agreement, Judge Broderick moved in the Fall of 1982 to phaseout the Special Master -- an action that caused some consternation among the plaintiffs. The Judge's order stated that the dissolution of OSM did not signal a diminution of vigilance on his part. Moreover, the order made no mention of the Hearing Master -- an entity that will almost certainly continue, perhaps even in an expanded capacity.
Certain individual cases were also noteworthy during this period including the April Saures case in Allegheny County. The Saures case was similar to the Schmidt case, ruled on earlier by the Pennsylvania Supreme Court. As in the Schmidt case, the Pennsylvania Commonwealth Court found that the state must pay 100% of the cost of community living arrangements for April Sauers. This ruling further reinforced the interpretations of the State Supreme Court and the Third Circuit regarding the requirements of the Mental Health and Mental Retardation Act of 1966.
Within the Pennhurst class, one group of individuals -- persons who are non-ambulatory -- have been particularly difficult to place. The problem is locating accessible housing that meets fire and life safety standards. One non-ambulatory community-based class member, M.H., has been on the waiting list for sometime for an accessible and appropriate placement. Planning began for M.H. and a site was selected last fall. The site was eventually rejected by the Commonwealth because of life safety problems. When the family was finally notified that M.H. would not be placed, a motion was filed in district court seeking emergency relief for M.H. In the meantime, the county shifted its energies to the placement of ambulatory class members -- a choice that some observers felt was necessitated by the court's placement deadlines.
Pennhurst-Specific Developments. It was evident during this most recent phase that the five county defendants had a sense of urgency in meeting the June 30 1982 deadline for community placements. Although only one county met all of its required placements by June 30, the remaining counties were very close to full compliance. There was concern, however, that in order to meet the deadline certain counties "cut corners." Specifically, it was suggested that trial visits for certain class members were not long enough. This issue was eventually brought to the attention of the Hearing Master and OMR. Counties were notified that all procedural safeguards for class members had to be observed, but certain county staff maintained that short cuts had not been taken; on the contrary, they asserted that a great deal of time had gone into developing appropriate placements. Other complimenting the deadline were not new. Zoning disputes, community resistance and ICF/MR delays all presented obstacles to meeting full compliance with the March 2nd Order.
With respect to the remaining part of the implementation order, the Judge denied the Commonwealth's request to eliminate the 100 "out-of-region" placements. Although the Commonwealth contended that there was inadequate funding to implement the out-of-region placements, Judge Broderick strongly disagreed and pointed to the Marcy and Harrisburg Mental Retardation unit deinstitutionalization efforts as evidence of the state's ability to fund institutional reductions in other regions. He gave the Commonwealth until September 30, 1982 to comply with placement requirements.
Following the Judge's ruling, planning for the out-of-region placements was accelerated. By June 1982, 90 persons had been identified as candidates for movement to a total of 14 counties. There was some concern that the three month time limit would be insufficient to guarantee adequate planning and that the needs of these clients could strain the resources of some of the smaller rural counties. Once the out-of-region placements are made, the question of who and how those placements will be monitored must be addressed.
During this phase, the Special Management Unit (SMU) in the Southeast Region became operational and took over monitoring and IHP reviews from the office of the Special Master. SMU staff agreed to monitor each Pennhurst placement once a year and to monitor the conditions of each class member within 120 days of placement out of Pennhurst. Although there were some concerns with certain aspects of the monitoring process, most of those interviewed were pleased with the SMU's and Temple's efforts.
General System Developments. Funding issues dominated general system developments during this time period. The new 1982-83 budget passed by the legislature contained few surprises for mental retardation except that funding for interim care and CLAs was separated into two line items. Several county staff expressed reservations about this move while provider representatives and others were pleased to see that interim care -- the primary funding source for private licensed facilities (PLFs) -- was clearly visible in the state budget.
Two proposed per them ceilings for ICF/MRs were issued in June 1982. The ceilings were $87.70 for urban areas and $77.27 for non-urban areas. The proposed ceilings created an uproar in the provider community and dismay and frustration in some counties. A number of key areas were at the center of the controversy. They included: the lack of differentiation in the regulations among types of ICF/MRs and the clientele they served; the data used to develop the proposed caps; and the distinction between private and public ICF/MRs in the application of the ceilings. The Commonwealth ultimately postponed the final regulations through the end of the fiscal year or until an acceptable rate methodology had been developed. If applied as proposed, the ceilings could have eliminated many providers already on line to develop ICF/MRs for Pennhurst class members.
Finally, the implementation of the recently enacted licensing regulation for community-based residential facilities created some fiscal problems for certain residential programs. A few of the large PLFs requested increases in their rates because of the upgrading required to meet the standards. However, state and regional OMR staff noted that only a few providers increased per diems and that so far, there was no statewide trend.
Briefly, some of the major events that shaped the implementation of the federal district court decree during the final study period (1982-1983, 1983-1984, and 1984 to date) include the following:
Termination of the Office of the Special Master. As of the end of December 1982, the Office of the Special Master was officially terminated by the court. Many of the functions previously carried out by OSM were transferred to an entity set up by the Commonwealth defendants. The new monitoring mechanism, the Special Management Unit, has responsibility for review of transitional habilitation plans (TIHPs) and individual habilitation plans (IHPs), on-site monitoring, and supervision of the certified advocates. Monitoring of individual client progress is being carried out by the Temple University Developmental Disabilities Center. The Hearing Master continues to conduct hearings in those cases where there are exceptions to community placement or when reinstitutionalization is proposed.
Order of January 14, 1983. The January 14 order sets out an implementation schedule for the 18 month period ending on June 30, 1984. The order requires the counties and the Commonwealth to provide community living arrangements to 143 residents of Pennhurst, 81 other members of the plaintiffs class living in the Southeast Region, and 50 Pennhurst residents from outside the Southeast Region.
Community Services Waiver Application. As part of their plan of compliance to the court following the January 14, 1982, order, the Commonwealth noted that an application for a Medicaid waiver under the provision of the Omnibus Budget Reconciliation Act of 1981 had been submitted to the Health Care Financing Administration (HCFA) on behalf of Philadelphia and Delaware Counties. Following submission of the plan, applications were also submitted on behalf of Montgomery, Bucks, and Chester Counties. The combined impact of the implementation of the waiver and the judge's order would reduce the population of Pennhurst to 200 by fiscal year 1985-1986. As of early 1985, the waiver applications for the suburban counties had not been approved and the federal audit of the Philadelphia waiver resulted in numerous exceptions.
Supreme Court Arguments. During this period, the Pennhurst case was argued twice in the Supreme Court -- once in February and once in October. The February argument centered primarily around the ability of the district court to order a major reordering of the state mental retardation system ostensibly on the basis of a state law claim. The theoretical assertions revolved around the interpretation of the Eleventh Amendment and the principle of "comity" as they related to the facts in Pennhurst. The argument also addressed the powers of the Office of the Special Master. In June, the Supreme Court announced that they had been unable to reach a decision and that they would rehear the case. The second argument, in October, focused on the same issues, with the exception of the Special Master.
Legislative Task Force. Following a legislative investigation and a special report on the mental retardation system in the state, three legislative task groups were convened to develop revised state legislation. The three groups looked at placement procedures, definitions, and state and county responsibilities, respectively. The task groups included providers, consumers, and county officials in addition to legislators. The reports of the task force groups have been presented to the legislature, but to date no action has been taken.
Consolidated Order. In August 1983, Judge Broderick issued a "consolidated" order which encompasses -- where relevant -- the provisions of all previous orders, and adjusts provisions to reflect changes made over the past year (e.g., the dissolution of the Office of the Special Master). For instance, the responsibilities vested with OSM to develop county plans for class members have not been shifted to the state. OSM's previous monitoring responsibilities at Pennhurst were given to the Special Management Unit. The Hearing Master's responsibilities have been expanded to include jurisdiction over class members who are subject to involuntary commitment in mental health facilities. The Commonwealth and counties have filed exceptions to the order. The state's primary concern is that the Judge has essentially mandated the Special Management Unit as party of the decree and also expanded its powers beyond those to which the state agreed. The new order also requires the Commonwealth to develop and submit plans for class members and non-class members through the end of June 1985.
Commonwealth Plan. In November 1983, the Commonwealth defendants submitted a plan to Judge Broderick in response to the consolidated order. The plan described how placements would be carried out for the ensuing two and a half years. It was in this plan that state officials indicated that Pennhurst Center would be closed and that closure would be accomplished by 1986. The plan called for a small residual population that would be transferred to other institutions.
Supreme Court Decision. In its second opinion in the Pennhurst case, issued on January 23, 1984, the Supreme Court -- in a five-to-four decision held that the sovereign immunity principle of the Eleventh Amendment to the Constitution prohibits a federal district court from ordering Pennsylvania state officials to comply with state law. The ruling reversed the Third Circuit's earlier ruling (which followed the Court's 1981 decision noted above) that affirmed the district court's decree in Pennhurst based on state law grounds alone. The Supreme Court decision significantly altered the traditional jurisdiction of federal courts in state law matters and may force litigators with state as well as federal claims to file in state and federal court respectively. This second Supreme Court ruling in Pennhurst is in keeping with the decision in Romeo v. Youngberg insofar as its emphasis on judicial restraint and its deference to state law and state mental retardation professionals is concerned.
Consent Agreement. On July 12, 1984, ten years after the original lawsuit was filed, the Commonwealth and county defendants and the plaintiffs announced that they had reached consensus on the terms of a consent agreement. The only party not included in the agreement was the Parent Staff Association. The consent agreement included the schedule of placement included in the Commonwealth's November 1983 plan and spelled out ongoing responsibilities for the preparation of Individual Habilitation Plans, the conduct of monitoring and quality assurance, and the maintenance of a placement review forum for those protesting the provisions of an IHP. The agreement also narrowed the definition of the class by eliminating ongoing monitoring for those on the Pennhurst waiting list and those at risk of being institutionalized at Pennhurst.
Hearings on the Consent Agreement. In September 1984, Judge Broderick held a public hearing on the provisions of the consent agreement in order to determine whether there were any objections. The only group that lodged a protest was the Parent Staff Association which continued to maintain that Pennhurst State Center should not be closed. Because of the delay in the approval of the waiver (on which much of the financing for class member placements depended) the Judge postponed his decision on the agreement. After one postponement, the Judge scheduled another hearing in November which was held in his chambers. At that time, the Commonwealth voiced its concern about its ability to carry out the agreement because of the lack of approval of the state's waiver applications. Finally, on December 5, 1984, the Judge held another hearing at which time the Commonwealth announced its intention to comply with the agreement regardless of the disposition of the federal waiver. It is expected that the Judge will soon announce his approval of the consent agreement.
New Role for Pennhurst. In January, 1985, Governor Richard Thornburgh announced that the Pennhurst State Center facility would be converted into a state medical facility for veterans by 1986.
As part of the qualitative assessment of the conduct of the Pennhurst litigation, four special analyses were conducted. These focussed assessments, termed "implementation analyses," were directed at particular aspects of the environment in which the decree was being carried out. The subject of each analysis was selected jointly by project staff and the Pennhurst Advisory Council. Candidate issues were drawn from the findings of the Historical Overviews described in Chapter III. The implementation analyses have made it possible to go beyond the broader historical analysis of the litigation to a fuller exploration of one or more key element in the use of litigation to bring about social change. The analyses have also allowed staff to examine factors in the implementation of the Pennhurst decree against the backdrop of political, sociological, organizational, and legal theories surrounding social change as discussed in a wide body of literature. Finally, in the last three years of the study, the implementation analyses have expanded to include comparative analyses in other states thereby increasing the relevance of the study and its utility to state and federal policy makers.
The following section describes three of the analyses including the methods used to secure information and the major findings reported. The fourth analyses, which summarizes the impact of the decree, comprises Chapter 10 of the report.
Nature of the Issue. The use of special masters appointed by the courts to supervise the implementation of broad-based structural reform is a relatively new phenomenon -- particularly in the area of public health and human services. Masters are judicial deputies appointed by the court to assist in the conduct of complex lawsuits. These officers traditionally have been utilized to superintend such things as the complex calculation of damages -, or to aid in corporate reorganization and dissolution. Further, masters have conventionally been used to oversee remedies directed at the private sector. More recently, masters have been used to carry out injunctions against public sector agencies such as school districts in desegration cases and state governments in prison reform cases. The appropriate role and function of masters in litigation directed at reform of large bureaucratic programs is a subject of much debate and controversy. The purpose of the Implementation Analysis for Year 1 was to shed some light on the issue by focussing on the case history of the master appointed to supervise the district court's decree in the Halderman v. Pennhurst State School and Hospital case.
Method and Objectives. In order to provide a comprehensive picture of the Office of the Special Master (OSM) in Pennhurst, the analysis encompassed both the legal context within which the master functions and the larger bureaucratic milieu which is the object of the court's intervention. Because the appointment of the master in Pennhurst is part of a larger legal tradition, OSM was also assessed in light of the experience of other masters in related litigation.
The analysis drew both from the legal literature on complex litigation and compliance mechanisms, and from the political science and public administration literature on implementation and program change. Further, the analysis encompassed interviews and document reviews conducted as part of the preparation of the Historical Overviews for Year 1. Thus, the analysis provided two perspectives -- one that concentrated on the legal expectations and parameters that characterize compliance mechanisms like the Office of the Special Master, and the other directed at system reform and bureaucratic change.
(1) Lack of Consent
Almost all of the cases in the mental disabilities field have ultimately been settled by consent decrees. A consent disposition has important implications for the efficacy of the compliance mechanism selected, the strategies that it employs and the resources it requires to bring about change. The significance of consent is highlighted in a report by the external Court Monitor appointed by the federal district court in Massachusetts to ensure compliance in five institutional class action suits (Horowitz, 1979):
It may be useful to clarify here the significance of the fact that the decrees were entered by consent of the parties. The spirit of seeking agreement has been fundamental to the success of the litigation to this point. Despite the far-reaching powers of the federal court, there can be no doubt that better and quicker results are achieved when all parties make an effort to cooperate and reach a common ground. (p. 4)
Achievement of a consent decree in institutional litigation does not necessarily guarantee the success of reform or even the spirit of cooperation connoted by a consent disposition. David D. Gregory (1980), Special Master in the Wuori v. Zitnay case in Maine, illustrates this point in a report to the district court:
The State's failure to comply with the Court's (consent] decree remains substantial . . . The State could have made much greater achievements if all State agencies bound by the decree had given their active, informed cooperation. The administrative complexity of carrying out the decree in the absence of just such cooperation has prolonged the time needed for compliance without bring any countervailing benefit to the state. (p. 1)
The fact that Judge Broderick could not persuade the parties in the Pennhurst. case to negotiate a consent decree also had an impact on the remedy adopted by the court. In the absence of consent or of any proposed orders from the defendants that the court could adopt, the character of the initial and subsequent orders has been significantly influenced by the plaintiffs. As a result, the defendants view the orders as instrusive and unrealistic and have little stake in the remedy since they have not participated in its development.
(2) Limited Enforcement Powers
Given documented problems of other court appointed implementation mechanisms in institutional and deinstitutional litigation, the plaintiffs in the Pennhurst case attempted to structure a remedy that embodied comprehensive planning and compliance duties. The master mechanism ordered by Judge Broderick encompasses both broad and individual planning responsibilities, needs assessment activities, monitoring tasks to ensure compliance with basic standards at the institution, and a variety of other responsibilities ranging from the recruitment and training of case managers to the certification of advocates for individual clients.
Notwithstanding the broad powers vested in the Office of the Special Master in Pennhurst, the ability of the Master to enforce compliance with the decree has been hampered because of the limited sanctions available to the court. The only real sanction is the contempt power which, in cases like Pennhurst, is generally regarded as a last resort -- in part because it must be directed at an individual or individuals within the broader bureaucracy implicated in the litigation. By focussing the punishment for non-compliance on one actor, the larger, more complicated wrongdoing is ignored. The ability of the court to enforce a complex decree is further complicated by the court's lack of power to reach through the bureaucracy to the legislature which is ultimately responsible for providing funds for the reform. Though some judges, such as Johnson in Wyatt v. Stickney (1972) have threatened to circumvent the legislature by attaching public lands or taking some other action that would inhibit the legislature's ability to control specific public funds, by and large courts have been unwilling to take the legislature on directly.
The court is thus limited to negative and to some extent blunt powers in enforcing its decrees. It has no bonuses or rewards to hand out to compliant defendants except the ultimate disappearance of the court and the master from the scene once the aims of the decree have been fulfilled. What the court, and therefore the master, are left with is some form of psychological reinforcement or judicial back-patting when the defendants have done well.
(3) Involvement in Individual Cases
The Special Master's compliance functions reflect a broad and deep involvement in the day-to-day implementation of the decree. The Master's responsibilities begin with the condition of the class in the institution, carry through the initiation of individualized habilitation plans, and continue to placement in the community and beyond. Compliance activities entail review and approval of both individual and collective plans for class members. They span such substantive areas as quality assurance, program development, client advocacy, institutional operations, program design, client and family grievances, fiscal auditing. and staff training. In short, OSM"s compliance functions touch on almost every aspect of the traditional delivery system for mentally retarded individuals. Because of the deinstitutionalization thrust of the decree, however, the institutional compliance functions of the Special Master in Pennhurst are relatively limited and focused primarily on life safety, sanitation and other mechanistic aspects of the program at the institution.
The individuated nature of the remedy in Pennhurst is a significant factor in diverting the attention of the Special Master from the broader structural aspects of the decree. Involvement in individual cases siphons off energy and places the master squarely in the middle of debates reflecting conflicts in professional judgment. It creates a sort of schizophrenia in the operation making it difficult to be both detached planner and general system monitor, and also analyst and arbiter of particular cases.
(4) Separate and Countervailing Agency
The establishment of OSM as an agency separate and removed from the bureaucracy to manage the implementation of the Pennhurst decree was directly motivated by the plaintiffs' frustration with past bureaucratic performance. The creation of new agencies to solve old problems is a tactic frequently used in government as evidenced by the establishment of special White House commissions, Congressional task forces, and elite semi-autonomous bureaus reporting directly to agency administrators. The isolation of such enterprises from the ongoing bureaucratic machine, however, has drawbacks. As Pressman and Wildavsky (1979) report in their book Implementation:
The cost of independence from ordinary bureaucratic constraints turns out to be loss of contact with the very political forces necessary to preserve the thrust of the organization. (p. 129)
In the case of the Office of the Special Master, the isolation and separateness of the agency created conflicts and tensions both because of its perceived favored position, and also because it ultimately relied on the bureaucracy to carry out the specifics of implementation. It must guide the course of implementation, but it cannot become the bureaucracy without jeopardizing its autonomous and unique character -- and ultimately its moral and legal authority.
Establishing a working relationship with the bureaucracy in order to accomplish the ends of the litigation has been difficult for OSM. Part of the problem is that OSM staff are perceived as being highly ideological and unbending. The reaction of a bureaucracy to this sort of "cause oriented" group is described by Eugene Bardach (1977) in The Implementation Game:
A not insignificant number of policies and programs originate in the desire to extirpate real or imagined evil. Such policies create implementation opportunities for activists whom many political interests will perceive as "hotheads," "extemists," or "zealots." A couterreformation then sets in. A political coalition emerges to scrutinize, criticize, and in some cases to terrorize the agency charged with assaulting the stipulated evils. (p.93)
In the case of OSM. however, the problem is not so much the actual values of OSM staff -- many of whom previously worked in the system -- but the seemingly rigid values and time tables included in the decree. Nonetheless, the perception of OSM staff by those forced to comply with the decree is not that different from the picture presented by Bardach.
The various structural and political factors have conspired to create a "we-they" mindset in OSM and conversely in the bureaucracy. Polarization of OSM is the result of its continually frustrated attempts to influence, let alone move, the bureaucracy to make those changes necessary to facilitate the deinstitutionalization process. On the other hand, the bureaucracy is increasingly alienated from what it sees as a "foreign" agency with power to direct its actions but totally outside of its control.
(5) Lack of Control Over Policy-Making
Though the Master has a quasi-policy making function in that she suggests proposed orders and devises related "policies," she is not a policy maker in the broadest sense. The sources of broader policies that affect the system are the Governor, the Department of Public Welfare and the legislature. The implementation literature argues strongly that the separation of policy making from the operationalization of a program is fatal to the success of reform. This principle is not directly relevant to OSM's situation, since OSM does not monopolize policy making in the system and is not directly responsible for implementation, but the principle does have some resonance. The need for connectedness and coherence between policy and implementation is as relevant in court ordered change as it is in legislative or bureaucratic change.
(6) Conflicts with the Bureaucracy
OSM seems inextricably drawn into areas traditionally reserved for the bureaucracy because of a perceived failure on the part of mental retardation administrators and the very implementation instrumentalities established by the decree. Though OSM can never become the bureaucracy, the court placed it squarely in the center of the system. As a result, it is difficult to tell where even the most prudent Master would draw the line between his or her authority and bureaucratic turf.
(7) Lack of an Overall Plan
Though the court order does not specify that the Master is responsible for developing an overall plan or task description to guide implementation, several of the parties have expressed the need for such a document. In particular, county personnel -- who are responsible for the bulk of implementation detail -- see a distinct need for such a document. They argue that an overall plan would be particularly useful in spelling out the expectations of the Master including the schedule of implementation and the specific actors designated to carry out particular tasks.
In Judge Broderick's original order, OSM was given the responsibility to develop county plans for the Southeast Region. OSM and others argued that the development of detailed county plans should not be the responsibility of the Master. It is the county administrators who are most familiar with the specific problems at the local level and it should therefore be their responsibility to prepare the plans. Further, if the goal of the litigation is to institute new practices, the counties should adopt plan preparation as an ongoing responsibility. It would appear that almost everyone, including the court, has accepted the inappropriateness of the Master's role in this area. However, no substitute plan requirement was adopted. As a result, there are also no plans to guide the implementation of county responsibilities under the decree.
(8) Compliance v. Planning Duties
The bulk of the activity conducted by the Office of the Special Master falls into the area of compliance. The extent of compliance responsibilities is far broader than the range of general planning duties. It is in part this imbalance between general system functions and compliance detail that explain the drain of OSM resources into individualized crises and particularistic controversies.
(9) Constraints to Compliance
The Master's ability to secure an acceptable level of compliance from the counties is complicated by the nature of state law and the counties' position in the overall mental retardation delivery system. Though counties have the responsibility for carrying out the law at the local level, the bulk of the funding comes from the state as do the policies that govern program content. OSM's ability, therefore, to influence and goad the counties into compliance has distinct limitations.
Further, though OSM can apply pressure to the counties to generate residential and support services for the class, counties are reliant on the private sector to provide needed services. The county system in Pennsylvania is based on purchase of service arrangements with the county administrator and his staff performing only administrative, monitoring and fiduciary functions. Thus the success of deinstitutionalization goals is to a large extent dependent on the service marketplace.
(10) Conflicts with Case Management Functions
OSM's involvement in individual cases may undermine the role of the county case manager. According to some case managers interviewed, continued involvement of OSM in the details of implementation has been aggravating. From their point of view, the Master is seeking "perfection" from a complex and already strained system. Further, the tenacity with which OSM staff have carried out their functions in this area leaves them vulnerable to accusations that they are merely substituting their own judgments for those of county bureaucrats -rather than ensuring overall compliance with systemic norms.
Nature of the Issue. In a departure from conventional right-to-treatment cases, the Pennhurst suit and others like it question the legitimacy of the institution itself. These second generation cases assert the rights of institutionalized mentally retarded persons to equal treatment and freedom from alleged discrimination enforced on them in large custodial facilities. Remedies in these lawsuits became even more complex as they reached into the less walled off and more complicated realms of community-based systems of care. As the second and third offspring of Wyatt have evolved, the interests of more and more groups have been implicated in court actions including institutional employees, parents of institutionalized children, parents of children in the community, community caretakers, and other human service providers and administrators.
The drama and controversy surrounding cases in this field have drawn considerable attention to the legal theories and strategies that characterize the litigation. However, very little attention has been paid to the complicated interaction between the nature of court mandates for reform and the constellation of resource, leadership, organizational, political, and systemic variables that exist within a particular state. Even the impact of seemingly unidimensional right to treatment suits -- one wrong, one remedy -will vary depending on the complexity and internal dymanics of a particular state. As more and more divergent interests become drawn into a decree, the character of the state system becomes key to an understanding of the role of litigation in creating the changes desired by plaintiffs.
Method and Objectives. The response of states to litigation in the area of mental disabilities has been varied and wide-ranging. Even within a particular state, the official position vis-a-vis the court can shift in response to changes in the level of resources, the force of public opinion, a turn-over in political leadership, and pressures of competing constituencies. Some states have readily entered into consent agreements with plaintiffs. Some states, even after consent agreements have been signed, have resisted the court's jurisdiction. Other states have begun to reach the limits of cooperation under consent agreements and are attempting to terminate the court's oversight. Still other states, like Pennsylvania, continue to contest the court's right to intervene in the state system.
The purpose of the Implementation Analysis for Year 2 was to explore the factors that dictate a state's reaction to more complex forms of mental disabilities litigation. By using Pennsylvania as a case example and contrasting it with selected comparison states, it was possible to gain insight into state policy-making, the influence of particular constituencies, the internal constraints that exist within a system, the cohesiveness of state leadership, and the relative openness of a system to external changes. A close examination of externally imposed deinstitutionalization mandates in Pennsylvania and other states also sheds light on the tensions surrounding community placement and the limits of the state's ability to hasten its activates in a politically charged atmosphere.
Specifically, the analysis accomplished the following:
In order to provide a framework for the analysis, a set of initial hypotheses which seemed likely to explain at least some of a state's reaction to broad-based litigation was developed. The hypotheses can be stated as follows: "A state's reaction to litigation will vary according to --"
In order to gather material for the analysis, several steps were taken. First, information was sought from the Commonwealth's Deputy Attorney General assigned to Pennhurst, the current Deputy Secretary of Mental Retardation, past Deputy Secretaries of Mental Retardation who held their positions during relevant stages of the litigation, plaintiffs' lawyers, county officials, Pennhurst State Center staff, and representatives of consumer organizations.
In addition, two states were selected for in-depth comparisons -- Maine and Michigan. These states were selected for several reasons. First, they both are currently the targets of suits that are roughly similar to Pennhurst. Second, unlike Pennsylvania, they both have entered into consent agreements. Third, Maine and Michigan represent two distinct types of states; Maine is a fairly rural state with characteristics very different from Pennsylvania, and Michigan is an industrial state with characteristics similar to those of Pennsylvania. In each state, key actors were identified including state program officials, institutional administrators, consumer representatives state legal representatives, plaintiffs' lawyers, and local program staff.
Because the issues to be considered in the analysis centered around organizational behavior, project staff also reviewed the public administration literature regarding the response of organizations to externally generated change. Though there is very little written on the response of state organizations to changes embodied in mental disabilities litigation, the general principles and theories advanced in the literature were helpful in describing the phenomena under analysis. Finally, staff reviewed materials from the two comparison states and other states facing similar court mandates including Minnesota, Nebraska, New York, Massachusetts, and the District of Columbia.
Findings. Before proceeding to a summary of the usefulness of the various hypotheses, it should be noted that there were sensible explanations for the state's posture that do not necessarily bear on internal political or systemic factors. Organizational theorists assert that it is perfectly rational for a complex organization to resist competing control over its traditional domains. Though human service organizations tend to have less control over their environments than do organizations in the private sector, their reactions to intrusions in those areas they do control are similar. Further, though it can also be expected that organizations will conform with the law, the law in these cases is by no means settled. In many ways, it was inevitable that one or more states would ultimately test the Constitutional and statutory underpinnings of institutional litigation in this field.
The analysis can be divided into two parts -- factors affecting consent and non-consent, and factors influencing progress in the implementation of court decrees. No one factor can be isolated as necessarily the most prominent and not all of the variables proved useful in explaining the reasons why Pennsylvania's reaction differed from that of Maine and Michigan. A summary of the relevance of the initial variables that formed the hypotheses follows:
Level of sophistication and development of the existing state mental retardation system -- This factor did not prove very helpful in explaining the distinction between Pennsylvania on the one hand and Maine and Michigan on the other. Though Maine's system at the time of the suit was not fully developed, certainly the Michigan system could be seen as relatively complex and sophisticated. The more interesting factor that emerged, which is somewhat related, is the extent of shared ideology among key staff in the mental retardation agency in Pennsylvania and their sense of efficacy in creating system change.
Extent of public pressure for reform -- Certainly in Michigan the pressure in the press and from the public weighed in favor of expedited negotiations. in Maine, the pressure was more diffuse and in Pennsylvania the pressure was more sporadic. This factor may be a partial explanation for consent but does not necessarily explain progress once the agreement is reached.
Explicit or implicit agenda of state officials -- This factor appears to be important both with regard to consent and progress in implementation--a fact that is born out in the comparison states and in the literature. To the extent that state officials see litigation as a means of furthering their programmatic agendas, the chances of consent and progress are heightened.
Nature of the relationship between state program officials -- This factor appears to be important in the forging of a consent decree. In the two comparison states, state lawyers were more governed by the program agenda of state agency officials than was the case in Pennsylvania.
Orientation of the state's political leadership -- This factor has a somewhat hazy relationship to the events under analysis. If orientation means political party, there appears to be no relationship between party identification, and the inclination to settle. In Pennsylvania, the case now spans two political administrations, and neither showed any inclination to consent. What is clear is that the aims of the Governor play a key role in the decision to consent.
Extent of previous litigation in the state -- Though it cannot be directly shown that the cumulative effect of multiple suits in a state will eventually turn state officials against consent decrees, anecdotal information clearly suggests that enthusiasm wanes and wariness increases after prolonged experience with complex consent decrees.
Judicial strategies employed by the federal judge in contested and uncontested cases -- This factor requires substantially more exploration in more cases before any real conclusions can be drawn. At least tentatively, it does appear that the judges in Maine and Michigan were more successful at cajoling the parties into consent--and into fairly regular progress--but it is not clear whether the other factors suggested outweigh the judicial influence in all three cases.
Nature of the decree and the monitoring mechanisms established -- This factor leads to a circular argument which is not terribly useful in explaining the differences among states. Since the nature of the decree and the compliance mechanism are directly related to whether or not there is consent, the analysis becomes a tautology.
Strategies employed by the plaintiffs -- This factor has potential utility for explaining the behavior of state defendants, but the limited amount of information in this analysis is not conclusive. If the defendants' perception of the lawyers themselves are taken into account, then this factor plus the strategies employed did tend to establish expectations among the defendants in Pennsylvania regarding the "implacability" of the plaintiffs' attorneys.
Level and distribution of state resources -- This factor is not particularly satisfactory in explaining the decision to consent among the three states -- at least at the time such decisions are made. Michigan's level of funding, if anything, was lower than what was available in Pennsylvania and certainly the economic future of that state was much more precarious. Level of funding may, however, bear on the degree of progress a state is able to make in implementing the decree. Further, the extent to which funding for the decree is obtained at the expense of other parts of the system may ultimately constrain compliance.
Though Pennsylvania is treated in this analysis as an exception to the trend of settlement in mental disabilities cases, the posture of the commonwealth may increasingly become the rule. The question is whether those settled cases, if they were brought now, would result in consent agreements. Of the cases brought recently, more are going to trial, and consent agreements are more aggressively negotiated by the defendants.
Even without a shift in the case law, many state officials are increasingly reluctant to submit control over aspects of the service system to federal court oversight. In part, this reluctance stems from direct experience with other consent decrees and in part it is the result of a growing consensus among such individuals that the price paid for consent is not worth the benefits that may be conferred on the system. One state official among the several states contacted for this analysis was asked whether he would support consent if he had to do it all over again and his answer was a reluctant "no."
Finally, the increasing resistance to federal court intervention is also strongly influenced by the gloomy financial picture emerging at the federal level and in several states. As long as resources were relatively flexible, there was enough "play" in the system to accommodate comprehensive consent agreements. As resources become short, meeting court requirements may be accomplished at the expense of expansion or improvement in other parts of the system. The uncertainty surrounding future cut-backs in federal funds also may mean that many state officials will be loathe to contemplate significant system reform projects.
Another related fiscal issue has to do with the Medicaid program. Those states that have certified a significant number of institutional beds for Title XIX reimbursement may resist court-mandated, deinstitutionalization unless they can be assured that the Title XIX funds will follow the clients into the community. In states where there is an aggressive ICF/MR program in the community, this shift may be accomplished with no substantial loss to the state treasury. However, in states where community programs are funded primarily with state dollars, deinstitutionalization will result in a direct loss of federal funding and a concommitant drain on scarce state funds. The rumored cap on Medicaid may even diminish the ability of those states with community ICF/MR facilities to expand the program given the reluctance of providers to invest funds in the face of an uncertain potential for reimbursement.
Growing fiscal concerns have also resulted in increased attention by state legislatures to the fiscal impact of litigation. Whereas in the past state legislatures were only somewhat involved in the development of litigating strategies, today more and more legislatures are demanding a role in implementation. Their potential resistance to funding complex decrees poses serious problems for implementation and forces the issue of federal court jurisdiction over legislative bodies.
Nature of the Issue. The first two Implementation Analyses concentrated on key actors in the litigation -- the Office of the Special Master in the first year, and the Department of Public Welfare in second year. The topic for Year 3 covered a range of issues -- both as they emerged within the context of the Pennhurst litigation in Pennsylvania and in other comparison states.
The Pennhurst litigation has focussed a spotlight both on the implementation of public law litigation in the field of metal retardation and. also on the stresses and strains afflicting the mental retardation system in general -- particularly in the face of funding cut-backs and increasing concerns regarding the allocation of scarce resources. The Historical Overviews highlighted several issues that bear further assessment and exploration. They included:
Method and Objectives. The Implementation Analysis for Year 3 had several major objectives:
To gather the information necessary for this analysis, HSRI first identified four comparison states -- Maine, Michigan, Minnesota and Massachusetts. The major characteristic of each state was the presence of litigation directed at some aspect of the mental retardation system. The first two states were included in order to provide continuity with the Implementation Analysis for Year 3. The second two states were selected in order to broaden the base of analysis and because the litigation in those states is longstanding.
Prior to site visits, each state was contacted and pertinent court related and program materials were requested. The names of key system actor were secured and interviews were scheduled. A specialized interview guide was prepared to ensure that all relevant topics were covered. Each site visit lasted approximately two days. Material on the Pennsylvania portion of the analysis was gathered during the five day site visit conducted prior to the preparation of Historical Overview VI.
(1) Legislative Backlash.
In Pennsylvania, one of the major changes in the political landscape in which the mental retardation system functions is intensified legislative scrutiny. Whereas in the past the legislature had, within reason, relied on the Department of Public Welfare to set the tone and direction for the mental retardation program, insistent complaints from parents and others stimulated the legislature to conduct its own investigation of the management of the system. Late in 1982, the Pennsylvania Senate passed a resolution establishing a five member investigatory committee to review the operations of the Office of Mental Retardation. The committee looked into allegations of mismanagement within the Office of Mental Retardation, and in the community system generally.
The final report of the committee is primarily focussed on community living arrangements in the state. Though the committee finds them to be the most "home like" of all facilities visited, the report concludes that there is a need for "additional planning, preparation, and safeguards," and that it is time to "take stock."
The major recommendation by the Senate Committee was the formation of a Senate Task Force to design needed changes in the Mental Health and Mental Retardation Act of 1966. In making its recommendation, the Committee notes that "the legal base upon which the State's MR system is built may no longer be adequate." In addition to problems in the delivery of services, the Committee also appears to have been strongly influenced by the Pennhurst litigation. In reviewing the actions of both the Third Circuit and the Pennsylvania Supreme Court, the Committee states:
. . . these decisions now interpret the Act to entitle all of the MR population to the above-stated treatment without regard for the availability of funds and services exists. Intervention of the courts has created additional legal and manpower costs; has limited the available choices of professionals, parents and MR clients; has made regional and statewide planning more difficult; and has encouraged a division among Pennsylvania's advocacy groups.
Though attempts to amend the 1966 Act have consistently been unsuccessful, it is possible that the combination of the litigation and the growing dissatisfaction among some parents provide sufficient momentum to those seeking to put the brakes on deinstitutionalization through revisions of the state statute.
In each of the four comparison states, legislative attitudes toward the mental retardation system generally and to related litigation were explored. In all four states, legislators were supportive of services for mentally retarded persons and did not appear to question continued development of community-based services. In Michigan, for example, legislators had appropriated $3 million in new funding to provide services to "underserved" persons in the community. This investment in the face of Michigan's dire financial condition underscores the legislature's continued commitment. In Maine, though there has been no significant increase in state funding for community services this year, legislators remain pleased with the progress being made toward expansion of community services.
In Minnesota, a state which has also been hard hit by the recession, legislators are concerned about how to make the most of shrinking resources, but these hard fiscal realities do not appear to have dampened their enthusiasm for community-based services. In Massachusetts, the legislature is clearly concerned about the conduct of the system, but its criticisms are directed at the state's management of the system rather than at the viability of community programs.
Though legislators in the four states do not seem to share the concerns about the community system expressed by their opposite numbers in Pennsylvania, they all share a certain restiveness about the continued presence of the federal court in the management of state mental retardation programs. In Minnesota, legislators complained that even after the recent stipulation in the Welsh suit that expands reforms to all of the state's institutions, the plaintiffs continue to bring the defendants before the court over various enforcement details. In Massachusetts, the Chairman of the House Ways and Means Committee has formed a special subcommittee on "Federal and Court Consent Decrees." The purpose of the committee is to assess the impact of the court's intervention and to explore the state department's management of the funds provided by the legislature to meet the requirements of the decree.
In Maine and Michigan, the level of hostility is not as prominent but individual legislators are still concerned with the court's continued presence. In Maine, legislators are perhaps more sanguine because the state has already been released from half of the provisions of the decree in the Wuori suit. In Michigan, there is no significant disagreement among legislators regarding the aims of the decree, though individual legislator are unhappy that they were not involved in the negotiations.
(2) Union Influence
In Pennsylvania, the American Federation of State, County and Municipal Employees (AFSCME) is a significant actor in the political environment of the mental retardation system. Actions of the union have taken many forms including the use of litigation to attempt to block institutional closures and institutional phase-downs; financial support for the Parent/Staff Association, a defendant intervenor in the Pennhurst suit; and legislative lobbying, including successful opposition to zoning legislation that would have opened up residential neighborhoods to small group living arrangements for mentally retarded persona.
The intensity of AFSCME's activities definitely increased once the deinstitutionalization character of Judge Broderick's decree became clear. The nature of the litigation in the four comparison states, however, is somewhat different than the Pennhurst case in Pennsylvania. In Massachusetts, for instance, the five class action suits are all directed at institutional improvement and have resulted in at least a doubling of staff to client ratios. In Maine, though the consent agreement required the movement of some residents of the state mental retardation center to the community, the increased staffing standards in the decree offset the need for any lay-offs of state personnel.
In Minnesota, the state AFSCME chapter considered joining the plaintiffs in the Welsh suit in order to press for institutional improvement. Even though the defendants have now signed a stipulation agreement that includes a reduction in institutional census. AFSCME spokespersons do not see any abnormal reductions in force at the institutions. The situation in Michigan comes the closest to the situation in Pennsylvania since the litigation has resulted in the planned closure of a state institution. AFSCME in that state did attempt to intervene in the suit, but the Judge rejected their petition. Since that time, union officials have brought in staff from their national headquarters to try and persuade legislators and others to stem the tide of deinstitutionalization. To date Michigan AFSCME has not been as effective as their counterparts in Pennsylvania though they have been successful at negotiating a 12.5% ceiling on institutional staff lay-offs.
(3) Role of Enforcement Mechanisms
The creation of the Office of the Special Master in Pennsylvania caused a great deal of consternation both because of the extent of its responsibilities and the amount of resources devoted to its operations. Since its inception, OSM has been viewed by the state defendants in particular as an intruder into traditional state prerogatives. In part, OSM's problematic relationships with the defendants had to do with its multiple mandates and the individuated nature of much of its compliance mission. It was also a very large target given its $900,000 budget at the height of its powers.
Again, the situation in the comparison states is very different. For one thing, the litigation in all of the states visited has been settled by consent agreement. As mentioned earlier, the presence of consent has a direct bearing on the nature of the compliance mechanism established by the court. As a result, the court-appointed officers in the four states have responsibilities that are much more removed from the day-to-day operations of the system and the resources at their disposal are much more limited than those allocated to the Office of the Special Master in Pennhurst.
This is not to say that there were no tensions between court officials and state defendants. In Maine, state defendants became upset with the attitude of the court monitor in the Wuori case because of what they asserted was his failure to acknowledge the positive accomplishments of the state in meeting the requirements of the decree. The monitor finally resigned in favor of another individual whose personal style is less confrontational. It should be noted, however, that many of those in the state feel that the initial court-appointed official had the right approach for that phase of the litigation, and that the approach of the recent monitor is consistent with the requirements of the later stages of the litigation.
In Minnesota, though the Welsh case has been active since 1972, it is only recently that the court appointed a monitor. By and large, relationships with the state defendants have been smooth though, as mentioned earlier, the patience of the legislature with the court is beginning to wear thin.
In the other two states, relationships between court compliance officers and state defendants appear to be fairly positive. In Michigan, the monitor has eschewed obvious demonstrations of authority in favor of an "illusion of power." In Massachusetts, most seem to accept the monitor's role and appear to direct most of their attention to the actions of the Judge. Some legislators in particular have been concerned with the Judge's involvement in the system -particularly his decision to subpoena the Chairman of the Senate Ways and Means Committee.
(4) Schism in Parents Groups
The Pennhurst litigation appears to have exacerbated if not created enions among the parents of mentally retarded persons in Pennsylvania. Because of the frank deinstitutionalizatl'on character of the remedy, proinstitution parents were forced to take sides and they ultimately formed a separate organization and became opposing parties in the case. Given the community orientation of the Office of Mental Retardation in Pennsylvania, this polarization may have occurred in any event, but perhaps not as quickly nor as intensely. In order to determine whether the apparent schism in Pennsylvania was repeated in other states -- as the result of litigation and/or state deinstitutionalization policies -- parents group representatives in the four comparison states were interviewed.
In assessing the experience in the other four states, it should be kept in mind that, with the exception of Michigan, litigation had resulted in substantial institutional improvement. In Massachusetts, the five remedies are almost entirely comprised of standards for institutional reform. Parents in that state are somewhat unified, although the father of one of the named plaintiffs remains an independent agent somewhat critical of the state parents group. Unlike the situation in Pennsylvania, it is the community parents in Massachusetts who feel some resentment toward the institutional parents because of the diversion of resources to support state center programs.
In Michigan, the Plymouth suit was originally brought by parents of Plymouth residents who were concerned about institutional conditions. The state ARC eventually joined the suit and more recently the defendants have signed a stipulation to close the facility. Though Plymouth parents felt somewhat left out of the negotiation process and were initially hesitant about the impact of closure, they admit that the viability of the facility is in serious doubt. When asked whether they had ever thought of aligning themselves with the institution's employees to stop closure. A parent spokesperson gave an unequivocal "no" -- especially in light of the abuses attributed to some personnel at Plymouth. Though there is no npen schism between the state association and this local group, there is very little communication or sense of solidarity of purpose.
In Maine, where the litigation has resulted in both institutional improvement and deinstitutionalization, parents interviewed seemed pleased with the results. When the consent was first signed, however, there was concern among some institutional parents regarding the movement of their relatives to the community. According to those interviewed, this resistance to placement was diminished in large part because of the intervention of the state commssioner who personally worked with parents to orient them to the nature of community programs. Though there is no vocal division among parents in Maine, there is also no state parent organization. Recent attempts to resuscitate the dying state ARC failed. The collapse of the ARC, however, appears to have less to do with philosophical differences and more to do with previous mismanagement.
In Minnesota, parents appear to have made a conscious effort to accommodate the sometimes divergent views of institutional and community parents in order to hold the organization together. The litigation in that state does not appear to have exacerbated relationships among parents in part because it has evolved slowly and now includes mandates regarding both institutional improvement and community services.
Interestingly, relationships among parents appeared somewhat more strained in those states -- Massachusetts and Michigan -- where the parents organizations had become plaintiffs in the litigation. Further, all parent group representatives reported a decline in vitality in their organizations ironically because of their past successes. Now that public education has been extended to all handicapped children, for instance, recruitment of the parents of young children has fallen off.
Bardach, E. The Implementation Game. Cambridge, MA: The MIT Press, 1977.
Gregory, D.D. Report of the Special Master to the United States District Court for the District of Maine: Community Standards, Appendix of the Court's Decree, April 1980, P. 1.
Horowitz, S.G. Report to the Court: March 12, 1979 - June 1, 1979. Report submitted by the District Court Monitor to the United States District Court, District of Massachusetts, p.4.
Pressman, J. & Wildavsky, A. Implementation (2nd edition). Berkeley, CA: University of California Press, 1979.
Wyatt V. Stickney, 344 F. Supp. 373 and 387 (M.D. Ala. 1972), aff'd sub. nom., Wyatt v. Aderholt, 503 F. 2d. 1305 (5th Cir. 1974).
|
Individual Progress Case Study: Growth in the Community Robert was delivered prematurely, at 6 1/2 months, in 1961; he weighed just over 3 pounds, and spent two months in an incubator. Due to retrolental fibroplasia at 6 months, Robert became blind. From that point on, Robert was developmentally delayed. He developed a seizure disorder and was not toilet trained until he was 5 years old. His parents enrolled him in a school for the blind, but he was asked to leave within a year, as he had begun to lose bowel control. For the next 3 years, Robert went to another school, where he learned to dress and undress, toilet himself, and speak in simple sentences. Robert's parents were going through a divorce, and his school was too far away, so his parents institutionalized him at Pennhurst. The early records indicate that Robert began to regress soon after admission. He lost his ability to speak, began having toileting accidents, and began to bite and slap himself and others when he was upset. Since Robert was under 21 in June 1979 (when Judge Broderick signed the "school-age children's order"), he was slated to be one of the early movers. Because of parental objections, Robert did not move until the summer of 1982, and even then his parents were less than thrilled. The changes in Robert in the 2 years since his placement have been remarkable. When the Case Studies Coordinator visited Robert in his group home most recently, he seemed very different. His clothes fit properly and were, in fact, quite stylish. His hair was well trimmed and neat, and he was smiling, something that had not been the case in the 12 visits with Robert while he was at Pennhurst. In addition, as staff pointed out, Robert had, no open wounds on his hands, which had been the prime target of self abuse in the past. In place of the open wounds were scars, a reminder of Robert's past behavior. There had been quite a change in Robert's home, as well. All over the house one could find soft sculpture on the walls to both stimulate and orient Robert in the house. He was also using a cane and, with it, was able to move about the house independent of staff. During the visit, Robert signed "bathroom" to the staff person and proceeded to the bathroom without help. When he returned, staff praised him and Robert, smiling, looked quite pleased with himself. Knowing he had achieved a major accomplishment, Robert approached the staff person and signed the words "please" and "cookie." |
For more than a century, states have maintained large) segregated, congregate care institutions for people with mental retardation. More recently, residential alternatives closer to home have been developed for such individuals. The Pennhurst Longitudinal Study investigated whether people were better off, in terms of their own individual behavioral development, after making the transition from an institution to a community residence.
The places where people went in the Pennhurst case are called Community Living Arrangements (CLAS). These are very small programs, usually housing only three but almost never more than six residents. CLAs are almost always in regular residential housing stock, and are staffed continuously when the people who live there are present. All residents leave every weekday to go to some variety of day program or work or school. Staff coverage is provided either according to the live-in plus part-time-help model or the shift model, with the preponderance of programs using the shift model. Service providers are private entities, about 90% are non-profit, and they range from very small (one CLA site) to quite large (40 CLA sites).
Beyond this basic CLA model, which has been in place in Pennsylvania since the early 1970s, certain additional programmatic and procedural elements were required by the Federal court for Pennhurst class members. The court mandated case managers with caseloads not to exceed 30, ordered that Individual Habilitation Plans (IHPs) be written in a collaborative way involving all concerned professionals and nonprofessionals, and also that those plans be reviewed and approved by a special unit before implementation, and finally that a special unit be designated to monitor the well being of the people and the services rendered to them.
Similar community service settings have been proliferating rapidly across the country (Janicki, Mayeda, & Epple, 1983). But to the extent that a given state's community service configuration differs from the model described above, the power to generalize from our Pennhurst Study findings to that state is decreased. As an extreme example, our research would probably have little to say about a state in which the community service system that is composed of 15-bed, specially constructed or renovated facilities located in mixed zoning areas.
The deinstitutionalization of Pennhurst Center should be seen in the national context of declining institutional populations and increasing community residential facility populations. Exhibit 4-1 on the next page shows the changes in public institution populations from 1960 to the present. Clearly, there has been a strong trend away from institutional care, but the figure also reveals that as of this writing about 100,000 people still live in public institutions. Whether it would be possible to serve those people in a "better" way, at the same or lower public cost, is an essential question addressed by the Pennhurst Study.
In the sense of Campbell (1967) in his classic article "Reforms as Experiments," the Pennhurst Study was an evaluation of a social experiment. The reform (experiment) in this case was conducted by a Federal court. On March 17, 1978, Judge Raymond J. Broderick of the Federal court for the Eastern District of Pennsylvania ordered that all the people living at Pennhurst (among others) move to alternative CLAs. Evidence and expert testimony had convinced the judge that people would be better off out of Pennhurst Center but no one was really certain. The issue of deinstitutionalization was controversial and provoked broad public concern.
| EXHIBIT 4-1. Population of Public Institutions |
 |
Prior research had established firmly that deinstitutionalization of people with mental illness had in many states been a failure (Bassuk & Gerson, 1978). In the field of mental illness, the decline in institutional populations began in 1955 (long before it began in mental retardation). People in many instances were "released" from mental institutions with no place to go, no backups, no supports, and nothing to do during the day. The bulk of public opinion about deinstitutionalization was formed by that flawed policy. The politicians who voice concern about the homeless, the street people, the vent people are, in the vast majority of cases, talking about people who were released from mental health, not mental retardation, institutions.
Institutions for people with mental health problems were generally not very pleasant places to live during the 1950s (Goffman, 1961). Public and professional outrage over institutional conditions surely lent momentum to the trend toward institutional discharges. Perhaps an even more powerful catalyst was the development of powerful new medications that could ameliorate the effects of many forms of mental illness. The first of these medications was approved for general use by the Food & Drug Administration in, not coincidentally, 1955. It appears that many people were released from facilities with a supply of medications and little else.
In the field of mental retardation, in contrast, the situation is by no means parallel. When a person with serious intellectual impairment is considered for release, it is clear to everyone that the individual will still need round the clock supervision. There are no chemical or other substitutes for creation of a place to live with staff and therapeutic, activities. Thus the Pennhurst Study was not revisiting an old question. The question was, in Pennsylvania, under this court order at this time, with these Pennhurst residents who had mental retardation, would community placement (deinstitutionalization) be beneficial?
In the first section of our quantitative research on this question, we were concerned with behavioral growth and development. This area merited primary attention because several ideological trends and practical program models were converging toward the "reduction of dependency" as the central goal of services. This concept was based, in part, on a growing realization among professionals in the field that all people could grow and learn (Gold, 1973). New behavioral technologies were being used to impart skills such as independent toileting to people who professionals had thought were incapable of learning such skills.
In the Federal standards for reimbursement under Title XIX, Intermediate Care Facilities for the Mentally Retarded, the phrase is was "active treatment." Active treatment implies interventions that are designed to be far more than custodial. The requirement is meant to facilitate gradual but continual increases in independent functioning. The Accreditation Council on Mental Retardation and Developmental Disabilities, formerly a part of the Joint Commission on Accreditation of Hospitals, has supported that notion in conjunction with an emphasis on the developmental model.
The single most influential principle in the field of mental retardation in the past decade has been the principle of normalization. In his original formulation, Wolfensberger (1972) defined normalization as:
"Utilization of means which are as culturally normative as possible in order to establish and/or maintain behaviors and characteristics which are as culturally normative as possible" (page 28).
The definition of normalization has evolved since 1972 but the original formulation held sway through most of the 1970s. The principle strongly implied, through the phrase "in order to," that one of the two central purposes of services was to increase peoples' behavioral repertoires to encompass skills and patterns displayed by average citizens. (the other purpose was to do this in ways that did not degrade people or emphasize their differences from average citizens.) Subsequent treatments of normalization (Wolfensberger & Glenn, 1975) also stressed a "developmental growth orientation" and the "intensity of relevant programming" to foster behavioral development.
The first part of the 1972 definition clearly meant that the principle of normalization was incompatible with segregated, large-scale institutional care because such settings could never be considered "as cultural normative as possible." If people moved from an extremely deviant and non-normative segregated setting to a more normative and valued living arrangement, then normalization predicted that favorable changes in behavior would follow. In specific terms, then, the principle predicted that people moving from Pennhurst to CLAs would display more normative (higher adaptive and lower maladaptive) behaviors.
Thus several standards and philosophies of service highlighted the importance of behavioral outcomes. Because the technology to measure the adaptive behavior of individuals was already well developed in 1978, the question of behavioral benefits of deinstitutionalization became the central focus in the Pennhurst Study.
In 1978 there was an extreme paucity of reported research concerning the behavioral benefits of deinstitutionalization. We knew of only a handful: Aanes & Moen (1976), Brown (1978), Fiorelli & Thurman (published in 1979, but conducted in 1977-1978 at Temple University), Isett & Spreat (1978), and Schroeder & Henes (1978). Each one reported behavioral improvements after community placement, but each study was small, short term, and limited in generalizability. In this area, then, the results of the Pennhurst Study became the most extensive body of knowledge in the country.
More recently, comprehensive reviews of the policy of deinstitutionalization (Willer & Intagliata, 1984) and of research about outcomes (Craig & McCarver, 1984) have been published. The Pennhurst studies figured prominently in both. Because of the availability of these recent reviews, we will not present an extensive literature review here.
In this chapter, there are two studies. The first is a replication of our earlier study (Conroy, Efthimiou, & Lemanowicz, 1982) using the matched comparison design, which tests whether similar people, some who leave Pennhurst and some who stay, display different amounts of behavioral growth over time. But that study concerned 70 of the first people to leave Pennhurst; here, we will report on 191. The second is the longitudinal design. This design, the best scientific approach available to us, measures a person's growth while living at Pennhurst, then measures that same person's growth upon community placement and while living in the community. This enables us to test whether the same person displays more rapid behavioral growth in one setting than the other.
Both of these designs are quasi-experimental; neither is as powerful scientifically as a true experiment. In a true experiment, as noted by Campbell (1967), the reformer (in this case the judge) would have ordered that some number of people, say 100, be chosen by lottery to be deinstitutionalized first. This "random assignment" would enable scientists to generalize what was learned about these first 100, and predict confidently that the remaining 1054 people would have similar outcomes. Although this was not done (and may never be), the combination of the two strong quasi-experimental designs from the Pennhurst Study comes very close to the level of confidence a true experiment would provide.
Because there are two major studies to describe, but both used the same instruments and drew from the same population of subjects, we will begin with a description of general Methods that were applicable to both studies. Then the specific methods and results of each study will be presented, followed by a general discussion of both sets of results.
Subjects: General
The people of primary interest in all aspects of the Pennhurst Study were the 1154 people who lived at Pennhurst Center on the date of Judge Broderick's original Order, which was March 17, 1978. Their ages ranged from nine to 82 years with an average of 39, and they had lived at Pennhurst for an average of 24 years. Sixty-four per cent of the people were male. Thirty-three per cent had some history of seizures, 13% had visual impairments, 4% had hearing impairments, and 18% were unable to walk. Medical problems of a severe, life-threatening nature were reported for only eight individuals, or under 1%.
In terms of level of functioning, 54% were labeled profoundly retarded, 31% severely, 11% moderately, and 4% mildly retarded. For 9%, I.Q. was reported as unmeasurable; for the others, the range was from 3 to 87, with an average of 23. Just over 50% were completely or nearly nonverbal, 47% were less than fully toilet trained, and 40% were reported to threaten or do physical violence toward others. On the Behavior Development Survey, the adaptive behavior scores ranged from 0 to 120, with an average of 51 points; maladaptive behavior scores ranged from 3 to 22, with an average of 17 points.
Instruments: General
The Behavior Development Survey (BDS) contained our measures of individual functioning. Changes over time provided a measure of developmental growth. The behavioral items on the survey were taken from the American Association on Mental Deficiency's Adaptive Behavior Scale (ABS), by the UCLA Neuropsychiatric Research Group at Lanterman State Hospital, on the basis of mathematical criteria and reliability. The resulting shortened research version of the scale contained 32 items on adaptive behavior and 11 items on maladaptive behavior. According to Arndt (1981), the best way to treat these data is as two simple sum scores, one reflecting adaptive behavior and the other maladaptive behavior.
The adaptive behavior sum score has been found to be highly reliable (Conroy, 1980), with test-retest reliability of .96, and interrater reliability of .94. For the maladaptive behavior section, although test-retest reliability is good at about .90, interrater reliability is barely adequate at about .65 to .70 (Isett & Spreat, 1979; Conroy, Efthimiou, & Lemanowicz, 1981). The relatively "noisy" measure of maladaptive behavior implies that it is more difficult to detect changes; they must be quite large to be detected.
For the present study, we extended the instrument by adding items covering individual characteristics, family relationships, friendships, medical status, the individual habilitation plan, program goals, and type and amount of services delivered. The full modified BDS was designed to be a comprehensive tool for monitoring the status, needs, services, and outcomes of individuals in the mental retardation service system. The BDS was designed to be collected by interviewing the direct care and other Personnel who knew the individual best, combined with examination of records where necessary. Each BDS required about 40 minutes with the respondent(s).
Although the behavioral items on the BDS were not changed, the other sections were revised continually during the five years of the study. The 1984 version of the BDS is presented in Appendix 4-1.
Procedures: General
In September 1978 a BDS was completed for every person at Pennhurst by teams of institutional staff members most familiar with the individuals. Each team usually included a direct-care worker, a psychologist, and a nurse. Written instructions were provided, and the Temple University Developmental Disabilities Center's Evaluation & Research team was available on site to answer questions about the form. A total of 1113 forms were completed (41 people had already left Pennhurst). This supplied the baseline data for the entire five year study.
In subsequent years, BDSs were collected by project field staff by direct interviews with interdisciplinary groups of direct care and other staff who knew the individuals best. Records were used to verify the data in the sections on written plan, demographics, health, and services. Exhibit 4-2 below displays the record of BDS data collection for the whole study.
| EXHIBIT 4-2. BDS Data Collection | ||
|---|---|---|
| Year | At Pennhurst | In CLAs |
| 1978 | 1113 | 0 |
| 1980 | 713 | 70 |
| 1982 | 0 | 223 |
| 1983 | 618 | 408 |
| 1984 | 0 | 474 |
Data were not collected at Pennhurst in every year because the focus of interest was the effects of community placement. Originally, the study design did not call for any Pennhurst data after 1978. The Temple team added this facet after the study began because it made possible the matched comparison designs.
Design
The matched comparison design was implemented by identifying all the people in CLAs for whom baseline BDS data were available; for each one we then tried to find a person who was still at Pennhurst, and who was the same sex and was also very similar in initial adaptive behavior, maladaptive behavior, and age. For both groups ("movers" and "stayers") we compared 1983 BDS data to the 1978 baseline data, investigating whether one group had changed more than the other.
The matched comparison design is quasi-experimental. Specifically, it is a prepost nonequivalent control group design with subjects matched on pretest scores and several other variables. The weaknesses of the design are that no matching can be perfect,-and that no adequate matches may be available for some people, so that we can wind up with biased samples.
Our objectives were to compare the behavioral changes of matched samples of institutionalized and deinstitutionalized people and to identify, in a preliminary way, specific variables that might be associated with individual growth.
Subjects
Prerelocation (1978) and postrelocation (1983) data were available for 340 people who were placed in CLAs under federal court order. Each "mover" was matched as closely as possible with a person who was still at the institution in 1983, and there were 618 such "stayers." Individuals were matched on the bases of (1) gender, (2) chronological age +5 years, (c) prerelocation (1978) Adaptive Behavior total score +5 points, and (d) prerelocation Maladaptive Behavior Total Score +3 points. The matching process located excellent matches for 191 of the 340 movers. Perfect gender matches were found in all cases (134 males, 57 females); means for the two groups on the other matching variables are shown in Exhibit 4-3. No significant differences were found between the movers and stayers on the matching variables (using simple t-tests).
| EXHIBIT 4-3. Adequacy of Matching | ||
|---|---|---|
| Variables | Movers | Stayers |
| Matching variables | ||
| 1978 Adaptive Behavior | 54.8 | 55.0 |
| 1983 Maladaptive Behavior | 18.3 | 18.1 |
| Age (in 1978) | 38.1 | 37.7 |
| Other variables* | ||
| Vision | 3.6 | 3.5 |
| Hearing | 3.9 | 3.8 |
| Ambulation | 3.4 | 3.4 |
| Years at Pennhurst (in 1978) | 24.3 | 23.8 |
| * Vision, hearing, and ambulation are on scales from 1 (extreme impairment) to 4 (no impairment). | ||
Both the movers and the stayers displayed an average 1978 adaptive behavior score of 55 points (the scale ranges from 0 to 128), which was very close to the overall population's average of 51. In maladaptive behavior, both groups scored about 18 points, again close to the population average of 17 points. The average age for both groups (in 1978) was 38 years, similar to the population average of 39 years.
Group differences were examined on some other variables as well. Secondary conditions, including vision, hearing, and ambulation were compared using simple t-tests; none were significantly different. These results seemed to indicate a lack of "creaming" (i.e., selecting people to leave the institution specifically because of less serious secondary disabilities) in selection of the movers. No difference was found between movers and stayers in the number of years they had lived at the institution. Both groups averaged 24 years, the same as the population.
Thus, although not chosen by lottery, the people in this matched comparison study reflected the characteristics of the population quite well.
Group Comparisons of Behavioral Change
Several methods of statistical analysis were used in the prior matched comparison of developmental growth (Conroy, et al., 1982); all led to the same conclusion as the simple t-test. Here, we present only the simple t-test because it is the most straightforward. As Exhibit 4-4 shows, the 191 people who were placed in community settings were functioning at a higher level of adaptive behavior in 1983 than were their matched peers who had remained at Pennhurst.
| EXHIBIT 4-4. Behavior Changes Among Movers and Stayers | |||
|---|---|---|---|
| 1978 | 1983 | Change | |
| Adaptive Behavior | |||
| Movers | 54.8 | 66.3 | +11.5 |
| Stayers | 55.0 | 55.7 | +0.7 |
| Maladaptive Behavior | |||
| Movers | 18.3 | 18.0 | -0.3 |
| Stayers | 18.1 | 18.2 | +0.1 |
| * Higher scores are favorable for both. | |||
A t-test on the 1983 adaptive behavior total scores of the two groups was significant (t = 3.94, (380), p = .001). The results in maladaptive behavior showed only very slight changes in both groups, and the t-test revealed no significant difference between the movers and stayers in 1983.
This analysis indicated that the deinstitutionalized group had improved in adaptive behavior by more than 11 points over a five year period, while the group which remained at Pennhurst gained less than one point. Neither group changed significantly in maladaptive behavior.
Group Comparison of Service Provided
Service data were collected in 1983 on the BDS for both the movers and the stayers. The amount of developmentally oriented service rendered in the prior month at t he living area was obtained. These services included training (e.g., academic, mobility, social, interaction, community living, etc.), skills development (dressing, eating, hygiene), therapy (physical, occupational, speech, etc.), behavior modification (to reduce maladaptive behavior), and supervised recreation. We also measured time spent at the day program (vocational, educational, etc.). Exhibit 4-5 presents average hours of service per person per month for the two groups.
| EXHIBIT 4-5. Hours of Service Per Month Reported in 1983 | ||
|---|---|---|
| Movers | Stayers | |
| Services at Living Area | 104.5 | 156.0 |
| Day program | 120.7 | 33.1 |
| TOTAL | 225.2 | 189.1 |
As the table shows, people living at Pennhurst received more service on their living areas each month than their counterparts in the community. However, the movers spent more time at the day program and received more total service. On the average, the movers received 8.0 hours of service per day and the stayers received 6.8 hours of service per day.
Correlates of Adaptive Behavior Gains Among Movers
Because a substantial change in adaptive behavior was found only for the movers, we examined factors correlated with growth among the movers. Change in adaptive behavior was compared by Pearson correlations with 23 variables, including personal characteristics (sex, age, etc.), functioning level, secondary conditions (vision, hearing, ambulation, seizures), medical information, family contact, and service data. The results appear in Exhibit 4-6.
Three variables displayed significant correlations with adaptive behavior gains upon deinstitutionalization. They were year of admission to community living arrangement, ambulation, and beginning adaptive behavior total score.
These results suggested that (1) people who had been in CLAs the longest showed the most overall growth, (2) people who could not walk displayed more growth than those who could, and (3) people who started out with lower levels of adaptive behavior showed larger gains than did people who initially had more skills.
| EXHIBIT 4-6. Correlates of Adaptive Behavior Gains Among Movers | ||
|---|---|---|
| r | p | |
| Year of admission to community living arrangement | -0.25 | 0.001 |
| Ambulation (1978)* | -0.23 | 0.001 |
| Adaptive behavior total score (1978)** | -0.21 | 0.001 |
| Number of goals in written plan | -0.11 | 0.057 |
| Weeks since case manager last visited | 0.11 | 0.058 |
| Level of retardation (1 = not retarded, 5 = profound) | -0.11 | 0.072 |
| IQ | 0.11 | 0.190 |
| Change of address in past year*** | 0.10 | 0.083 |
| Medical needs* | 0.09 | 0.098 |
| Sex (0 = female, 1 = male) | 0.08 | 0.137 |
| Family contact (1 = weekly, 5 = never) | -0.08 | 0.143 |
| Vision (1978)* | -0.07 | 0.169 |
| Maladaptive behavior total score (1978)** | 0.06 | 0.187 |
| Number of residents at the site | -0.06 | 0.225 |
| Year of admission to Pennhurst | 0.05 | 0.228 |
| Amount of behavior modification used | -0.05 | 0.231 |
| Months since last medical exam | 0.05 | 0.247 |
| Seizure frequency | -0.05 | 0.247 |
| Amount of developmental service received | -0.04 | 0.280 |
| Hearing (1978)* | 0.03 | 0.358 |
| Year of birth | 0.02 | 0.366 |
| *
Scale of 1 (extreme impairment) to 4 (no impairment). ** Higher scores are favorable. *** 0 = not, 1 = yes. |
||
Subjects
In 1984, we visited 474 people who left Pennhurst Center under court order at their new homes in CLAs. The information we had collected about these people since 1978 formed the data set for the longitudinal analyses of growth and development. Again, for convenience we will adopt the "movers" and "stayers" terminology.
In mid 1984 there were about 450 Stayers still living at Pennhurst. Ninety-two of the remaining 138 people (the original 1154 minus 474 minus 450) had died, 77 of them while still at Pennhurst and 15 in CLAs; 32 had gone to other congregate care facilities, and the other 14 had returned to the natural family at family choice.
The movers were living in small CLAs. Most, 63%, lived in three person CLAs. Another 1% were living in a CLA by themselves, 19% had just one housemate, 11% were in CLAs with a total of four to six people, and 6% were in settings with a total of seven to 11 people.
Because many past deinstitutionalization activities have resulted in creaming" or selection of only the highest functioning people for placement, an immediate question was how the movers compared to the original population of 1154 people. In prior years of the Pennhurst Study, we had found only trivial differences between Movers and Stayers; people being placed were just about the same as those still awaiting placement in the areas of adaptive and maladaptive behavior, age, level of retardation, and secondary handicaps. As our data set grew in numbers, some of the differences reached statistical significance, but they were still not large in magnitude, as shown in Exhibit 4-7.
| EXHIBIT 4-7. Comparison of Movers’ Characteristics to Those of the Original Population of 1154 People | ||
|---|---|---|
| Movers | Population | |
| 1978 Adaptive Behavior | 59 | 51* |
| 1978 Maladaptive Behavior | 18 | 17* |
| Age in 1978 | 37 | 39* |
| Years at Pennhurst | 21 | 24* |
| Vision (1 to 4 scale) | 3.7 | 3.5* |
| Hearing (1 to 4 scale) | 3.8 | 3.8 |
| Ambulation (1 to 4 scale) | 3.4 | 3.3 |
| * t-test significance, p < 0.01. | ||
The statistically significant differences meant that the people placed in CLAs by 1984 were slightly higher in adaptive behavior, had slightly fewer maladaptive behaviors, were about two years younger and had spent three fewer years at Pennhurst, and were slightly less likely to have a visual impairment, than the average person who lived at Pennhurst in 1978. These differences suggest that, strictly speaking, our findings for the people placed so far will not necessarily hold true for those to be placed in the future. However, the differences are small, and we think it is very likely that future placements will have outcomes very similar to those we have observed.
Design
The longitudinal approach is, in this case, really a family of analyses of the form called "interrupted time series" by Campbell (1967). We observed the behavior of people repeatedly, both before and after they moved to CLAs. The move to the CLA is the "interruption" in the time series. If significant changes are observed right at the time of the "interruption," then those changes are unlikely to be coincidental.
The strength of the design is enhanced by using all possible time series configurations available in the data set. We have done so. We collected BDS data (as previously displayed in Exhibit 4-1) in 1978, 1980, 1982, 1983, and 1984. For some individuals, we collected a BDS in all five years; these were people who were still at Pennhurst in September 1980, and went to CLAs in late 1980 or by the middle of 1981, so that we saw them in CLAs in 1982 (we only collected data for people after they had been out for six months or more). For other people, who moved in 1983, the 1982 CLA data point did not exist; for them, there were just four observations. When all of the permutations are examined simultaneously, we can see whether the results are consistent across all the ways of analyzing behavior change.
Adaptive Behavior
The overall results of the family of longitudinal analyses for adaptive behavior are presented in Exhibit 4-8 in numeric form. We will summarize the findings and then provide more detail on two of the clearest and most meaningful analyses. The overall questions are, again, did people change behaviorally upon deinstitutionalization, and did that pattern of change continue after placement?
| EXHIBIT 4-8. Longitudinal Results: Adaptive Behavior | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Design | Year | (N) | ||||||||
| 1978 (PC) | 1980 (PC) | 1982 (CLA) | 1983 (CLA) | 1984 (CLA) | ||||||
| 1. | 51.4 | 51.6 | ***** | 59.9 | ***** | 65.2 | 65.1 | (92) | ||
| 2. | 53.0 | 53.6 | ****************************** | 63.8 | 65.1 | (176) | ||||
| 3. | 60.8 | ****************************** | 69.1 | ***** | 73.8 | 74.4 | (163) | |||
| 4. | 52.4 | 53.0 | **************************************************** | 64.8 | (200) | |||||
| 5. | 60.5 | ******************************************************** | 71.3 | 72.2 | (326) | |||||
| 6. | 59.3 | ****************************************************************************** | 70.7 | (383) | ||||||
| * Entries connected by asterisks were significantly different by paired t-tests at p < 0.001. | ||||||||||
Exhibit 4-8 indicates the five years of data collection across the top. The subheading "PC" means that the data in those columns were collected at Pennhurst Center, and the "CLA" subheading means the data were from CLAs. Overall, the table shows that significant gains never occurred within Pennhurst, always occurred upon CLA placement, and sometimes gains continued even after placement. Notably, none of the designs revealed significant growth among people in CLAs between 1983 and 1984.
In design 1, which included all five data points, the right hand column shows that N = 92, which means that there were 92 people who were at Pennhurst in 1978 and 1980, and then moved to a CLA in time for us to visit them in 1982 and 1983 and 1984. The asterisks show where significant increases in adaptive behavior occurred: for this design, significant increases were observed from 1980 to 1982 (initial CLA placement) and from 1982 to 1983 (advances continued after placement). The gains appeared to level off after 1983. Exhibit 4-9 presents these findings visually. Exhibit 4-9 also shows what is evident from all the designs in Exhibit 4-8: there was no statistically significant growth in this measure of adaptive behavior among these individuals while they were living at Pennhurst. The second longitudinal design included everyone for whom we had baseline 1978 data, who were still at Pennhurst in 1980, and who went to a CLA between 1980 and late 1982. There were 176 people in this category, and, as can be seen in Exhibit 4-10, they also made large gains in adaptive behavior upon community placement. The gain from 1983 to 1984, within the CLAs, was not statistically significant in this analysis.
Design 3 revealed the large initial gains, and also showed a continuation of growth within the community settings. Designs 4,5, and 6 further confirmed the lack of growth within Pennhurst and the sudden gains upon placement.
In sum, the adaptive behavior data showed clear and large gains among people who went to CLAs. After placement they were doing more things independently or with less help. Because this could have been the result of the change in environmental demands between the institution and the CLAs, it was important to test for continued growth after placement. In two of the longitudinal analyses (designs 1 and 3 in Exhibit 4-8), such continued growth was observed. In the first of those analyses, the post-placement growth rate was just as rapid as the large gains upon placement. These adaptive behavior findings, especially among people who had been institutionalized an average of 24 years, seemed to us to tell a very positive story about human potential that had laid dormant among these people with mental retardation.
| EXHIBIT 4-9. Adaptive Behavior Growth: 5 Observations, 92 People |
 |
| EXHIBIT 4-10. Adaptive Behavior Growth: 4 Observations, 176 People |
 |
Maladaptive Behavior
The results of the longitudinal analyses of changes in our measure of maladaptive behavior were that there was no significant change when people went to CLAs. The data are presented in Exhibit 4-11.
| EXHIBIT 4-11. Longitudinal Results: Maladaptive Behavior | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Design | Year | N | ||||||||
| 1978 (PC) | 1980 (PC) | 1982 (CLA) | 1983 (CLA) | 1984 (CLA) | ||||||
| 1. | 17.2 | 17.0 | 17.2 | 17.7 | 17.8 | 93 | ||||
| 2. | 17.2 | 17.3 | 17.6 | 17.7 | 179 | |||||
| 3. | 17.9 | 18.1 | ***** | 18.6 | 18.6 | 165 | ||||
| 4. | 17.3 | 17.3 | 17.6 | 203 | ||||||
| 5. | 18.1 | 18.3 | 18.5 | 326 | ||||||
| 6. | 18.0 | 18.2 | 386 | |||||||
| * Entries connected by asterisks were significantly different by paired t-tests at p < 0.05. | ||||||||||
Exhibit 4-11 represents over 5 years of trying to detect any change on this scale, and the only one noted was statistically weak and was within-CLA rather than a change upon placement. It is possible that there was no improvement in the maladaptive behavior area among these people over the years. But it is equally possible that our scale was not sensitive or reliable enough to detect genuine changes.
As noted previously, the maladaptive behavior scale suffers from a lack of interrater reliability. Different respondents do not agree very well on what constitutes, for example, "Rebelliousness." This makes it difficult to attain statistical significance; the "signal" (behavioral change) must be very "loud" (large in magnitude) to be heard over the "noise" (random error of measurement). Indeed, it is at least suggestive that all of the rows in Exhibit 4-11 show increased scores after CLA placement, and thereafter maintainance or further increases; even though the trends do not reach statistical significance, we suspect that changes may be taking place.
In summary, however, we are not statistically scientifically able to report any significant benefits of deinstitutionalization in the area of reduction of maladaptive behaviors.
Longitudinal Changes in Service Delivery Patterns
The services section of the BDS was developed only after 1978, so there were no baseline data on services rendered to the population. In 1980, at Pennhurst, we did collect services information, and also in the community in subsequent years. This enabled longitudinal analysis of changes in the amount and pattern of services rendered to people. This time, we were asking the question "Is this person receiving more or less or different services in the community than s/he formerly received at Pennhurst?"
This is different from the matched comparison analysis, which asked whether two groups of similar people were receiving different services in 1983. In the longitudinal approach, we ask whether a person in the community in 1984 is receiving more or less or different services than that same person previously received at Pennhurst in 1980.
The results were much like those of the matched comparison. The summary figures are given in Exhibit 4-12.
| EXHIBIT 4-12. Hours of
Service Per Month Reported at Pennhurst in 1980 and in CLAs in
1984 (N=207) |
||
|---|---|---|
| 1980 (PC) | 1984 (CLA) | |
| Services at Living Area | 139 | 95 |
| Day program | 48 | 119 |
| TOTAL | 187 | 214 |
The decrease in hours of service per month delivered via the residential program was significant (t=5.17, (206), p<.001), meaning that the community service system delivered fewer hours of developmentally oriented programming, at the place where the person slept, than did the institution.
The community system delivered more than twice the amount of day programming, away from the place where the person slept, than the institution (t=19.6, (205), p<.001). When the two forms of service were combined into a total index, the 1984 community service system was delivering a larger quantity of service to these people than they had previously received at Pennhurst in 1980 (t=4.15, (205), p<.001).
As an exploration of an urgent contemporary issue in service delivery, we tested whether the 207 people in our data set who had been at Pennhurst in 1980 and were in CLAs in 1984 had shown any change in the number of medications administered to them on a daily basis, other than topical ointments and vitamins. At Pennhurst in 1980, these people had received an average of 2.1 medications each day; in 1984 in CLAs, they received an average of 1.7. The decrease was significant (t=3.22, (206), p<.001).
The overall results of five years of investigation into the behavioral consequences of deinstitutionalization are clear: in terms of adaptive behavior, the average person who left Pennhurst is better off. The average person is now about 11 points higher on our 128 point scale of adaptive behavior than s/he was while at Pennhurst. Matched people still living at Pennhurst did not show significant improvements. Moreover, the dramatic and sudden increases in adaptive behavior after CLA placement did not stop and level off; for at least a year after placement, the average person continued to display significant developmental growth.
The evidence suggests, however, that gains begin to level off at some point, usually a year or more after placement. It seems to us that the lack of significant growth from 1983 to 1984 demands attention and continued study. We will continue this investigation with support from the Commonwealth of Pennsylvania.
We should reiterate here however, that during the course of the study we did detect favorable behavior changes among the people living at Pennhurst. When all the people at Pennhurst are included in the analysis, we do attain statistical significance, as reported by Lemanowicz, Conroy, & Feinstein (1984). These gains amounted to just over 1 point in adaptive behavior and under 1 point in maladaptive behavior. This finding is mentioned here because it suggests that, unlike the situation at Pennhurst at the time of the trial in 1977, people have not been regressing while residing at the institution. At the trial, evidence indicated that the average person at Pennhurst had lost skills during his/her time there. In more recent years, then, that situation has changed. Any visitor can tell in a brief tour that Pennhurst has improved over the years, and it may be that our findings of growth are quantitative reflections of that fact.
Nevertheless, the results of the two designs presented here do establish the quantitative superiority of CLA settings in fostering adaptive behavior expression and growth. People who have gone to CLAs have gained literally 10 times as much as the people who still await placement.
The limitations of the two designs should be kept in mind, and, even more important, our caution about generalization of these results to other areas or states is very important. To the extent that a community service system is similar to the Pennsylvania model, such generalization is warranted with moderate caution. But for systems unlike the one implemented for the Pennhurst class members, it would be extremely hazardous to assume that our findings will apply.
In addition to the elementary finding that people are better off in terms of behavior, we also noted that the pattern and amount of developmentally oriented services rendered had changed. The patterns were that the institution delivered more service at the living area, while the community system delivered more service at the day program, and more service overall (6.8 versus 8.0 hours per day for Pennhurst and CLAs respectively). Thus we conclude that the people who have left Pennhurst are also better off in terms of the amount of developmentally-oriented service rendered to them. We hope that further evaluative studies will address the quality and consequences of various kinds of day program.
We also examined medication use, and found that the average person who had been placed was receiving fewer daily medications than previously at Pennhurst. This would usually be regarded as a favorable outcome, because there has been a great deal of concern in the field of mental retardation about overuse and misuse of many kinds of medications, particularly those used for behavior control, and particularly when they may have serious and permanent side effects such as tardive dyskinesia. (We should also note that, from 1980 to 1983, since the reorganization of medical services at Pennhurst under the auspices of a private corporation, the average person at Pennhurst is also receiving fewer medications.)
Other than the essential findings that people are better off in terms of behavior and services, we believe the most important outcome of our years of work in this area is that we have developed a technology for quantitative monitoring of the well being of people in dispersed, decentralized community service systems. Many observers have suggested, over the years, that the difficulties in monitoring community services would be enormous compared to the ease of monitoring all the people in one place at an institution. This has been offered as a major argument against deinstitutionalization.
In fact, quantitative monitoring is not a difficult process at all, nor does it need to be terribly costly. The Temple part of the team has embarked on a long term partnership with the Pennsylvania Office of Mental Retardation to continue monitoring the Pennhurst class members when the Federal funds for this study expire, and to expand that monitoring as rapidly as possible to other people in community settings. Although our once a year monitoring visits are no substitute for frequent case manager visits, active family participation, fiscal controls, and alert neighbors, the quantitative information about individual growth (or regression), individual services, family opinions, and environments yields a rich basis for individual corrective actions and for systematic analysis and planning.
Aanes, D., & Moen, M. (1976). Adaptive behavior changes of group home residents. Mental Retardation, 14, 36-40.
Arndt, S. (1981). A general measure of adaptive behavior. American Journal of Mental Deficiency, 85, 554-556.
Bassuk, E.L., & Gerson, S. (1978). Deinstitutionalization and mental health services. Scientific American, 238, 46-53.
Brown, M. (1978). Evaluation report: Post-institutional placement project. New York: New York University Institute of Rehabilitation Medicine.
Campbell, D.T. (1967). Reforms as experiments. American Psychologist, 24, 409-429.
Conroy, J. (1980). Reliability of the Behavior Development Survey (Technical Report 80-1-1). Philadelphia: Temple University Developmental Disabilities Center.
Conroy, J., Efthimiou, J., & Lemanowicz, J. (1981). Reliability of the Behavior Development Survey: Maladaptive behavior section (Pennhurst Study Brief Report No. 11). Philadelphia: Temple University Developmental Disabilities Center.
Conroy, J., Efthimiou, J., & Lemanowicz, J. (1982). A matched comparison of the developmental growth of institutionalized and deinstitutionalized mentally retarded clients. American Journal of Mental Deficiency, 86, 581-587.
Conroy, J., Feinstein, C., & Lemanowicz, J. (1981). Medical needs of clients: Perceptions of relatives and of staff (Pennhurst Study Brief Report No. 6). Philadelphia: Temple University Developmental Disabilities Center.
Craig, E.M., & McCarvar, R.B. (1984). Community placement and adjustment of deinstitutionalized clients: Issues and findings. In: N. Ellis & N. Bray (Eds.), International review of research in mental retardation, Volume 12. Orlando: Academic Press.
Fiorelli, J., & Thurman, S.K. (1979). Client behavior in more and less normalized residential settings. Education and Training of the Mentally Retarded, 14, 85-94.
Goffman, E. (1961). Asylums. New York: Doubleday-Anchor.
Gold, M.W. (1973). Research on the vocational habilitation of the retarded: The present, the future. In N.R. Ellis (Ed.), International review of research in mental retardation, Volume 6. New York: Academic Press.
Isett, R., & Spreat, S. (1979). Test-retest and interrater reliability of the AAMD Adaptive Behavior Scale. American Journal of Mental Deficiency, 84, 93-95.
Isett, R., & Spreat, S. (1978). Progress accountability and deinstitutionalization. Technical report 78-3-3. Philadephia: Temple University/Woodhaven Center.
Lemanowicz, J.A., Conroy, J.W., & Feinstein, C.S. (1984). A quantitative overview of Pennhurst Center: 1978 to 1983 (Pennhurst Study Brief Report #16). Philadelphia: Temple University Developmental Disabilities Center.
Schroeder, S.R., & Henes, C. (1978). Assessment of progress of institutionalized and deinstitutionalized retarded adults: A matched-control comparison. Mental Retardation, 16, 147-148.
Willer, B., & Intagliata, J. (1984). An overview of the social policy of deinstitutionalization. In: N. Ellis & N. Bray (Eds.), International review of research in mental retardation, Volume 12. Orlando: Academic Press.
Wolfensberger, W. (1972). The principle of normalization in human services. Toronto: National Institute on Mental Retardation.
Wolfensberger, W., & Glenn, L. (1975). Program analysis of services 3: A method for the quantitative evaluation of human services. Toronto: National Institute on Mental Retardation.
|
Consumer Interview Case Study: If I Were a Rich Man Steve moved to the community after having lived at Pennhurst for 27 years. While Steve reported having been very happy at Pennhurst, he is even happier in his new group home. When he was asked what he liked about the group home, he talked about how good the staff were to him and how they had put a bell in his bedroom so that if he needed help during the night he could just ring and the staff person would come (Steve is non-ambulatory). Steve also talked about how good he felt having been able to visit his Aunt Sue when she was in the hospital. When asked if he missed Pennhurst or any of the people, Stave said no. After thinking for a moment he said that he does miss a few of his friends, but not very much. When he was at a Speaking for Ourselves meeting be saw a few of his friends from Pennhurst who were now also living in group homes. Steve explained that Speaking for Ourselves is a place where you talk about a lot of things, like Pennhurst closing, and if you have a problem or something is bothering you they try to help you figure it out. When asked what he would wish for if he had one wish, Steve responded, "I wish for people to live with me who are nice and kind to me like these people." Bruce would like to stay in his group home. He moved there about 6 months ago, after having lived at Pennhurst for 28 years. He likes living in the community, because he gets to see his sister and her family and he works and earns money. (Bruce works on a pressing machine that steams and presses cardboard.) When asked how his group home differs from Pennhurst, Bruce said, "Pennhurst was alright, I grew up in that place. We have different hours of getting up and going to sleep here. We have Saturdays and Sundays to ourselves. This is more home; there is no big crowd, just a few people." When asked what he would wish for if he had one wish, Bruce replied, "I wish I was a millionaire." |
Among the many ways that the well-being of people with mental retardation may be assessed, one that stands out in importance and in difficulty is to ask the people themselves. In the Pennhurst Study, we were determined to address the feelings of the people themselves to the maximum extent possible.
It seemed particularly important to avoid the common error of assuming that only parents and professionals can make valid judgments about whether a person with mental retardation is better off. As Seltzer (1980) pointed out, "A critical, yet often ignored, aspect of retarded persons' community adjustment is their perceptions about their environments and the psychological sense of well being or discomfort derived from their living environments." However, Sigelman, et al. (1979) stated that, despite a trend toward allowing and encouraging people with mental retardation to speak for themselves, “...virtually nothing is known about the reliability and validity of information gained through survey research" (p. 1) with them.
It was clear at the outset that the methodological difficulties were considerable. For example, because we knew that nearly half of the people living at Pennhurst Center were nearly or completely nonverbal, we knew that the views of the people who were able to speak would not necessarily represent the views of those who were unable to speak.
In addition to this problem of representativeness, prior studies had suggested that some people with mental retardation had difficulty in expressing themselves in a consistent fashion. Despite these problems, it was decided that the effort to measure changes in individual satisfaction after movement from the institution to small community based living arrangements was demanded by the nature of the study.
The consumer interviews part of the study, then, addressed two objectives. First, it was designed to ascertain whether people who moved from Pennhurst into community living arrangements (CLAs) were pleased with the change, and whether there was any change in their self-expressed satisfaction and happiness. Second, because of questions about the reliability and validity of such interviews, the study was also intended to shed new light on the methodological problems inherent in soliciting direct consumer input.
Moreover, the study was, unexpectedly, able to investigate changes over time in the self-expressed satisfaction and happiness of people who remained at Pennhurst during the years in which the population of the facility dropped from 1154 to 450.
Consent Procedures
At the outset, it was determined that this phase of the study required extremely careful attention to the rights and privacy of the individuals themselves, because this was practically the only part of the effort that demanded direct contact. Certainly, if an individual said that s/he was not willing to be interviewed, then no interview would be done. But there were others who might have an important viewpoint regarding the advisability of the person's participation as well: program staff and families. We considered all of these parties. The only people we interviewed were those (a) who appeared, from prior data, to be capable of responding to verbal interview, (b) for whom staff judged there would be no significant risk to the person, (c) for whom written informed consent was obtained (either from families, or, in the case of people who had no family but were capable of giving their own informed consent, from the people themselves), and (d) who agreed on their own behalf when approached by our interviewers.
Our extreme caution in safeguarding rights and privacy in this part of the study grew at least partly from the knowledge that, in past years, people living in institutional settings have been part of studies that would never have been approved if the subjects had not been labeled mentally retarded.
Design
Interviews were designed to be administered to a sample of people still at Pennhurst in 1980, and then again after as each person was placed into a community living arrangement (CLA). The "pre" interviews at Pennhurst and the post" interviews in community settings asked the same standardized questions about resident satisfaction with the living situation, activities and services received, and general self-reported aspects of "happiness."
This simple pre-post consumer interviews design had not been implemented previously in any study of deinstitutionalization known to us. Even the pioneering work of Edgerton (1967), and Edgerton & Bercovici (1976) was based on interviews that began only after people had moved into community living. In related work, Birenbaum & Seiffer (1976) and Birenbaum & Re (1979) followed and interviewed adults for four years, and utilized a standardized questionnaire, but again the study began only after placement into community settings.
In our design, we waited about six months after each person's placement, and then conducted the post-placement interview. The first post-placement interview occurred in early 1981, the last in mid-1984.
We expected, on the basis of prior literature, that the people with the most functional skills (especially verbal) would probably be among the first to move to CLAs. Because the people in this part of the Pennhurst Study had verbal skills, we thought that, by the end of the study, most would be in CLAs. In fact, when the study was finished, only about half of the people in our Consumer Interview sample had left Pennhurst. (For convenience, this group will be referred to as "movers.")
This presented an opportunity to reinterview the people who were still at Pennhurst in 1984 ("stayers") and to check for changes in their self-reported satisfaction and happiness. This was not viewed as a control group, because there was no matching or random assignment, but rather as a convenient but non-equivalent group for whom the results would also be of interest. As institutional populations decrease during moves toward closure, it is important to know how such a situation affects the people who still live in those facilities. The results of interviews with the two groups, movers and stayers, were not intended to be compared to one another; they were two separate studies, each with its own set of policy implications.
Subjects
The sample of people interviewed in this part of the Pennhurst Study was not representative of the 1154 people who lived at Pennhurst in 1978, nor was it representative of all the people who moved to CLAs. Again, this was because the interview method itself biased the sample by excluding all people who were not able to communicate verbally (or by signing). Nevertheless, every effort was made to select a sample of people that would reflect the diverse elements of the verbal portion of the Pennhurst population.
Subject selection took place in Spring of 1980, after all design and instrument development was completed. The first stage of selection was to decide which people would be eligible for inclusion. Naturally, the people who had already left Pennhurst could not be included. It was also decided for economic reasons that, of the people still at Pennhurst, only the people who were originally from the greater Philadelphia area (the five southeastern counties of Pennsylvania) would be candidates.
Using this decision rule, there were 713 candidates for inclusion in the consumer interviews. These were all the people who lived at Pennhurst in May 1980, and who came from the Southeast Region of Pennsylvania. We then examined Behavior Development Survey data (collected at Pennhurst in 1978) to identify all the people who were reported to possess moderate or good verbal skills. There were 287 such individuals.
From these 287, we wished to select a representative sample. In the view of the Temple team, the best such sample would have been simple random. However, a consultant retained as an outside methodological reviewer by the government required a stratified sample of 60 people, with approximately 15 from each labeling category for level of retardation: mild, moderate, 'severe, and profound.
In our first stage of probabilistic selection, we oversampled from each of the four categories. By simple random selection, about 25 were taken from the moderate, severe, and profound categories; all 19 people labeled mild were taken. In all, 92 people were selected at this stage. The oversampling was in anticipation of losses due to our strict consent procedures.
Because we were only able to secure complete consent and valid interviews with 35 of these 92 people, a second stage of sample selection was initiated, by similar rules, in which 51 additional people were drawn. In all, then, we drew 143 candidates for interviews in this part of the study. By the completion of the baseline surveys, we had interviewed 56 people who lived at Pennhurst in the summer of 1980. The disposition of the sample is displayed in Exhibit 5-1.
Exhibit 5-1 shows that the people we interviewed were not representative of all the people at Pennhurst, nor even of all verbal people. People with fewer functional abilities were underrepresented from either point of view.
| EXHIBIT 5-1. Disposition of the Consumer Interviews Sample | ||||
|---|---|---|---|---|
| Reported Level of Retardation | ||||
| Mild | Moderate | Severe | Profound | |
| The universe of 713 | 21 (3%) | 55 (8%) | 197 (28%) | 440 (62%) |
| The 287 verbal people | 19 (7%) | 52 (18%) | 136 (47%) | 80 (28%) |
| The 143 drawn in sample | 19 (13%) | 45 (30%) | 43 (30%) | 36 (25%) |
| The 56 completed baseline interviews | 12 (21%) | 15 (27%) | 22 (39%) | 7 (13%) |
| The 30 Movers | 7 (23%) | 8 (27%) | 12 (43%) | 2 (7%) |
| The 26 Stayers | 5 (19%) | 7 (27%) | 9 (35%) | 5 (19%) |
The table also shows that, between stayers and movers, the differences in level of retardation were small but noticeable; again, these two groups were not treated as controls or comparisons.
Because of the way subjects were selected in this part of the Pennhurst Study, the consumer interviews should be viewed as (a) a case study of changes in the self-reported well-being of a specific group of deinstitutionalized people, (b) a case study of changes in the self-reported well-being of a specific group of people living in an institution as it phases down, and (c) an exploration of reliability and validity issues in direct consumer surveys.
Instruments
An extensive search for prior work in this area was initiated in 1979. The study team obtained copies of instruments used before, analyzed all available literature, and telephoned many of the researchers who had conducted such work. A draft instrument was developed from this groundwork in 1979. It was pilot tested and revised. In Spring of 1980 it was tested again, this time comparing telephone interviews to face-to-face interviews (Conroy & Beyer, 1979). The third revision was piloted during Summer 1980, and an entire new section was added to assess respondents' ability to label their own feelings accurately.
In the process of instrument development, the weight of prior research demanded primary attention to reliability. Sigelman, Winer, Schoenrock & Hromas (1978) focused on the problems of responsiveness, reliability or consistency, and response bias. The difficulties they noted were considerable; the suggestion they offered was that any such interview effort should include alternative format questions and checks for consistency. Winer, Sigelman, Schoenrock, Spanhel, & Hromas (1978) compared responsiveness to Yes-No, Either-Or, Multiple-Choice, and Open-Ended questions. The Yes-No format appeared to yield the highest proportion of responses and also the highest consistency. Yet Sigelman, Budd, Spanhel, & Schoenrock (1981) suggested that Yes-No questions were problematic because of a common tendency to say "Yes" to all questions, regardless of content; this was called the acquiescence phenomenon.
Our interview was designed with these studies in mind. It contained, in its final form, 12 Yes-No, 3 Either-Or, 4 Open-Ended, and an entire separate section of 7 Multiple-Choice (Likert scale) items with five facial drawings (big smile, small smile, neutral, small frown, big frown) to assist in labeling the way people felt about various issues. The questionnaire is included as Appendix 5-1.
An important facet of the interview instrument was the fact that there were six pairs of redundant questions. They were designed specifically as checks for consistency on the most important questions. For example, we asked, "Do you like living here?" (a Yes-No question), and later in the interview we asked "Would you like to leave here and live somewhere else?" (another Yes-No), and also "Which [face] is most like how you feel about living here?" (a Multiple-choice item with visual aids). These check items were intended to give the most weight to consistent responses.
Procedures
Interviews were generally scheduled by contacting the residential staff and then the individuals themselves. Appointments were made by telephone. The interview data were collected directly on the form in Appendix 5-1. Researchers at Temple edited the forms and entered the data directly onto mainframe disk storage, and conducted analyses using the Statistical Package for the Social Sciences (SPSS).
Internal Consistency: Acquiescence and Nay-Saying
The problem of acquiescence was first noted by Rosen, Floor, & Zistein (1974) in connection with interviews of people with mental retardation. More recently, it was investigated by Sigelman, Budd, Spanhel, & Schoenrock (1981). Their article, titled "When in doubt, say Yes," concluded that many people with retardation were likely to say "Yes" to any question that was not clear, concrete, and immediate. They speculated that this was part of a general tendency to avoid responses that "normal people might interpret as negative, resistive, or rebellious. In related work, Sigelman, et al., (1979) found a smaller number of people who acted in the opposite way, saying "No" to all questions - a phenomenon called nay-saying.
In their samples, Sigelman and colleagues found an acquiescence rate of 44% on Yes-No items, and a nay-saying rate of 4%. Because of their work, we included check questions for five of the Yes-No questions. They are shown in Exhibit 5-2, along with the results as to consistency.
Exhibit 5-2 shows, in the column headed "Acq," the number of people who displayed acquiescence on each item pair. This means that they said "Yes" to the Yes-No question, but then contradicted that answer on the check question. The column headed "Nay" works the same way for people who said "No" and later contradicted that answer.
| EXHIBIT 5-2. Acquiescence and Nay-Saying, Pre and Post | |||
|---|---|---|---|
| Question | # of Responses | # Acq. | # Nay |
| YES-NO VERSUS YES-NO | |||
| Q3: Do you want to keep on living here? | Pre:
55 Post: 53 |
16 8 |
0 0 |
| Q16: If you could, would you like to leave here and live somewhere else? | |||
| YES-NO VERSUS SCALE | |||
| Q1: Do you like living here? | Pre:
48 Post: 46 |
6 1 |
6 0 |
| Q7B: Which face is most like how you feel about living here? | |||
| Q13: Do you like your day program? | Pre:
46 Post: 42 |
4 1 |
2 0 |
| Q10B: Which face is most like how you feel about your day program? | |||
| Q2: Do you like the people who work here? | Pre:
48 Post: 45 |
6 2 |
1 1 |
| Q11B: Which face is most like how you feel about the staff? | |||
| YES-NO VERSUS EITHER-OR | |||
| Q2: Do you like the people who work here? | Pre:
54 Post: 53 |
3 1 |
5 1 |
| Q7: Are people here mean or nice? | |||
| OVERALL | Pre: 251 Post: 239 |
35 13 |
14 2 |
There is a lot of information in Exhibit 5-2, but there are really just three main points. First, our overall rate of acquiescence in the baseline interviews at Pennhurst was 35 occurrences out of 251 possible occurrences, or 14%. This was much less than the rate of 44% reported by Sigelman et al. (1981). Second, our baseline rate of nay-saying was 6%, about the same as the Sigelman et al. rate of 4%. Third, our rates of inconsistent responses declined sharply in the post interviews; the rate of acquiescence in the post-test was 5% and the nay-saying rate was 1%. This decline was statistically significant (even by the relatively conservative nonparametric Wilcoxon T test, p<.001). Further investigations revealed that significant declines in inconsistencies occurred among the movers and among the stayers.
Internal Consistency: Recency
Spanhel, Sigelman, Schoenrock, Winer, & Hromas (1978) reported that 28% of the responses of institutionalized children to Either-Or items were inconsistent because of "recency." For example, when asked "Are you big or small?" and later "Are you small or big?," 19% of the children chose the most recently heard option both times (small the first time and big the second time), and another 9% chose the first option offered both times. Our questionnaire contained one pair of questions to check recency:
Q8: Are you usually happy or sad?
Q15: Are you usually sad or happy?
In the baseline interviews, 53 people responded to both items. Among Item, 3 people chose the first option on both questions, and 8 people chose the second option on both questions, for a combined "recency" inconsistency rate of 21%. This was somewhat less than in the prior work of Spanhel et al., probably partly because of our screening procedures and partly because Spanhel et al. were dealing exclusively with children. In our second round of interviews, there were 54 people who responded to both questions. None of them chose the first option on both questions, and nine chose the second item on both questions (17%). This was not significantly different from the baseline recency rate.
Changes in Satisfaction: The Movers
Of the 56 people interviewed at Pennhurst in the 1980 baseline, 30 had moved to community living arrangements (CLAs) and had been reinterviewed there by 1984 (movers). This section presents our findings for these movers.
In the baseline, 18 of the 30 movers had said "Yes" in answer to the question "Do you like living here?" However, as shown in the upper part of Exhibit 5-3, four of those 18 later contradicted themselves on the check question by indicating that they felt "Sad" or "Very Sad" about living there. In the table, these four can be seen in the "Yes" column (one sad and three very sad). The people who were consistent in their responses are marked with an asterisk; those who contradicted themselves are marked with parentheses.
| EXHIBIT 5-3. Movers’ Satisfaction with Where They Live | ||||
|---|---|---|---|---|
| PRE: At Pennhurst | Q1: Do You Like Living Here? | |||
| Yes | In Between | No | ||
| Q7B: Which face is most like how you feel about living here? | Very Happy | 9* | 0 | (1) |
| Happy | 3* | 0 | (2) | |
| Neutral | 2 | 3* | 1 | |
| Sad | (1) | 0 | 0* | |
| Very Sad | (3) | 1 | 1* | |
| (3 people did not respond) | ||||
| POST: In Community Living Arrangements | Q1: Do You Like Living Here? | |||
| Yes | In Between | No | ||
| Q7B: Which face is most like how you feel about living here? | Very Happy | 21* | 1 | (0) |
| Happy | 1* | 0 | (0) | |
| Neutral | 1 | 1* | 1 | |
| Sad | (0) | 0 | 0* | |
| Very Sad | (0) | 0 | 0* | |
| (4 people did not respond) | ||||
The table revealed that these verbal individuals had increased in their self-reported level of satisfaction with their living arrangements, but the data in the table must be interpreted carefully. In the baseline, at Pennhurst, 12 people, or 40% of the sample, reliably expressed satisfaction with living there; conversely, one person (3%) was reliably dissatisfied. Later, in CLAs, 22 people, or 73% of the sample, reliably expressed satisfaction, and no one was consistently dissatisfied. By this measure, satisfaction had almost doubled. On the facial picture scale item, the increase in expressed was tested both with the parametric t-test (t=4.30, (24), p<.001) and with the nonparametric Wilcoxon T (p<.001).
A condensed presentation of the responses of the movers to the check question described above, and to the five other sets of check questions, is given in Exhibit 5-4.
| EXHIBIT 5-4. Summary of Movers’ Reliably Expressed Satisfaction Before and After CLA Placement | ||||
|---|---|---|---|---|
| Before | After | Change | ||
| Satisfaction with Living Arrangement (Q1 and Q7B) | Satisfied | 40% | 73% | +33% |
| Dissatisfied | 3% | 0% | -3% | |
| Desire to Move (Q3 and Q16) | Satisfied | 43 | 63 | +20 |
| Dissatisfied | 17 | 7 | -10 | |
| General Happiness (QB and Q15) | Satisfied | 67 | 67 | 0 |
| Dissatisfied | 3 | 0 | -3 | |
| Satisfaction with Staff (Q2 and Q7) | Satisfied | 60 | 80 | +20 |
| Dissatisfied | 0 | 0 | 0 | |
| Satisfaction with Staff (Q2 and Q11B) | Satisfied | 53 | 63 | +10 |
| Dissatisfied | 7 | 0 | -7 | |
| Satisfaction with Day Program (Q13 and Q10B) | Satisfied | 53 | 53 | 0 |
| Dissatisfied | 0 | 7 | +7 | |
The figures in Exhibit 5-4 reflect only the consistent responses, and all the percentages are taken as fractions of the entire 30 people in the movers group. We have already discussed the first change in the table, Living Arrangement. The second change was in Desire to Move, which decreased; at baseline 17% wanted to move and after relocation it was 7% (Wilcoxon T, p<.01). On the General Happiness questions (Are you usually happy or sad), the table shows that there was no change. About two thirds reliably said they were usually happy, both while they were at Pennhurst, and later in the CLAs. On both sets of check questions about staff, the proportion of people who reliably reported positive feelings increased after CLA placement (Wilcoxon T, p<.05). Finally, there were no significant changes in satisfaction with the day programs; although not statistically significant, it is worth noting that this was the only area in which there was increased dissatisfaction; two people reliably expressed dissatisfaction with their community based day programs.
Thus, in four of the six areas of satisfaction in which the consistency and reliability of responses could be checked, satisfaction increased; in the other two areas, satisfaction was unchanged.
There were also a number of questions for which there were no check questions. There were no significant changes from pre to post relocation for "Do you have any real good friends?" or "Do you ever see anyone in your family?" or "Do you make any money?" A significant increase was noted for "Do you have a girlfriend/boyfriend?" from 10 people saying "Yes" in the baseline to 17 saying "Yes" after relocation to CLAs.
The smile face Likert scale items were of special interest, and further analyses of change were undertaken. The special interest arose from prior reports of failure of this question format (Winer, et al., 1978) because too few people could respond to it at all; yet, if it could work, the data from a five point scale might be more useful than simple Yes-No answers. As has already been noted, in our sample, the smile face format worked fairly well; response rates did not drop Much below those of the Yes-No and Either-Or formats. It was therefore possible to treat the seven smile face items as ordinal scales, calculating average scores on each one before and after relocation, and to use routine statistical tests of significance of change. For each item, a score of 1 meant the "big frown" face, and a "5" meant the "big smile." Thus higher scores were more positive. The results are presented in Exhibit 5-5.
| EXHIBIT 5-5. Changes on Smile Scale Items After Relocation | |||
|---|---|---|---|
| Which fact is most like how …” | Mean
Score Before |
Mean
Score After |
Significance of t (T) |
| Q7B you feel about living here? | 3.4 | 4.7 | 0.001 (0.001) |
| Q8B the staff feel about you? | 3.5 | 4.3 | 0.021 (0.028) |
| Q9B the other residents feel about you? | 3.4 | 4.0 | 0.076 (0.096) |
| Q10B you feel about your day program? | 3.9 | 3.9 | 1.00 (1.00) |
| Q11B you feel about the staff? | 3.8 | 4.3 | 0.109 (0.140) |
| Q12B you feel about the other residents? | 3.3 | 4.3 | 0.021 (0.026) |
| Q13B you feel about yourself? | 4.1 | 4.3 | 0.484 (0.469) |
| OVERALL SCALE | 25.5 | 30.0 | 0.001 (0.003) |
The test of significance of change from before relocation to after was the simple paired t-test. The sample size was often less than 30 because not everyone answered every question; therefore we also ran the nonparametric Wilcoxon T tests. Significances of the Wilcoxons are shown in the Parentheses. The t and the Wilcoxon were nearly identical in each case.
The largest and most significant change was in how people felt about where they lived, which became more positive in the CLAs. Significant changes were also noted in people's beliefs about how staff felt about them, and how people felt about the other residents. When all seven Likert items were added up to form a single satisfaction scale, the change on this "overall scale" was also significant. Results on the overall scale showed that the movers were more satisfied in three of these seven areas, and also overall, after they moved into the CLAs.
Changes in Satisfaction: The Stayers
For the 26 people we interviewed in 1980 at Pennhurst who were still living at Pennhurst in 1984, it was of interest to find out whether they had changed in any areas of satisfaction/happiness. Certainly, the four years had been eventful ones in the history of Pennhurst. The population declined from about 1000 to about 450 in those years, some buildings had closed, some staff had been furloughed, and it had been announced by the Department of Public Welfare that Pennhurst definitely would close. For these reasons, we conducted reinterviews with the 26 people in the Summer of 1984.
Exhibit 5-6 shows a summary of the changes in satisfaction among Stayers on the items for which we had check questions.
| EXHIBIT 5-6. Summary of Stayers’ Reliably Expressed Satisfaction in 1980 and in 1984 | ||||
|---|---|---|---|---|
| 1980 | 1984 | Change | ||
| Satisfaction with Living Arrangement (Q1 and Q7B) | Satisfied | 42% | 35% | -7% |
| Dissatisfied | 12% | 27% | +15% | |
| Desire to Move (Q3 and Q16) | Satisfied | 35 | 27 | -8 |
| Dissatisfied | 27 | 35 | +8 | |
| General Happiness (Q8 and Q15) | Satisfied | 50 | 58 | +8 |
| Dissatisfied | 8 | 15 | +7 | |
| Satisfaction with Staff (Q2 and Q7) | Satisfied | 65 | 69 | +4 |
| Dissatisfied | 4 | 0 | -4 | |
| Satisfaction with Staff (Q2 and Q11B) | Satisfied | 38 | 50 | +12 |
| Dissatisfied | 12 | 0 | -12 | |
| Satisfaction with Day Program (Q13 and Q10B) | Satisfied | 38 | 58 | +20 |
| Dissatisfied | 4 | 4 | 0 | |
The data in Exhibit 5-6 indicate, if anything, a slight decrease in satisfaction with the living situation, as evidenced by the consistent responses to the first two pairs of check questions, on which satisfaction decreased slightly and dissatisfaction increased slightly. General happiness appeared to increase for some, and decrease just as much for others. Changes regarding satisfaction with staff were all in a positive direction. The largest change was an increase in satisfaction with the day program. Statistical tests, however, showed that none of these changes were significant.
The unchecked items regarding good friends, girlfriends and boyfriends, family contact, and making money were also examined for change from 1980 to 1984. There were no significant changes in these areas.
| EXHIBIT 5-7. Changes on Smile Face Scale Items Among Stayers, 1980-1984 | |||
|---|---|---|---|
| Which fact is most like how …” | Mean Score Before | Mean Score After | Significance of t (T) |
| Q7B you feel about living here? | 3.7 | 2.9 | 0.074 (0.075) |
| Q8B the staff feel about you? | 3.4 | 3.6 | 0.709 (0.638) |
| Q9B the other residents feel about you? | 3.9 | 4.0 | 0.774 (0.790) |
| Q10B you feel about your day program? | 3.7 | 4.5 | 0.111 (0.139) |
| Q11B you feel about the staff? | 3.6 | 4.2 | 0.276 (0.272) |
| Q12B you feel about the other residents? | 3.7 | 3.6 | 0.822 (0.875) |
| Q13B you feel about yourself? | 3.5 | 4.0 | 0.394 (0.394) |
| OVERALL SCALE | 26.8 | 26.9 | 0.940 (0.638) |
As we did for the Movers, we treated the face scale items as ordinal data and computed averages and tests of change over time. The results of this analysis for the stayers are presented in Exhibit 5-7.
Remarkably, non of the changes reached even the .05 level of statistics significance. The decreased satisfaction with the living situation came close, as did the rise in satisfaction with the day program. But strictly speaking, we cannot infer that there were any real changes in these measures of satisfaction.
The central question of the Pennhurst Longitudinal Study for the Temple University part of the effort was "Are people better off?" In the consumer interviews section of the study, the answer seems to be that the people (in our sample of verbal people) who have moved to CLAs are, in fact, better off. They are better off in terms of their own verbally expressed satisfaction with various areas of their lives, particularly with the place where they live.
In our explorations of reliability, we found generally higher consistency than in prior work, but we certainly agree with the body of work by Sigelman and colleagues that it is essential to include check questions in this kind of work. Hence asking questions in several ways, and in several formats, is important. Answers given to varied formats must be compared, and then the presentation of the results should give weight to the consistent, reliable responses. We believe that the extra effort required to perform quality interview work with people with mental retardation is amply justified.
This study revealed no strong preference as to the best question formats to use with people with mental retardation. Probably because of our preselection of people with verbal skills, nearly everyone was able to respond to all the formats (Yes-No, Either-Or, Multiple choice, Open-ended) most of the time.
Our surprising finding of sharply reduced inconsistency rates on the second interview was of considerable, although subsidiary, interest. Many explanations for the phenomenon are possible, including the idea that the first interview may have been the first time the people were asked for their opinions in a formal way by a stranger, and that, with even a little practice, they became more able to respond in such a situation. Another concerns the possibility of increased trust of, and rapport with, our interviewer. Similarly, it is possible that our interviewer gained in skill in probing answers by the time of the second interviews. If any of these explanations were the case, they could pose a threat to the validity of the increased satisfaction findings since improved ability to respond to interviews, or improved openness, or improved interviewer technique could all be potential explanations for the changes in satisfaction. However, both the movers and the stayers displayed sharp reductions in contradictions, but only one group showed the increases in satisfaction, so there does not seem to be a direct threat to validity in this area.
Finally, the stayers in this sample did not change significantly in their self-expressed satisfaction. There was a suggestion of increased satisfaction with the day program; one would hope that the decreased population of the Pennhurst Center has enabled more people to attend day programs, and to receive more individual attention when they do.
Originally, we did not expect to be able to investigate changes among the stayers, which could help to illuminate the effects of institutional phase-downs, but the opportunity to do so was welcome. We hope that similar work will continue here and elsewhere, so that the feelings of people who have lived in facilities for decades are taken into account as those facilities are phased down.
Birenbaum, A. & Re, M.A. (1979). Resettling mentally retarded adults in the community -- almost 4 years later American Journal of Mental Deficiency, 83, 323-329.
Birenbaum, A. & Seiffer, S. (1976). Resettling retarded adults in a managed community. New York: Praeger Special Studies.
Conroy, J. & Beyer, F. (1979). A comparison of two interview methods with mentally retarded persons: Personal and telephone. Philadelphia: Temple University Developmental Disabilities Center.
Edgerton, R.B. (1969). Cloak of competence. Berkeley: University of California Press.
Edgerton, R.B. & Bercovici, S.M. (1976). The cloak of competence: Years later. American Journal of Mental Deficiency, 80, 485-497.
Rosen, M., Floor, L., & Zistein, L. (1974). Investigating the phenomenon of acquiescence in the mentally handicapped: I -- Theoretical model, test development, and normative data. British Journal of Mental Subnormality, 20, 58-68.
Seltzer, G. (1980). Residential satisfaction and community adjustment. Paper presented at the American Association on Mental Deficiency, San Francisco, May 1980.
Sigelman, C.K., Budd, E.C., Spanhel, C.L., Shoenrock, C.J. (1981). When in doubt, say year: Acquiescence in interviews with mentally retarded persons. Mental Retardation,19, 53-58.
Sigelman, C.K., Schoenrock, C.J., Spanhel, C.L., Hromas, S.G., Winer, J.L., Budd, E.C., & Martin, P.W. (1980). Surveying mentally retarded persons: responsiveness and response validity in three samples. American Journal of Mental Deficiency, 84, 479-486.
Sigelman, C.K., Schoenrock, C.J., Winer, J.L., Spanhel, C.L., Hromas, S.G., Martin, P.W., Budd, E.C. & Bensberg, G.J. (1979). Issues in interviewing mentally retarded persons: An empirical study. Lubbock, Texas: Research & Training Center in Mental Retardation, Texas Tech University.
Sigelman, C.K., Winer, J.L., Schoenrock, C.J., Hromas, C.L. (1978). Performance of the mentally retarded in interviews: Responsiveness to questions. Lubbock, Texas: Research & Training Center in Mental Retardation, Texas Tech University.
Spanhel, C.L., Sigelman, C.K., Schoenrock, C.J., Winer, J.L., Hromas, S.G. (1978). The feasibility of interviewing the mentally retarded: responsiveness, reliability, and response bias. Lubbock, Texas: Research & Training Center in Mental Retardation, Texas Tech University.
Winer, J.L., Sigelman, C.K., Schoenrock, C.J., Spanhel, C.L, Hromas, S.G. (1978). The performance of mentally retarded children on repeated and alternative format interview questions. Lubbock, Texas: Research & Training Center in Mental Retardation, Texas Tech University.
|
Assessment of Environments Case Study: Access to Generic Resources Joan left Pennhurst in June 1980, after having lived there since May 1969. Joan has Down's Syndrome and is legally blind. At years of age, Joan had eye surgery which revealed a congenital cataract in her "good eye." As she got older, her eyes began to atrophy, as did the muscles around them. Joan had significant instances of self-abusive behavior while at Pennhurst. Her self-abusive behaviors consisted of face-slapping, mainly around her eyes. In addition, Joan has been known to spit at and pinch others. Joan's move to the community in the summer of 1980 was fairly uneventful. She moved into her new home in the suburbs with 2 other women, one of whom had lived at Pennhurst and the other of whom had lived at another state-operated mental retardation facility. Joan seemed to adjust to her new home fairly well. She learned new skills at a steady pace, yet her inappropriate behaviors remained the same. Over the next two years, Joan's self-abusive behaviors increased steadily, especially face-slapping to the area around her eyes. The community doctor believed there was no medical problem and did not deem it necessary to bring Joan into his office for a visit. In December of 1982 a new project director took over Joan's program. When she assessed Joan's behavior problems, she made several changes, including bringing in a new house team leader and getting a new behaviorist and general practitioner. Joan's parents were quite upset with the regression their daughter was experiencing, and contacted staff on a daily basis. The project director met with the Harris's and suggested that Joan's problems with self-abuse may have been due to irritation in or around her eyes. When the project director suggested an evaluation at Wills Eye Hospital, Mr. and Mrs. Harris were hesitant, as they believed that Pennhurst had exhausted all options with regard to Joan's vision or lack thereof. After some coaxing, the Harris's consented to an evaluation at Wills. The evaluation concluded that, due to the atrophy of Joan's eyes and the muscles surrounding them, her upper and lower eyelids had grown inward, causing her irritation and pain. The opthamologist suggested that Joan should be considered for prostheses to alleviate the irritation. In February 1983, after numerous fittings and close communication with one of the only occularists in the city, Joan received her prostheses. Over the past year Joan's behavior has improved considerably. The incidence of self-abuse has decreased appreciably, and when Joan does slap herself it is never around her eye area. Joan seems very happy with her new eyes, and, most important, she is no longer in pain. |
In this part of the Pennhurst Study, we address the question of whether people are "better off" in terms of the qualities of the places in which they live. We have consistently used the phrase "qualities" of environments to emphasize the fact that there is no generally accepted measure of quality; instead, there are many measures of environmental quality in use, and we have used several.
In the first part of this chapter, we describe the methods and results of our investigation of differences between Pennhurst and the CLAs in terms of normalization and individualization. In the second part, we present a summary of our efforts to identify and measure aspects of community residential settings that are correlated with developmental progress among the people living in them.
Instruments
Four dimensions of the environmental program quality of the service setting were measured at the institution: (1) PASS-3 (Program Analysis of Service Systems; Wolfensberger & Glenn, 1975), a widely used measure of normalization; (2) selected portions of the Accreditation Council Standards for Mental Retardation & Developmental Disabilities (ACMRDD), chosen by ACMRDD field experts to measure physical and (3) programmatic aspects of the environment; and (4) the Resident Management Survey (King, Raynes, & Tizard, 1971; Balla, 1976), which measured the extent to which treatment was institution-oriented versus individual-oriented, or, in other terms, the degree of individualization versus regimentation.
PASS-3 may be thought of as a quantification of the normalization ideology. It is the oldest and most widely used instrument for that purpose. As it is usually applied, about six to 15 person days are needed for a complete 50-item rating. Because our resources would not permit that level of effort for each of hundreds of CLAs, it seemed that PASS could not be included among our environmental measures. After considerable literature review and nationwide contact with experts, a solution was found. Flynn & Heal (1981) had developed a shortened version of PASS-3. They identified an 18-item subset that was correlated at r=.965 with the full 50-item PASS-3 scale. We concluded that the 18-item short form, administered by highly experienced raters, would be ideal for this study.
The ACMRDD standards consisted of 807 Yes-No items. In August 1979 the project engaged Mr. Terry Perl, former head of the Survey Procedures Committee of ACMRDD, and Mr. William Snauffer, director of a corporation that employed experienced ACMRDD field surveyors, as consultants. The purpose o the consultation was to reduce the ACMRDD standards to two subsets, focused on physical standards and program standards. From the full 807 standards, 323 were selected as core items representing physical and programmatic aspects of environments. The core item checklist was pilot-tested at a residential school in Maryland. Two survey teams of four members each performed independent evaluations in order to assess inter-team reliability. The consultants then selected 41 items concerning the physical environment and 106 for programming that were most readily applicable to both institutional and community programs.
After the institutional assessments were conducted, and the data analyzed, it was decided that use of the modified physical and program standards of ACMRDD be terminated. Our attempts to identify any relationship between individual growth within the institution and either ACMRDD environmental score had met with no success. After trying simple correlations, partial correlations controlling for individual characteristics, and multiple regression of various forms, we had not been able to detect a relationship. Moreover, ACMRDD central office staff and at least one board member took strong exception to this experimental study of the standards. Among the public objections were the contention that the standards should not be considered as a scale, that the institutional cottage sampling was inadequate, that the specific items selected and the way they were selected were questionable, and that one of our methods of statistical analysis was misleading. Our repeated offer to provide the data tape for ACMRDD to conduct its own search for a relationship between growth and the ACMRDD characteristics of the living area received no response. In this atmosphere, and because the ACMRDD standards were extremely labor-intensive and expensive to collect, we decided to abandon all efforts to validate the utility of those standards.
The Resident Management Survey (RMS) was designed to differentiate institution-oriented from individual-oriented care practices. King, et al. (1971) used this scale to compare care practices in institutions (size 121-1650), voluntary homes (50-93), and hostels or group homes (12-41). They found that the instrument was a sensitive measure of individualized versus regimented treatment, with the group homes being the most individualized and the institutions the least. McCormick, Balla, and Zigler (1975) later replicated these findings and extended them cross-culturally. More recently, the instrument was adapted for wide use in conjunction with the Individualized Data Base at UCLA. Because of its wide use, prior findings, and the theoretical importance of the RMS in comparing institutions to community settings, it was included.
Sample
At the institution. In October 1978 there were 45 living areas at Pennhurst. The first principle of our approach was that we should not do one environmental rating for the whole institution since there was likely to be considerable variation among living areas. We could not, however, rate every living area. Therefore, it was necessary to crosstabulate the characteristics of the people in the living areas (using our 1978 Pennhurst behavioral and demographic data) and look for natural clusters of similar living areas. When this analysis was performed, Pennhurst fell into 10 clusters of living areas. We then randomly selected one living area to represent each cluster. We wanted to be able to assign a normalization score, two ACMRDD scores, and an RMS score to each individual's living area as accurately as possible.
In the community. With respect to environmental ratings in the community, sampling was not possible. We had no data at the beginning of the study to even test the clustering idea. Therefore, each CLA was rated along all environmental dimensions.
We decided to add three other environmental quality instruments before we began the community phase of data collection: the Life Safety Codes Instrument, Characteristics of the Treatment Environment (CTE), and Characteristics of the Physical Environment (CPE).
The Life Safety Codes Instrument was developed by the Evaluation & Research Group at Temple University's Developmental Disabilities Center. It recorded adherence to life safety codes, emergency procedures, staff preparation for emergencies, and so forth. This instrument also contained selected items from the standards for intermediate care facilities.
Characteristics of the Treatment Environment (Jackson,1969) was developed to measure the degree to which autonomy and activity are encouraged in the residential setting. It was revised in 1977 (Silverstein, McClain, Hubbell and Brownlee, 1977). Silverstein et. al. identified 10 items from Jackson's original scale that produced the highest item-factor correlations with the scale's two factors: autonomy and activity. This instrument was designed to be collected by interview with appropriate CLA staff.
Characteristics of the Physical Environment was developed by the Developmental Disabilities Project on Residential Services and Community Adjustment at the University of Minnesota (1981). This instrument measured the degree to which the environment was home-like. Each of five rooms was assessed on a five-point scale with "1" indicating a very home-like environment and "5' indicating a very non-home-like environment. This instrument was designed to be completed by the site reviewer after direct observation of the residence.
Procedures
At the Institution. For Normalization and RMS ratings, it was desirable to locate a number of people highly familiar with PASS-3, because normalization assessment in the field presupposed intensive training. We were supplied with a list of 18 persons who were not only familiar with PASS-3, but were qualified as PASS-3 Team Leaders or Assistant Team Leaders. A training workshop was held in September 1979. The 18-item short form of PASS-3 (which we will henceforth call the Normalization Instrument, because it is not actually PASS-3) and the RMS were presented and explained. The normalization and RMS assessments in the institution were performed by two-person teams in September 1979. The interrater agreement appeared to be sufficiently high (Flynn & Heal, 1981) to justify later reduction of field team size in the community to one rater per site. This was seen to be cost effective, as well as less intrusive.
The condensed ACMRDD surveys were performed by a team of three qualified and experienced surveyors. For the Physical Standards section, the surveyors performed an on-site inspection to complete their checklist of 41 items. For the Program Standards section, the Principal Investigator was asked to draw a small simple random sample of three to six individuals in each selected living area. The surveyors assessed the individual records of each individual thus drawn, visited each unit, interviewed staff, and completed their 106-item Program Standards checklist for each individual.
The institutional environmental data were coded and keypunched and entered into the computer record of each person at Pennhurst. Each individual was given a normalization score, an RMS score, an ACMRDD physical standards score, and an ACMRDD program standards score.
In the Community. At the second training session, held in early 1982, site reviewers were retrained in Normalization and the RMS, and were trained in the use of the three new environmental instruments (CTE, CPE, Life Safety Codes). The three new instruments added approximately 1/2 hour to the review.
In March 1982, data collection began in the community. As of that time, approximately 200 people had been relocated from Pennhurst to the community. One site reviewer went to each site where a former Pennhurst resident lived; each reviewer collected the Normalization Scale, the RMS, the CTE, the CPE, and the Life Safety Codes instruments (in addition to a Behavior Development Survey for each individual). Once the data were collected, they were entered onto the record of each individual, thus enabling comparison between institutional and community scores on the environmental instruments.
Within Pennhurst
In 1980, the Behavior Development Survey was collected for all 713 individuals who remained at Pennhurst, and whose county of origin was one of the five counties in the Southeast. Region of Pennsylvania (Bucks, Cheater, Delaware, Montgomery, and Philadelphia). Comparison of those BDS scores to the ones collected in 1978, revealed that people had gained an average of 1.24 points in adaptive behavior.
The environmental variables were tested for relationship to the amount of behavioral growth displayed by the people in the Pennhurst living areas. In one approach, we examined simple correlations, in a second approach we used partial correlations controlling for 1978 adaptive behavior, and in a third we used several forms of multiple regression. In the regression analyses, we forced individual characteristics to enter the equation first, because the nature of the question we were asking was whether environmental variables could account for individual growth above and beyond the growth that was accounted for by unchangeable individual characteristics (e.g., sex, age, or level of retardation).
Above and beyond the growth that could be explained by unchangeable individual characteristics, we identified a few programmatic variables that showed suggestions of statistical significance, depending on the choice of statistical technique. The analyses suggested that, individual characteristics being equal, greater time in day program could make a difference in growth, as could individualized treatment (as measured by the RMS) and fewer medications daily. In addition, other forms of analysis implied some effects of compliance with the ICF standards, smaller living areas, more staff, and residential continuity. However, regardless of the statistical procedures used, these programmatic variables could not account for very much of the variation in growth among the people living at Pennhurst. Compared to unchangeable individual characteristics, these program and environmental variables appeared to be relatively weak in predicting, or explaining, variations in individual growth.
1982 Community Data
The results of the first round of data collection in the community are summarized in Exhibit 6-1, which gives the average Normalization scores and RMS scores for individuals from the five counties while they were residing at Pennhurst and once they had moved to the community.
| EXHIBIT 6-1. Average Normalization and RMS Scores for Institution and Community by County | |||||
|---|---|---|---|---|---|
| County | N | Normalization | RMS | ||
| Pennhurst | CLA | Pennhurst | CLA | ||
| 1 | 14 | -239 | 152 | 54 | 66 |
| 2 | 22 | -237 | 163 | 55 | 66 |
| 3 | 29 | -247 | 110 | 60 | 64 |
| 4 | 34 | -226 | 177 | 58 | 64 |
| 5 | 58 | -226 | 207 | 58 | 65 |
| Average scores | -232 | 172 | 58 | 65 | |
| Average change | 404 | 7 | |||
The people who had moved into CLAs had clearly experienced a large increase in the degree of normalization (from -232 to +172), and a significant increase in the degree of individualization (from 58 to 65), as measured by our short version of PASS-3 and by the RMS. The conclusion from these measures was that people who had moved to CLAs were better off in terms of these two environmental qualities.
The county tabulation shows, in addition, that there were significant variations among Normalization scores received by CLA programs in different counties; however there was practically no variation among RMS scores in CLAs in different counties. This illustrates why the RMS was eliminated from the CLA data collection process; all CLAs were at or very near the top of the scale.
Instruments
For the second round of community data collection, the environmental assessment package was revised. We decided to keep the normalization measure derived from PASS-3, because the questions it addressed were basic and essential, and were not addressed by the other environmental measures. The Resident Management Survey was dropped in 1983, after the institution to community changes had been assessed. The RMS was replaced with the Group Home Management Schedule (Pratt,1969), another measure of individualization versus regimentation, but designed to be more sensitive to the less obvious variations among community programs.
The Characteristics of the Treatment Environment was also dropped in 1983 for the same reason the RMS was eliminated: almost all CLAs received the highest possible score. The study team decided that replacement of this instrument was unnecessary because our normalization scale covered the same or similar areas.
The Characteristics of the Physical Environment was replaced by the Physical Quality Instrument (PQI) (taken from a modification of the MEAP Rating Scale created by Seltzer, 1982, and further modified by our group). The Physical Quality Instrument was found to be a more thorough measure of the pleasantness of the residential site. This instrument also assessed the physical quality of the neighborhood in which the homes were located. As with the CPE, the PQI was completed by the site reviewer after the site review, including a tour of every room.
Procedures
Since approximately 100 additional individuals had been relocated from Pennhurst to the community, it was necessary to recruit additional site reviewers to complete the environmental assessments. We recruited 10 more individuals who had been PASS trained through the Commonwealth of Pennsylvania's PASS training program. At a four day training session held late in 1983, site reviewers were retrained in the Normalization Scale, and were trained in the use of the three other environmental instruments (GHMS, PQI, Life Safety Codes). The three instruments took approximately 1/2 hour to administer, the same as in the previous year.
Data collection occurred late in 1983 and early in 1984. One site reviewer went to each site where a former Pennhurst resident lived; each reviewer collected the BDS, the normalization scale, the GHMS, the Life Safety Codes Instrument, and the PQI.
One of the original aims of the Pennhurst Study was to explore the differences in environmental qualities between institution and community, and we did so. Equally important was the question of what environmental qualities in community programs would "make a difference." That is, it was important for policy makers and program operators alike to know how programs could best be designed to foster individual growth and development. We therefore used the Pennhurst Study data set to investigate whether any of our environmental quality measures were associated with individual growth and development among people living in CLAs.
The analysis presented here was based on the data collected for all people in CLAs in 1983 and 1984. This was the most recent information available, and it also included the largest number of people. We further selected people who were living at exactly the same CLA, with the same street address, in 1983 and 1984, in order to eliminate people who had changed environments, even if they only moved to an apartment across the hall. This assured us that the physical environment, at least, was relatively constant. There were 320 people in the study's data base who met these criteria.
The index of individual growth and development was the change in adaptive behavior from 1983 to 1984 as measured by the Behavior Development Survey. The 320 individuals in the analyses had gained an average of 2.0 points between 1983 and 1984; actual changes ranged from a 45 point loss to a 34 point gain.
Literally hundreds of variables were available to test for association with growth, but our interest in this analysis was in the environmental variables. The first analysis was a simple Pearson correlation. The variables selected were the following environmental measures: number of other residents living at the site, hours of developmental service, hours of day programming, Group Home Management Schedule score, Physical Quality score, Normalization Score, Characteristics of the Treatment Environment score and total staff hours.
Of all the variables entered into the Pearson correlation, shown in Exhibit 6-2, the only significant correlation was between adaptive behavior growth and the Group Home Management Schedule score (r=-.20, 314 df,p=.001). This suggested that individuals living in more regimented settings gained more in the area of adaptive behavior. This was a paradoxical finding, because the prevailing wisdom indicated that more regimentation would inhibit growth.
| EXHIBIT 6-2. Associations Between Environmental Measures and Individual Growth Within CLAs | |
|---|---|
| Simple Correlation with
1983-1984 Gains in Adaptive Behavior Score |
|
| Number of residents | -0.09 |
| Hours of developmentally-oriented service per day | 0.05 |
| Hours of day program per day | 0.00 |
| Group Home Management Schedule | -0.20* |
| Physical Quality Instrument | -0.06 |
| Normalization Instrument | 0.04 |
| Total staff hours per week | -0.06 |
| Characteristics of the Treatment Environment | -0.10 |
It is a well established fact that much of the variance in gain scores, no matter what the context, can be accounted for by initial scores (Cronbach & Furby, 1970). In our case, the 1983 adaptive behavior scores were significantly correlated with the gain scores (r=-.24, (318), p<.001). The negative correlation meant that people who started out with lower scores were likely to gain the most, and people who started out with higher scores were likely to gain the least.
We also know that many environmental measures are correlated with the level of functioning of the people living in the environments; for example, the Group Home Management Schedule measures the degree to which the environment fosters expression of individual choices, as opposed to having blanket rules and regimentation for all. But obviously there would tend to be less evidence of individual freedom of choice among people with profound mental retardation than among people with mild mental retardation. Hence such a measure might yield higher scores for settings with higher functioning people.
If the environmental measures are correlated with initial adaptive behavior, and initial adaptive behavior is negatively correlated with gain scores, then possible relationships between the environmental variables and gain may be masked if we rely solely on simple Pearson correlations. It is useful to try to remove this confounding influence from the analysis. One mathematical way of doing so is to use partial correlations. A partial correlation gives a measure of the relationship between two variables while adjusting for the effects of one or more additional variable. (As an example, suppose it is found that there is no correlation between the number of firefighters and the speed of putting a fire out. Should the mayor cut the number of firefighters? No, because partial correlation shows a strong relationship between the number of firefighters and the speed of extinguishing, when we adjust for the size of the fire.)
| EXHIBIT 6-3. Partial Correlations Between Environmental Measures and Individual Growth Within CLAs, Controlling for 1983 Adaptive Behavior | ||
|---|---|---|
| Simple Correlation with 1983 Adaptive Behavior |
Partial Correlation with Gain in Adaptive Behavior |
|
| Number of residents | -0.14* | -0.13* |
| Hours of developmentally-oriented service per day | -0.04 | 0.04 |
| Hours of day program per day | 0.20** | 0.05 |
| Group Home Management Schedule | 0.43** | -0.12* |
| Physical Quality Instrument | 0.02 | -0.05 |
| Normalization Instrument | 0.31** | 0.12* |
| Total staff hours per week | -0.47** | -0.20** |
| Characteristics of the Treatment Environment | 0.47** | 0.00 |
| *
p<0.05; ** p<0.001 |
||
To see how strongly our environmental measures were influenced by the level of functioning of the people in the settings, we computed the correlations of each of the environmental measures with initial (1983) adaptive behavior. Positive correlations were found with the amount of day programming, the Group Home Management Schedule, the normalization scale, and the Characteristics of the Treatment Environment. Negative correlations were found with the number of other residents and the total number of staff hours per week. The simple correlations are shown at the left of Exhibit 6-3.
The simple correlations in Exhibit 6-3 indicated that some of what the environmental variables were measuring was the functional level of the people living in the environments. This is not a desirable property for a measure of environmental quality. The data specifically showed that higher functioning individuals: (a) were in smaller settings, (b) were in settings that required fewer total staff hours, (c) received more day programming, and (d) lived in less regimented settings where normalization, autonomy and activity were encouraged.
Exhibit 6-3 also presents the results of the partial correlation analysis. When the confounding effect of the relationship between the environmental measures and the adaptive behavior of the people in the setting was removed, four partial correlations between environmental measures and adaptive behavior gain were significant: number of residents (r=-.13, (308), p=.019), Group Home Management Schedule score (r=.12, (308), p=.030), Normalization score r=.12, (308), p=.032), and staff hours per week (r=.20, (308), p=.001).
These partial correlations suggested that, when controlling for differences in initial adaptive behavior scores, the people who tended to make larger gains within the CLAs were those who lived: (a) in smaller CLAs; (b) in more regimented CLAs; (c) in CLAs with higher normalization scores; and (d) in CLAs in which fewer total staff hours per week were expended.
Findings (b) and (d) were puzzling, so both were explored further. Both the Group Home Management Schedule and the total number of staff hours were correlated with the size of the CLA, and possibly both were acting through size to product misleading partial correlations. However, partial correlations of the Group Home Management Schedule and staff hours with growth, controlling for initial adaptive behavior and size, were still significant, and about the same magnitude.
We stress, however, that none of these partial correlations were overwhelmingly large; instead, they indicated significant, but weak, relationships. It should also be noted that these results do not represent a model of growth, since a series of partial correlations was used. These relationships may not lead to the same conclusions as multivariate techniques, and moreover the methods used here assume that the variables are all related in simple linear fashion. The validity of that assumption merits further investigation before drawing final conclusions about the nature of quality in services for people with mental retardation.
These five years of work on measurement of environmental qualities has been intriguing and rewarding, but has not produced any final list of things that "really matter" nor has it resolved all of the problems of measurement in this arena. At the time of this writing, however, the support of our entire behavioral and environmental assessment process has been taken over by the Commonwealth of Pennsylvania as a monitoring system. Hence although the five years of Federal funding are over, work in this crucial area will continue into the foreseeable future.
A few things did emerge that were very clear. One is that when people moved from Pennhurst into CLAs in Pennsylvania under this Federal court order, they went into settings that were much "better" in terms of our measures of normalization and individualization.
In fact, the change was so extreme that our measure of individualization (the RMS) ceased to be of value after people moved to CLAs. Practically all of the CLAs attained the highest possible score on that scale. This implies that institutional environments and small community environments are so different that it may be an error to try to use the same set of of standards for both. That could result in unrealistic demands for large institutional settings and unnecessarily low expectations from community programs. It is possible that we would be wiser to start from scratch in developing standards for community programs, rather than trying to tinker with and adapt the existing institution-oriented standards.
Another is that many so-called "environmental" measures are highly sensitive to the characteristics of the people living in the setting being rated. We hope that this will further impel the effort to develop standards and measures that are independent of the functional level of the people being served. Even our normalization measure, which definitely should be independent of individual functional level, was not. This need for "functional-level-free" measures of environmental quality is similar to the need for measures of individual intelligence that are free of "culture-bias."
We were not able to discern relationships between aspects of one of the most widely used set of program standards in the nation (ACMRDD) and individual growth and development within the institution in this study. That does not mean that a relationship does not exist, and we hope that others will investigate this issue in a rigorous scientific manner. It seems to us extremely important that programmatic standards should be shown to be associated with continual increases in the independent functioning of the people served. These comments apply also to the multitude of other standards and licensing instruments that are used at national, state, and local levels; we urge a great deal more scrutiny of validity (particularly predictive validity vis-a-vis growth).
During the research process, we were constantly reminded that growth is not the only criterion of a good environment. Our measures of Physical Quality and Life Safety, for example, were completely unrelated to people's functional level or their growth. Yet not one of our site reviewers would suggest dropping those measures. There is clearly a place for standards of comfort, safety, and other areas that may have nothing to do with individual development. Of course, they too must be demonstrably reliable.
We constantly tested the reliability of our environmental instruments. One concern involved our Normalization Scale, because preliminary analyses showed that the Normalization Score of a given CLA could change a large amount from one year to the next. Although such phenomena may be genuine, they may also arise from a lack of one or more kinds of reliability. (We must emphasize that this concern about reliability was only about our normalization measure as we applied it with single site reviewers -- our work did not use the full PASS-3 scale.) Our work will continue in this area.
Over the years, our impressions from service providers have led us to the conclusion that skepticism about the reliability of environmental measures and standards is a major problem. It seems understandable that the agencies object to any review in which the result depends on the orientation of the reviewer who is sent out that year. To the extent that they believe that luck is involved, providers will gradually become cynical and resentful. That is certainly not a desirable product of the quality assurance process. We therefore call for far greater attention to the interrater reliability of the existing standards, licensing, and environmental quality measures. Data collection instruments may need to be revised or replaced, and reviewer training may need to be intensified. We see considerable promise in the use of videotaped site reviews to train surveyors and to test their scoring accuracy.
We have continually perceived the quiet presence of a significant question about the entire issue of environmental measures and program standards. In simple terms, that question is to what extent is it feasible to measure qualities of the environment in a brief visit to a residential program? Some researchers have suggested that literally nothing useful can be learned about the quality of a program in less than several weeks of direct presence and observation. Because that is not likely to be practical for large systems of very small community residences, we must continue to face the question of how well we can measure things in brief visits, and whether we can establish that the things measured make a real difference in the lives of the people served.
Our investigation of environmental correlates of growth in community settings led to some provocative, if not conclusive, analyses. The data suggested that, adjusting for initial adaptive behavior, people in smaller settings tended to display more growth. People in more normalized settings tended to display more growth. The analysis suggested, however, that people in settings that were more regimented (as measured by the Group Home Management Schedule) did slightly better. Such a finding, although it could be accurate, is certainly not in line with the general trend of current professional theories in the field. We hope that others will investigate the possibility that a certain amount of structure is necessary and beneficial (although that amount varies according to the level of functioning of the people served), and that below this amount, less growth will occur. The question for scientific study is: how much "restrictiveness" is proper for which kinds of people?
Balla, D. (1976). Relationship of institution to quality of care: A review of the literature. American Journal of Mental Deficiency, 81, 117-124.
Cronbach, L., & Furby, L. (1970). How should we measure change -- Or should we? Psychological Bulleting, 74, 68-80.
Flynn, R.J., & Heal, L.W. (1981). A short form of PASS 3: A study of its structure, interrater reliability, and validity for assessing normalization. Evaluation Review, 5, 357-376.
Jackson, J. (1969). Factors of the treatment environment. Archives of General Psychiatry, 21, 39-45.
McCormick, M., Balla, D., & Zigler, E. (1975). Resident care practices in institutions for retarded persons. American Journal of Mental Deficiency, 80, 1-17.
King, R., Raynes, N., & Tizard, J. (1971). Patterns of residential care: Sociological studies in institutions for handicapped children. London: Routledge and Kegan Paul.
Rotegard, L., Bruininks, R., & Hill, B. (1981). Environmental characteristics of residential facilities for mentally retarded people. Minneapolis: University of Minnesota, Developmental Disabilities Project on Residential Services and Community Adjustment, Project Report #15.
Silverstein, A.B., McClain, R.E., Hubbell, M. & Browniee, L. (1977). Characteristics of the treatment environment: A factor-analytic study. Educational and Psychological Measurement, 37, 367-371.
Wolfenberger, W., & Glenn, L. (1975). Program Analysis of service systems 3: A method for the quantative evaluation of human services. Toronto: National Institute on Mental Retardation.
|
Family Impacts Case Study: A Case of Guilt Susan left Pennhurst in 1974, before the Pennhurst litigation began; therefore by current definition she is not a class member. She was included in the case studies as part of a comparison group of individuals who had been living in the community when the litigation started. Since Susan moved to the community she has lived in 2 different homes. As a result of Susan's ambulation problems, the provider felt it would be better for Susan to live in a ranch-style home. Susan's ambulation problems were caused by chronic phlebitis, which has required hospital care on several occasions. While Susan was at Pennhurst, she had little or no contact with her family. When she moved to the community, however, family visits increased dramatically. She has been visiting with her family in South Carolina at least 3-4 times a year over the past 10 years. Upon examination of records at Pennhurst and at the group home, it was quite clear that Susan's family (aunts and cousins) felt a great deal of guilt around her living at Pennhurst, yet they were unable to care for her themselves. It seems that, to alleviate their own guilt, they have become a doting family. In fact, they are perpetuating the myth of the sick, helpless, eternal child in Susan. Over the past few years gifts to Susan have been gifts that encourage dependence rather than foster independence. One year they gave her a single bed with bed rails. More recently, they gave her an ejection chair so that she wouldn't have to struggle to get up; she pushes a button and is lifted to an upright position. Staff and the provider agency have been unsuccessful in discouraging such gifts. While increased family contact is certainly a desired outcome, it should never occur at the expense of the individuals who are struggling so hard, in many cases, to achieve their independence. Certainly, in this case, Susan is getting mixed messages from her family and from staff and others around her. |
When Judge Broderick ordered Pennhurst closed it was obvious that the decision would have an impact on the lives and attitudes of families, as well as on the people themselves. Therefore, part of the Pennhurst Longitudinal Study has been an assessment of the impact of the court orders on families.
Strangely, no prior study of family reactions to and feelings about alternatives to institutional care included family contact before and after deinstitutionalization. The opportunity to do both was presented by the court decree plus the federal support of the Longitudinal Study. This study therefore became the first to examine changes in family feelings after relocation of a relative. Literature (described below) had already established firmly that most families of people living in institutions were opposed to community placement. The question we were able to address was, if people were placed anyway, how would family behaviors and opinions change?
Families of people in public institutions have been found to be very satisfied with the facilities, and opposed to changes such as community placement. One of the earliest reports of such satisfaction was from Klaber (1969). Surveying parents of people in institutions in Connecticut, he found more than three fourths of them were convinced that the facilities delivered excellent care. Later, Brockmeier (1975) reported similar levels of satisfaction, coupled with skepticism about community-based care, among families of people in Nebraska institutions. In Texas, Payne (1976) discovered the same situation. Overwhelming satisfaction was also reported by Willer, Intagliata, & Atkinson (1979) in New York state. Meyer (1980) found that over 70% of families were satisfied with an institution in Pennsylvania, and were against the idea of community placement. Our own initial findings in the Pennhurst Study were released in 1980, and showed the same pattern. Atthowe & Vitello (1982) detected similar feelings among families in New Jersey. In their survey, 54% expected no more than custodial care, and 91% said the institutional care was adequate or better.
Payne (1976) also identified a "deinstitutional backlash," a loosely knit countermovement of various local and state-wide associations of parents organized in support of institutions as opposed to community living arrangements (CLAs). While many families of people in institutions see group homes or community living arrangements (CLAs) as a viable way to care for some people, most prefer the institution for their own relatives (Atthowe & Vitello, 1982; Frohboese & Sales, 1980; Payne, 1976). Similarly, Ferrara (1979) documented that parents of children with mental retardation were much more supportive of normalization activities for children with mental retardation in the abstract than they were for their own children.
Many families believe that there are individuals with mental retardation who will never be able to achieve the level of independence they think is necessary for community living. Further, many families think it is damaging for professionals to create expectations that their children will achieve such independence (National Association for Retarded Citizens [NARC], 1977).
Families generally believe the decision to institutionalize their relatives was permanent and final. Atthowe and Vitello (1982) found that 84% of families believed that their child would stay institutionalized for life. Stedman (1977) suggested that deinstitutionalization of a relative with mental retardation forces the family to question whether institutionalization had been appropriate in the first place. To those families who institutionalized their children, deinstitutionalization represents a "painful revisitation" of the original decision (Willer et al., 1979).
Families also fear the implications of the concept of least restrictive alternative; they fear that their children will not be protected properly in small community settings (NARC, 1977). As Willer, et al. (1979) said:
In this instance, the individual is moved from a very secure situation where someone else, the state, is responsible for his safety and future. Alternative settings are, by definition, less restrictive, and the family is faced with the belief that increased risk of harm or abuse may occur. (p. 13)
Frohboese and Sales (1980) documented that families believed the state institution to be the least restrictive alternative feasible for their relatives. They perceived greater freedom of movement, independence, and safety within the institution.
Perhaps the greatest concern families have about deinstitutionalization is the permanence of the community settings (Frohboese & Sales, 1980). The question of permanence, in turn, is linked to funding and the duration amount, source, and intent of that funding. An analysis of funding history and current practices reveals that funding for institutions has continued for nearly 100 years, and federal assistance has grown significantly in the past decade. In contrast, funding for CLAs has come primarily from states and/or short-term federal demonstrations. Recent federal funding initiatives for community programs have not yet been tested fully (Braddock, Howes, & Hemp, 1984). For a family whose concern is that their relative be housed, fed, and clothed in the year 2020, institutional funding may appear to be a safer bet than CLA support.
Thus, a reasonably large array of research in many states shows that most families oppose community placement of their institutionalized relatives. The focus of the Pennhurst Longitudinal Study was to test whether attitudes of Pennhurst families fit this pattern, and then to take the unprecedented next step: test for changes among the same families after community placement of their relatives.
Respondents
There were 713 people residing at Pennhurst in May 1980 who originally came from the five southeastern counties of Pennsylvania. Of these residents, 630 had known relatives. Questionnaires were mailed to each of these 630 families for the Baseline Survey. After two mailings and extensive telephone follow-up, responses were received from 472 families (75%). One-fourth of the non-respondents were telephoned and asked a subset of the survey questions; it was determined that the 472 respondents were representative of the population of 630. (That is, the non-respondents did not differ from respondents in their answers to 19 key survey items, as measured by t-tests. Hence the sample was judged to be free of non-respondent bias.)
After the Baseline Survey, we telephoned the families of each of the next 134 people who moved to CLAs; only the families of people who had already been in a CIA for six months were telephoned for the Post-Relocation Survey. The telephone interviews were conducted between January 1981 and February.1984 in four waves.
In the first wave, conducted in early 1981, the 22 families of people who had been in CLAs for six months or more were interviewed. The second wave added 43 more families in mid-1982. In the third wave, in early 1983, there were 54 families of recently placed people, and in the final wave, in late 1983, we spoke with another 15 families. At the end of the Longitudinal Study, then, we had spoken with 134 families of people who had moved to CLAs. In all cases, we had spoken with families who had completed the baseline mail survey form and whose relatives had experienced the CLA setting for at least six, but less than 12, months.
Instruments
Two questionnaires -- one for the baseline survey and one for the post-relocation survey -- were developed for the family impacts study. The aims of these instruments were to assess initial attitudes toward deinstitutionalization, to measure changes in those attitudes after relocation of the relatives, and to identify demographic variables, such as education, sex, and race, which might possibly relate to attitudes.
Barnes, Krochalk, and Hutchinson (1976) conducted a comprehensive community residential care system study that included a mail survey of families/guardians of individuals with mental retardation. Their survey questionnaire assessed characteristics of the person with retardation, services needed to keep the person living at home, positions on philosophical issues, and the types of facilities preferred for out-of-home placement. Although no item from their questionnaire was used in ours, the Barnes, et al. instrument served as a model for development of the first draft of our questionnaire.
The first draft of the baseline survey was prepared in September, 1979. Two national experts in this field were consulted, and their reviews and recommendations for modification were received in December, 1979. At about the same time, the survey was pretested on nine family contacts whose relatives had moved recently from Pennhurst into the community. This group was selected because they had recently been in the same situation as the population of the study but would not be eligible for the before-and-after study. The pilot test provided feedback which led to improvements in the questionnaire.
Additional criticism and feedback was obtained from several psychologists, and a certified advocate from the Office of the Special Master for Pennhurst, and necessary modifications were made. The revised instrument was submitted to the Office of Management and Budget in February, 1980. In March, that agency requested additional changes. The final form of the instrument was approved in April, 1980. This baseline questionnaire is included as Appendix 7-1.
The Post-Relocation questionnaire was designed to measure changes in families' attitudes six months after relocation of their relatives with mental retardation. This post-questionnaire was simply a subset of the items on the baseline questionnaire. We also asked an open-ended question, intended to gather any perceptions, attitudes, or feelings not covered in the survey. This post-relocation questionnaire is included in Appendix 7-2.
The instruments contained many questions that addressed the attitudes of the respondent toward deinstitutionalization. The validity of single items can be questioned, because errors and misinterpretations can bias any particular response. This problem is reduced when many similar items are combined into a scale. Therefore the Attitudes Toward Deinstutionalization Scale (ATDS), a simple additive scale composed of 25 items, was constructed. It was well-structured and internally consistent (Cronbach's Alpha = .94). This scale ranged from 1 to 5; the higher the numerical value of the score, the greater the resistance toward deinstitutionalization. The items contained in this simple additive scale are marked with asterisks in Appendix 7-1.
Procedures
The overall design of the family impacts portion of the Longitudinal Study was pre-post. Families of Pennhurst residents were surveyed by mail in June 1980, before their relatives left the institution. As each resident left Pennhurst, his/her family was interviewed by telephone, approximately six months after the relocation; the six-month delay was intended to permit enough time for each family to develop familiarity with the CLA, and for transitional or temporary relocation phenomena to fade.
The decision to use telephone contact for the post-relocation questionnaire was reached only after careful consideration with government officials, consultants, and the project Advisory Committee. It was possible that the change in methods (pre = mail, post = phone) could influence the results. On the other hand, too small an "N" could call the entire family study into question. Not knowing how many people would actually move, and judging the minimum acceptable prepost sample size to be 100, we chose telephone follow-up because it assured collection of data from virtually 100% of the families of people who moved. By mail, we could only be confident of reaching about 70%. As it turned out, there were 136 families of people who moved and met our criteria for this part of the study. If we had done the post-relocation survey by mail, we might have obtained only 95 completed prepost interviews, rather than the 134 we actually received.
Baseline Study
The central and most striking finding of the baseline family study was the overwhelming opposition of the families to the idea of community placement. When asked the question, "If your relative were to be selected for movement from Pennhurst to the community, how likely would you be to agree with this decision?", the responses were as follows:
| Very likely to agree | 9% |
|---|---|
| Somewhat likely to agree | 5% |
| Unsure | 14% |
| Somewhat unlikely to agree | 9% |
| Very unlikely to agree | 63% |
Thus, 72% of the families of the people still living at Pennhurst in 1980 would have disagreed with any proposal for community placement of their relatives.
In addition, the families were very satisfied with services their relatives were receiving at Pennhurst. In answer to the question, "Overall, how satisfied are you with the services your relative has received from Pennhurst?" the following responses were given:
| Very satisfied | 54% |
|---|---|
| Somewhat satisfied | 29% |
| Neutral | 11% |
| Somewhat dissatisfied | 5% |
| Very dissatisfied | 2% |
Together, these two questions revealed a clear pattern of satisfaction with the institution, coupled with strong opposition to community placement. This was the primary finding of the baseline study.
Attitudes Related to Opposition. We were also interested in some of the reasons for these initial feelings of the families, and in some related opinions. Analysis revealed that families who were older, and whose relatives at Pennhurst were older, were more opposed to community placement. More educated families were more opposed, and white as opposed to non-white families were more opposed.
An attitude that appeared to be related to feelings about community placement was that 75% of families believed, strongly or somewhat strongly, that their relatives had no potential for further educational or psychological development. Moreover, a family's opinion in this area was not related to the relative's adaptive behavior, IQ, or level of retardation. The fact that it was not related suggested that this pessimistic attitude among the families was not necessarily grounded in empirical observation or rational thinking.
Three of the best known philosophical trends in service delivery in the 1970s and 1980s have been "normalization," the "least restrictive alternative," and "deinstitutionalization." Families were asked for their degree of agreement/disagreement with these ideas and the results showed that they were not in accord with these concepts. In fact, 32% agreed (strongly or somewhat) with "normalization," 36% agreed with "least restrictive alternative," and only 19% agreed with "deinstitutionalization." This lack of agreement suggested that there was a general pattern of suspicion and distrust of "new" ideas that might lead to change in the situation of the institutionalized relatives.
There were other opinions that were relevant to opposition to deinstitutionalization. For example, only 15% of families agreed (strongly or somewhat) that funding for community living arrangements was secure and permanent, and 61% disagreed. Permanence seemed to be a central issue for families, and they were clearly not confident of the permanence and security offered by the new CLA concept. Similarly, only 18% of families agreed that all needed services would be available in the community, and only 20% agreed that CLA staff would be sufficiently knowledgeable and skillful to handle all situations that might arise with their relatives.
One of the strongest predictors of a family's opposition to community placement was the family's perception of the intensity of the relative's need for medical care. If the family believed that the relative had great need for attention from doctors or nurses, then that family was likely to oppose community placement. A questionnaire item on medical needs was put into both the family survey and the Behavior Development Survey (BDS), which was our primary instrument for collection of information about individuals:
1 = would not survive without 24-hour medical care
2 = has life-threatening condition that requires very rapid access to medical care
3 = needs visiting nurse and/or regular visits to the doctor
4 = generally has no serious medical needs
In 1978 we collected the BDS for each person living at Pennhurst, from staff, including nurses, and from facility records. We were therefore able to compare the responses from the families to the responses from the facility. The comparison revealed that the facility responses and the family responses did not agree very much at all. Exhibit 7-1 presents the results from both sources.
| EXHIBIT 7-1. Medical Needs as Perceived by Families and by Facility Staff | ||||||
|---|---|---|---|---|---|---|
| Families | Facility Responses | |||||
| High
Need 1 |
2 |
3 |
Low
Need 4 |
Total | ||
| High Need | 1 | 5 | 7 | 43 | **57** | 112 |
| 2 | 2 | 3 | 20 | 21 | 46 | |
| 3 | 0 | 5 | 50 | 62 | 117 | |
| Low Need | 4 | 1 | 2 | 41 | 96 | 140 |
| TOTAL | 8 | 17 | 154 | 236 | 415 | |
The meaning of this table is simply that the families perceived much more intense medical needs among their relatives at Pennhurst than did the staff who were providing direct care. The entry in the table marked by asterisks is the most extreme case of this disparity in perceptions; it represents the fact that there were 57 people about whom the family reported that the person would not survive without 24-hour medical care, but about whom staff reported that there were no serious medical needs.
Post-Relocation Study
Representativeness of the sample. The 134 people in the Post-Relocation study were generally very similar to the average Pennhurst person, except in age. The 134 were five years younger on the average, and were admitted to Pennhurst about five years later than the average. In both adaptive and maladaptive behavior, the 134 people were not significantly different from the average of the Pennhurst population. The distribution of level of retardation labels was about the same for our sample and the population, as well. In both groups, 86% of the people were labeled severely or profoundly retarded.
Characteristics of the 134 family respondents. The 134 family respondents interviewed in the post-relocation study were not very different from the population of 472 families with regard to education, race, sex, and relationship, as Exhibit 7-2 shows.
| EXHIBIT 7-2. Demographic Characteristics of Families | ||
|---|---|---|
| Population of 472 Families |
Sample of 134 |
|
| Education: High school of more | 55% | 51% |
| Race: Non-white | 18% | 16% |
| Sex: Male | 63% | 51% |
| Relationship: Parent | 68% | 72% |
| EXHIBIT 7-3. Initial Attitudes: ATDS Scale |
 |
Furthermore, in analyzing the responses of the 134 family sample to the 25-item scale assessing attitudes toward deinstitutionalization, we found that this group did not differ in initial attitudes from the average Pennhurst family (472), as Exhibit 7-3 shows. This bar graph showed what the statistics also revealed: the 134 families were, to begin with, just as opposed to community placement as was the average Pennhurst family.
Because the 134 families initially were not different from the average Pennhurst family, we believe that the findings about attitude changes contained here are generalizable to the entire set of Pennhurst families. That is, what we have observed for 134 families will probably hold for the remaining hundreds of families.
General Pre-Post Changes
Satisfaction. The 134 families in this study were very satisfied with Pennhurst, but are now just as satisfied with the CLAs. The question we asked was, "Overall, how satisfied are you with the services your relative is receiving from (Pennhurst/the CLA)?" The responses were on a 5-point scale from "very satisfied" (1) to "very dissatisfied" (5).
The average baseline survey response of the 134 (in 1980, while their relatives were still at Pennhurst) was 1.7, which was identical to the average for all 472 families. After movement of the 134 relatives to CLAs, their families gave an average response of 1.5, which indicated that they were just as satisfied with the CLAs as they had been with Pennhurst. This was remarkable because the families had been so opposed to placement, and generally had not expected to be pleased by community services.
Family visits to relative. Families' visits to their relatives hardly changed. Initially, 42% of the 134 reported visiting their relatives at least once a month (similar to the 472, at 44%). After relocation, the figure was 52%, and, though this change was statistically significant, the substantive change was very small. Similarly, 13% of the 134 families reported that their relative came home for a visit at least once a month (much like the 472 at 11%), but this figure changed only 'to 15% after relocation of the relative to a CLA. Thus we found no confirmation of the notion that visits to or from the family would become more frequent upon deinstitutionalization.
Perception of medical needs. We obtained the families' perceptions of their relatives' medical needs from 126 of the 134 families in the pre-post study. In general, families perceived serious medical needs among their relatives in 1980, before relocation, and also after relocation. Families continued, for the most part, to view their relatives as being in need of frequent attention from doctors and nurses. On a scale of I to 4, families averaged 2.8 both before and after relocation.
Changes in Attitudes Toward Deinstitutionalization
Overall change. Our general measure of attitudes was the 25-item ATDS, described previously. This overall scale ranged from I (in favor of deinstitutionalization) to 5 (opposed). The average score of the 134 families before relocation was 3.5; the average score after relocation was 2.4. This change was highly significant (t = 12.94, = (114), p < .001). The families were much more positively disposed toward the complex of concepts related to deinstitutionalization after the relocation of their relatives had taken place.
Changes in particular attitudes. The most direct questions about the idea of community placement were measured on 5-point agreement scales. The questions were:
Baseline: If your relative were to be selected for movement from Pennhurst to the community, how likely would you be to agree with this decision?
Post Relocation: Overall since your relative was selected for movement from Pennhurst to the Community, how do you feel about that move?
From pre to post, the changes were dramatic, as shown in Exhibit 7-4.
| EXHIBIT 7-4. Agreement with Community Placement |
 |
The figure shows visually what the data revealed statistically: these 134 families had drastically changed their positions. Before relocation, 55% of the families were strongly opposed to community placement, and afterward, only 4% were still strongly opposed (the bars at the extreme right of the figure). Conversely, the bars at the left of the figure show that, before placement, only 19% agreed strongly with placement, and afterward, fully 66% strongly agreed.
Treating the same data statistically, as a pair of 5 point scales, the average score of the families before relocation was 3.8, indicating strong opposition. Afterward the average was a very positive 1.7, and the change was highly significant (t = 13.7, (130), p < .001).
Other Changes
There were a number of other areas in which potential changes from before to after relocation were of interest. In Question 13, we asked whether families believed changes would occur in 14 areas of their lives; after relocation, we asked whether changes had occurred in the same 14 areas.
Of the 14 items within Question 13, the 12 that showed significant pre-post changes are shown in Exhibit 7-5. Each item was on a scale ranging from 1 (large change for the better) to 5 (large change for the worse). For the 134 pre, the means indicate expectations; for the 134 post, the means reflect actual changes.
| EXHIBIT 7-5. Expected and Perceived Changes* in Family Life | ||
|---|---|---|
| 134
Pre (Expected) |
134
Post (Actual) |
|
| a. Your own social life | 3.5 | 2.8 |
| b. Your job | 3.5 | 2.9 |
| d. Family recreation activities | 3.4 | 2.8 |
| e. Your time alone | 3.5 | 3.0 |
| f. Your time with your spouse | 3.4 | 2.9 |
| h. Family vacation | 3.5 | 2.9 |
| i. Your general happiness | 3.7 | 2.1 |
| j. Your relative’s relationships with other people | 3.6 | 1.9 |
| k. Your relative’s general happiness | 3.6 | 1.7 |
| l. Your relative’s relationship with you | 3.1 | 2.5 |
| m. Your relative’s relationship with your spouse | 3.3 | 2.7 |
| n. Your relative’s relationship with brothers and sisters | 3.1 | 2.6 |
| * All changes were significant at the 0.001 level. | ||
The initial (expected) responses of the 134 families clustered about 3.5 at baseline, which meant they were basically pessimistic about expected changes. Their expectations were exceeded on each of the items shown in Exhibit 7-5. (The two areas in which changes were not significant were "Your spouse's job" and "Your time with your children living at home.") In many areas, the change from pre to post was from negative expectations to an actual observation of no change (e.g., "Your job" went from 3.5 to 2.9, and 2.9 is essentially no change).
In some areas, however, the differences were from negative expectations to post-relocation reports of distinctly positive observations. For example, the largest change reported by the families was in their relatives' general happiness, followed by changes in their relatives' relationships with other people and in the family respondents' own general happiness.
As a concrete example of the magnitude of these differences between expectations and actual experiences, we present the pre and post data for the Relative's General Happiness in greater detail in Exhibit 7-6. (We select this item because of its special interest for families who want to know whether people are perceived to be happier in community settings.)
| EXHIBIT 7-6. Change in Relative’s General Happiness | ||||||
|---|---|---|---|---|---|---|
| Pre (Expected) |
Post (Actual) |
|||||
| Much
Better 1. |
2. |
3. |
4. |
Much
Worse 5. |
||
| Much Better | 1. | 19 | 0 | 2 | 0 | 0 |
| 2. | 4 | 1 | 0 | 0 | 0 | |
| 3. | 7 | 3 | 5 | 0 | 0 | |
| 4. | 10 | 3 | 2 | 0 | 0 | |
| Much Worse | 5. | 20 | 10 | 11 | 4 | 0 |
Examining the diagonal from upper left to lower right on Exhibit 7-6, we see that there were 25 families (19 + 1 + 5) whose expectations matched their actual experience. For example, the 19 expected a large change for the better in their relatives' general happiness, and then reported seeing exactly that. Above the diagonal are the families whose expectations were disappointed. There were only two who expected a large change for the better, but saw no change. All the other families, (i.e., those below the diagonal), perceived that the happiness of their relatives had improved beyond their expectations. In fact, at the extreme lower left of the table, 20 families expected a large change for the worse, but actually saw a large change for the better.
| EXHIBIT 7-7. Agreement with Specific Ideas | ||
|---|---|---|
| 134
Pre (Expected) |
134
Post* (Actual) |
|
| 14. Relative will not progress beyond present level | 2.0 | 2.9 |
| 17. CLA personnel are knowledgeable and skillful | 3.5 | 1.9 |
| 18. CLA funding is secure | 3.8 | 2.7 |
| 19. All needed services are available in community | 3.8 | 1.8 |
| 20. Community placement does not add to family financial burden | 2.8 | 1.7 |
| 21. Normalization | 3.1 | 1.9 |
| 22. Least restrictive alternative | 2.9 | 1.6 |
| 23. Deinstitutionalization | 3.6 | 1.9 |
| * The significance of the prepost change for the 134 was p < 0.001 by paired t-test for all items. | ||
In another part of the survey (Questions 14 to 23), we posed a series of 10 specific statements concerning deinstitutionalization, and asked for responses from 1 (strongly agree) to 5 (strongly disagree). All items changed in a positive direction, and 8 were statistically significant. These 8 are presented in Exhibit 7-7.
These results followed one general pattern: the 134 families became significantly more positive about each area after their relatives moved to CLAs (note that on Question 14, agreement implied a negative attitude).
Two items in Exhibit 7-7 were of particular interest. Item 14 was important because it concerned the developmental model, (i.e., the belief that all people can grow and learn). This concept is one of the cornerstones of the new ideology in mental retardation services. The families initially tended toward rejection of the developmental model, and at post-test changed only to neutrality. Both in the institution and the community, then, it appeared that families were not responsive to this relatively new philosophy.
The second item of special interest from Exhibit 7-7, Item 18, concerned the security of CLA funding, a very important issue for families. The families initially tended to disagree somewhat that funding for CLAs was secure and permanent. After relocation, the 134 families changed their opinion, but only to approximate neutrality. Their anxieties on this issue were reduced, but by no means eliminated.
Qualitative Results
At the conclusion of the structured interview, we asked an open-ended question: "Is there anything else you would like us to know about your relative's recent move from Pennhurst?" Interviewers were instructed to take comments verbatim, and not to ask additional questions.
Upon analysis of these responses, the predominant tone indicated that the majority of the respondents expressed very positive feelings about the CLAs and the quality of service therein. A significant majority had not expected such services and were quite overwhelmed by the superior quality of the facilities. The general feeling was that the relatives had shown progress toward development of skills for independent living. Many respondents attributed this growth to the personalized attention and interest of the staff, which was greatly facilitated by the small size of the facility and a high staff-to-client ratio.
The respondents also reported that they enjoyed their visits to the CLAs. They had found visits to Pennhurst "scary" and were intimidated by converging crowds of other people who lived there. Other respondents felt the CLA setting was conducive to bringing younger siblings for visits. Previously, parents had not wanted to expose other children to the large, hospital-like, impersonal environment of Pennhurst. In addition, most respondents indicated that their relatives appeared happier at the CLAs. They enjoyed the small family and home-like environment and individual attention.
Though the general tone indicated a positive attitude toward community living, there were some objections to the move from Pennhurst. Some respondents felt it was "not safe" and "rather dangerous" for "these people" to "walk around alone." The implication was that persons with mental retardation need to be protected from the "normal" world; that they should not be free to walk around, since they are vulnerable. Given the level of functioning of these former Pennhurst residents (86% were labeled severely or profoundly retarded, and nearly half were non-verbal), this belief was understandable. One respondent opposed the move because CLAs did not have the advantage of having all the necessary facilities (medical, recreational, educational) on the premises and another respondent objected on the grounds that the court decision to move the relative to a CLA was a violation of parental rights.
In addition, there were many expressions of concern about the security of funding for the CLAs from both the respondents who approved the move and those who did not. A number of respondents feared they might have to assume financial responsibilities for which they had no resources. Also, there was some apprehension about the effect of staff turnover. One respondent felt that the staff could not possibly be permanent since they would want to "live their own lives,” and feared that this would be emotionally damaging to his/her relative.
The retrospective evaluation of Pennhurst from these 134 families was that it was too large and crowded a place to offer adequate care and growth opportunities for the people who lived there. It was felt to be a place where repetitive, institutional behaviors prevailed due not only to the large numbers of persons housed, but also to the chronic shortage of direct care and professional staff.
Although the CLAs were seen as addressing the needs of the clients more favorably than Pennhurst, there was strong concern that they would not have the permanence of a large institution like Pennhurst. This appeared to us to be the central counterpoint to the general extreme satisfaction expressed in the open-ended comments, and this paralleled the quantitative results of the survey.
The most striking result of the family survey was the overwhelmingly positive change in attitudes among the families of the people who left Pennhurst and went to live in community based settings. Also of significant interest were the attitudes which did not change. In this discussion, we will comment on both.
Before proceeding, it is important to stress the caveat that the Pennhurst results did not arise from families of individuals deinstitutionalized at random. We cannot be certain that the "sample" of 134 families were representative of the "population" of 630 families in every way, although we found that they were so in nearly every way we could measure. Although cautions against perfectly confident generalization to the population, or to facilities in other states, must be applied to the Pennhurst results, we believe that some general policy implications can be drawn.
The attitudes expressed by the families in the baseline study were consistent with the results obtained by Atthowe and Vitello (1982), Brockmeier (1975), Klaber (1969), Meyer (1980), Payne (1976), and Willer, et al. (1979). A large number of the families in our study disagreed strongly with deinstitutionalization, and a substantial number disagreed strongly with the principles of normalization and least restrictive alternative. The families in our study seemed to agree with families in other studies that the institution was the most appropriate environment for their relatives, and were generally very satisfied with Pennhurst. Our findings brought to mind Klaber's (1969) comments:
The parents…were convinced of the excellence of the facilities in which their children were placed…The praise lavished on the institutions was so extravagant as to suggest severe distortions of reality in this area.
Most of the families in the baseline study (75%) also believed that their relatives had reached the highest level of development possible. Evidently, families did not accept the idea that everyone, even persons with severe or profound mental retardation, can grow and develop (e.g., Gold, 1973). The families in the Baseline Study also believed their relatives had serious medical needs (although this belief was not confirmed by comparison to reports from Pennhurst staff).
These two attitudes, pessimism about future development and perception of serious medical needs, may be related. They could both have arisen from the advice given to the families, decades ago, by professionals. When the families in this study admitted their relatives to Pennhurst, an average of more than 20 years ago, there were no alternatives, and the professional with the most authority in these matters was usually a physician. The message most often given, we speculate, was on the order of 'IS/He will never be able to learn, and will always need medical care." These attitudes are not eradicated even today. Among both families and professionals, we would suggest vigorous efforts to provide the most up to date information; particularly for families of people who may move out of institutions. Any educational interventions should stress the idea that people can grow and learn, and should address medical concerns directly.
Families in the baseline study also were greatly concerned about the security of funding for community placements. They did not believe the funding for community alternatives was secure and permanent -- unlike the funding for institutions -- and felt they could not depend on the community service system to provide services for their relatives.
Six months after the relocation of their relatives, the 134 families in the Post-Relocation Study were more than satisfied with the community placement of their relatives. Many expressed astonishment over their own changes since the baseline survey. The families reported unexpected changes for the better in their lives, and in the lives of their relatives, especially with regard to the happiness of their relatives, their relatives' relationships with other people, and their own happiness. In addition, families' fears about the quality of CLA staff and about the availability of services in the community seem to have been allayed somewhat, although by no means complete y. An examination of the beliefs and attitudes which did not change is also revealing. For example, even after the relocation, families still believed their relatives had serious medical needs. Families also did not completely accept the developmental model (i.e., the idea that their relative would continue to grow and learn). Attitudes became more positive, but only to the point of neutrality, not to outright acceptance. In light of the increase the families had seen in their relatives' developmental abilities -- especially as reported in the open-ended comments -- this finding was difficult to interpret. In essence, most families were saying that they had recently witnessed large improvements, but they doubted that more was possible.
Families also expressed continued concern about the security of funding for community services. Their belief that funding was secure improved significantly, but only to the point of approximate neutrality. Their anxieties were reduced, but not eliminated. This appeared to us to be the central remaining issue among these families, who otherwise were generally surprised and pleased by the change from care in a large segregated public institution to small, more integrated settings in regular neighborhoods.
No matter how much families may prefer the services in the community, if they believe those services can be cut off for lack of funding next year -- or, more important, in ten or fifteen years -- families will not support community living. We believe this issue must be addressed on a federal level since the federal government has provided more fiscal support to institutions than to community service systems (Braddock, Howes, & Hemp, 1984). This has engendered a disincentive for states to develop community services.
Two final implications of our five years of research with families do not arise directly from data, but rather from years of impressions. The first is that any deinstitutionalization plan or effort should provide a formal forum through which families can express their feelings, especially their fears and their reasons for opposition. Although this need not guarantee that families have the power to veto community placement, impressions from formal family hearings in the Pennhurst arena strongly imply that many, perhaps most, families will become willing to "give it a try" after a formal and structured hearing designed to treat their concerns with dignity.
The second is that, in our experience with surveying these families, we have come to the conclusion that any monitoring or quality assurance system should include annual surveys of families. Particularly when conducted by a third party, such surveys can reveal information that families would not express otherwise. Many dissatisfactions go untold because families fear that state, county, or private providers will resent such statements and that the consequences might fall on the relatives. Our surveys in this area have been welcomed by families, they are very inexpensive to conduct, and they can help to raise red flags that would not reach official attention in any other way. In years to come, our surveys of families will continue as part of our permanent monitoring of Pennhurst class members (and others in the Commonwealth).
Atthowe, J.M., Jr., & Vitello, S.J. (1982). Deinstitutionalization: Family reaction and involvement. Unpublished manuscript, College of Medicine and Dentistry of New Jersey, Rutgers Medical School.
Barnes, D., Krochalk, P., & Hutchison, J. (1976). Comprehensive Community Residential Care System Study. Los Angeles: Exceptional Children’s Foundation.
Braddock, D., Howes, R., & Hemp, R. (1984). A summary of mental retardation and developmental disabilities expenditures in the United States: FY 1977-1984 (Preliminary working data). Chicago: University of Illinois at Chicago, Institute for the Study of Developmental Disabilities.
Brockmeier, W.E. (1975). Attitudes and opinions of relatives of institutionalized mentally retarded individuals toward institutional and non-institutional care and training. Lincoln: University of Nebraska, doctoral dissertation. Dissertation Abstracts International, 35, 5163A.
Conroy, J.W., Feinstein, C.S., Lemanowicz, J.A., & Kopatsis, M. In press.
Ferrara, D.M. (1979). Attitudes of parents of mentally retarded children toward normalization activities. American Journal of Mental Deficiency, 84, 145-51.
Frohboese, R., &. Sales, B.D. (1980). Parental opposition to deinstitutionalization. Law and Human Behavior, 4, 1-87.
Gold, M.W. (1973). Research on the vocational habilitation of the retarded: The present, the future. In N.R. Ellis (Ed.), International review of research in mental retardation, Volume VI. New York: Academic Press.
Klaber, M.M. (1969). The retarded and institutions for the retarded -- A preliminary research report. In S.B. Sarason and J. Doris (Eds.), Psychological problems in mental deficiency. New York: Harper & Row.
Meyer, R.J. (1980). Attitudes of parents of institutionalized mentally retarded individuals toward deinstitutionalization. American Journal of Mental Deficiency, 85(2), 184-187.
National Association for Retarded Citizens (1977). The parent/professional partnership: The partnership -- How to make it work. New York: NARC.
Payne, J.E. (1976). The deinstitutional backlash. Mental Retardation, 3, 43-45.
Stedman, D.J. (1977). Introduction. In J.L. Paul, D.J. Stedman, & G. Neugeld (Eds.), Deinstitutionalization: Program policy developments. Syracuse, NY: Syracuse University Press.
Willer, B.W., Intagliata, J.C., & Atkinson, A.C. (1979). Crisis for families of mentally retarded persons including the crisis of deinstitutionalization. British Journal of Mental Subnormality, Vol. 2, p. 2.
When a group home or community living arrangement (CLA) for people with mental retardation opens in a neighborhood, how do neighbors react? How many even know about it, and do their attitudes toward people with mental retardation change in any way? How do these attitudes compare to feelings about people with other kinds of differences? We have been exploring these questions in southeastern Pennsylvania for five years, as part of the Pennhurst Longitudinal Study, and some of the most interesting findings are reported in this chapter.
Attitudes about unfamiliar groups of people are generally characterized as stereotypes. As noted by Triandis (1971) and others, the strength of the stereotype is inversely related to knowledge about the group. The more one knows about a person or group, the less likely one is to develop stereotypes about them.
In most of this century, the practice of segregated institutional care has meant that people with severe or profound mental retardation rarely have been seen in public places. Despite the fact that institutional populations have been declining since 1967 (Lakin, 1979), and despite the literature and experience that demonstrate that people with severe degrees of retardation can live and grow in less segregated community settings (e.g., Bradley & Conroy, 1983), the public's knowledge about mental retardation is limited (Budoff, Siperstein, & Conant, 1979; Gottwald, 1970; Hollinger & Jones, 1970; Latimer, 1970). Therefore it is likely that public attitudes toward people with mental retardation are based on stereotypes. The question of whether these public attitudes can change, then, should be viewed from the perspective of theories on stereotypical attitudes.
Budoff et al. also implied an important function of attitudes, called the knowledge function. The knowledge function is related to the nee s o persons to maintain an-organized, stable, and meaningful structure of the world. These attitudes change when the existing attitude is insufficient for dealing with situations, whether because of new information or because of a new environment of some kind. This viewpoint is related to the situation of a group home opening in a neighborhood, in that stereotypes may prove to be of little value when a citizen directly encounters a new neighbor with mental retardation.
Researchers have assumed that knowledge about and contact with people with mental retardation affects attitudes toward such people, and that change of either knowledge or contact would change those attitudes. Much of the research on attitude change has been done in school settings rather than in the community (i.e., structured rather than unstructured contact), and previous studies of attitudes vary in their conclusions. Most of the studies indicating a relationship between contact and attitudes were concerned with contact that was structured in some way. Begab (1969), for example, found that increasing knowledge about retardation only effected a positive change in attitudes when education was coupled with direct contact with people with mental retardation.
There have been few studies which examined attitude change in communities where community living arrangements opened. Baker, Seltzer, and Seltzer (1974), and Sandler and Robinson (1981) examined the effects of preparing communities for the opening of a residence and suggested that preparation is likely to raise opposition, perhaps to the point of preventing the opening of the home. Mamula and Newman (1973) and O'Connor (1976) reported that, after initial opposition, communities tended to accept the residence. If opposition prevents homes from opening, however, there can be no opportunity for attitudes to become more positive. Sigelman (1976) suggested that a Machiavellian approach (i.e., establishment of homes without informing neighbors and without measuring attitudes) "has the advantage of preventing moves to block the home's opening" and that such an approach "may be no less effective in the long run than more elaborate strategies involving advance attitude sampling" (p. 26).
Because community acceptance has been portrayed in the media as a crucial issue in the opening and success of group homes and other community programs, the Pennhurst Longitudinal Study included an examination of the attitudes of members of the communities into which the people from Pennhurst moved. The situation offered the opportunity to conduct the first "before-and-after" interview study of neighbors. There was no preparation of neighborhoods; the only "intervention" was the actual opening of the community living arrangement (CLA).
The plan called for the assessment of neighbor attitudes toward people with mental retardation before and after CLAs opened in nearby houses or apartments. In addition, an exploration of the factors related to attitudes, and of factors related to changes in attitudes over time, was planned. The general research questions were: (1) What were the patterns of attitudes among the general public toward people with mental retardation living in their neighborhoods? (2) What factors and characteristics were associated with those attitudes, that is, were some kinds of neighbors more accepting than others? (3) Were there changes in attitudes after the CLAs entered the neighborhoods? and (4) Would there be any consistent pattern to, or predictors of, changes in attitudes?
Subjects
The locations of eight prospective CLA sites were obtained, and a one quarter to one half mile radius (depending on population density) was drawn around each. In each circle, 45 households were selected by a simple random selection procedure from crisscross telephone directories. A probabilistic, representative sample of the adults in those households was accomplished a procedure developed by Kish (1965). A table determined which household member was to be the interview subject, based on the number and kind of potentially eligible respondents in the household. In this table, the one selected would be varied from one interview to the next. No substitution was allowed. In this way, we achieved samples in each neighborhood that were close approximations to simple random samples (i.e., every person within the neighborhood had about the same chance of being interviewed).
There were 362 neighbors who were interviewed in the initial round. Their average age was 48 years. They were 87% white and 54% female, and 80% had a high school education or more. They had lived at their current addresses for an average of 16 years.
Design
This study of neighbor attitudes was designed to be the first to assess attitudes before and after opening of a CLA nearby. Initially, we intended to interview neighbors six months before the CLA entered the neighborhood, and then again six months after the opening. After completing that design, however, the research team determined to interview the neighbors again about a year later. The reason was that we had detected significant changes in attitudes at six months after CLA opening, and others (Mamula and Newman, 1973; O'Connor, 1976) had suggested, but had not quantitatively demonstrated, that such short term changes would vanish by about a year to a year and a half.
The national advisory committee for the study agreed that this was a worthwhile design modification, as did the government project officers. We therefore conducted a total of three waves of interviews with the original sample of neighbors.
Instruments
Because we did not want to inform respondents that a CLA was about to open in their neighborhoods, we could not ask the most direct questions -- such as "How do you feel about the group home that's going to open on your block next month?" Such questions would have destroyed the integrity of the study by giving information to many neighbors who otherwise might not have had it. More importantly, it could have engendered active opposition as suggested by Sigelman (1976). It was necessary to aim instead for general attitudes about people with mental retardation. No completely suitable instrument was found in the literature so a new instrument was developed.
We began by assembling 350 items from a dozen previously used scales. About two thirds of the items were immediately revised or rewritten to remove archaic language or to suit the conditions in Pennsylvania. We also wrote about 50 new items for our specific pre-post needs, then sorted all 400 into categories: (1) tolerance toward people with mental retardation in everyday settings; (2) knowledge about mental retardation; (3) general attitudes toward people with mental retardation; (4) frequency and locations of contact; and (5) fears and stereotypes. The research group then began to eliminate items within each category, keeping only the ones that appeared to be the most clear and concise ways to ask about each content area. After several cycles of review, we settled on a draft set of 50 items.
The draft interview contained many questions about attitudes toward people with mental retardation. We added items about people with physical disabilities, people of a different race, and people with mental illness. Again, these other questions were added to prevent respondents from coming away with the impression that they had been interviewed solely about mental retardation, or about new CLAs. We wished to avoid alerting the neighbors because of the possibility of resistance and because of experimenter effects. By alerting neighbors in a way that normally would not occur, we possibly could have altered their natural pattern of response to the eventual opening of the CLA. The extra questions also enabled comparisons of attitudes among the various groups, to give some idea of the magnitude and direction of the attitudes.
Following selection and refinement of items, the instrument was pretested. One of the most significant results of the pretest was the labeling of all questions about mental retardation with "mild" or "severe." This was done because nearly three-fourths of pretest respondents said, on at least one question, "It depends on how severe . . ." or a similar qualifier.
Following final review, the questionnaire contained 46 substantive items about various groups, and 34 concerned people with mental retardation. The instrument was submitted to the federal office of Management and Budget, and was approved by May 1980. At this final stage, it was designed to take approximately 15 minutes to administer. No mention of Pennhurst or the prospective CLA was contained in the interview.
After the baseline interviews, the form was shortened somewhat, and a few new items were added. After CLA opening, it was of interest to ask people whether they knew of any such programs in their neighborhoods, and if so, how long they had been in existence. The questionnaire is in Appendix 8-1
In order to provide a sensitive and reliable measure of general attitudes toward people with mental retardation, a scale was constructed from questionnaire items (ATTSCALE). All items were weighted equally, and a simple additive scale was constructed. Item selection was based on Cronbach's Alpha, a measure of one kind of reliability called internal consistency. By removing five questionnaire items (of the 18 chosen initially as candidates for a general attitudes scale on face validity grounds), Cronbach's Alpha attained a value of .78. This was a very acceptable value because Alpha provides a conservative estimate of reliability.
In addition, the interview included nine True-False items that were designed to be combined into a single scale of knowledge about mental retardation. It was intended to permit analysis of variations in attitudes, and in attitude change, according to how much people understood about mental retardation.
Procedures
Data collection was initiated with an introductory letter, followed by telephone screening and interviewing. The first interviews were conducted six months prior to the opening of the CLAs, in May-June 1980, and yielded 364 interviews. The second wave (an average of six months post opening) was conducted in the Spring of 1981, and yielded 287 interviews, 79% of the original sample. The third wave was conducted in the Spring of 1982, an average of about 20 months after CLA opening, and produced 252 interviews, 69% of the original sample.
Respondents who moved out of the boundaries of the CLA sample circle were not reinterviewed, and in one sample site the CLA did not open. The 34 sets of interviews from that neighborhood have been included only in the baseline results. The final data set, on which most of the results presented here are based, consisted of the remaining 218 respondents from whom all three interviews were obtained.
Specific Attitude Items
Respondents were asked how much they would be bothered if two to five people with mental retardation moved into the neighborhood. As a comparison, and to avoid sensitizing respondents, respondents were also asked how much they would be bothered if members of other groups moved into the neighborhood. The results are presented in Exhibit 8-1.
| EXHIBIT 8-1. Would Neighbors be “Bothered”? | ||||
|---|---|---|---|---|
| Question: How much would it bother you if 2 to 5 people who are […GROUP…] moved into your neighborhood? Would if bother you a lot, some, very little, or not at all? | ||||
| Group | A Lot | Some | Little | Not at All |
| Physically disabled | 2.5% | 27.5% | 11.9% | 80.6% |
| Mildly mentally retarded | 6.1 | 9.7 | 15.5 | 68.7 |
| Severely mentally retarded | 14.0 | 16.8 | 18.5 | 50.7 |
| Mentally ill | 16.4 | 25.6 | 15.8 | 42.1 |
| Of a different race from your own | 4.7 | 12.2 | 16.3 | 66.9 |
On this question, responses concerning people with mild mental retardation were most like those about people of a different race. Responses about people with severe mental retardation were most like those about people with mental illness. Also, responses showed sharply different levels of "bother" for mild versus severe mental retardation.
Respondents also were asked how much they thought the value of their property would be affected if two to five members of the same groups moved into the neighborhood. These results are presented in Exhibit 8-2.
| EXHIBIT 8-2. Neighbors’ Beliefs About Effects on Property Values | ||||
|---|---|---|---|---|
| Question: How much do you think the value of your house would change if 2 to 5 people who are […GROUP…] moved into your neighborhood? | ||||
| Group | A Lot | Some | Little | Not at All |
| Physically disabled | 7.2% | 12.6% | 15.9% | 64.3% |
| Mildly mentally retarded | 7.3 | 13.6 | 18.2 | 60.9 |
| Severely mentally retarded | 13.3 | 17.5 | 19.0 | 50.2 |
| Mentally ill | 15.5 | 21.0 | 18.6 | 44.8 |
| Of a different race from your own | 12.4 | 25.6 | 14.4 | 47.6 |
On the property values question, again, feelings about people with mild mental retardation were less intense than feelings about neighbors with severe mental retardation. This time, however, responses about people with mild retardation were most like those about people with a physical disability; responses about people with severe mental retardation were similar to those for people with mental illness and people of a different race.
| EXHIBIT 8-3. Attitudes Toward Different Groups |
 |
Comparison of the "bother" question to the property values question revealed an intriguing pattern. Exhibit 8-3 on the next page is structured to show the comparison as a bar graph, in which each bar represents the average value of the baseline responses to both questions.
The labels for the 5 groups in the figure have these meanings:
PHYS = People with Physical Disabilities
MMR = People with Mild Mental Retardation
RACE = People of a different Race than the respondent
SMR = People with Severe Mental Retardation
MI People with Mental Illness
In answer to either question (would you be bothered or would property values be affected), respondents could say:
[1 = Not at all] [2 = A little] [3 = Some] [4 = A lot]
These responses were treated as 4 point scales. The bar graph shows that, on the "bother" dimension, respondents said they would be bothered very little by new neighbors with physical disabilities, with mild mental retardation, or of a different race. They would be bothered much more by neighbors with severe mental retardation or mental illness.
The pattern for property value effects was different. Relatively 'mild' effects were projected for people with physical disabilities and for people with mild mental retardation. "Major" effects were projected for people of a different race, people with severe mental retardation, and people with mental illness. (Dividing the responses into minor and major effects is based on t-tests of differences among the mean scores for the five categories. The bars for property values in Exhibit 8-3 for PHYS and MMR were statistically indistinguishable from one another; the bars for RACE, SMR, and MI were also indistinguishable from one another. However the first two were statistically smaller than the latter three.)
An interesting facet of the bar graph data was that respondents were quite consistent for SMR and MI (they would be bothered considerably and property values would be affected considerably), they were fairly consistent for PHYS and MMR (they would not be bothered much, and property values would only be affected slightly), but they were not consistent for RACE. Here the respondents claimed that they would be bothered very little by new neighbors of a different race, yet they projected major property value impacts.
Factors Related to General Attitudes
The next step in analysis of the baseline data was an investigation of the factors that were related to the general attitudes of the neighbors toward people with mental retardation. As previously noted, we had composed such a scale by combining 13 items (ATTSCALE). The factors that might be related to this scale of general attitudes fell into two categories: unchangeable factors, such as the age and sex of the respondent, and changeable factors, or things that could conceivably be changed by social policy or experiences, such as knowledge about mental retardation or contact with people who have mental retardation.
Unchangeable factors. The baseline survey produced an array of demographic and descriptive data on the characteristics of each respondent. These characteristics were tested for relationship to general attitudes toward people with mental retardation via simple Pearson correlations and analyses of variance. The characteristics examined were age, sex, ethnicity (white/other), education, income, type of dwelling, length of time at this address, length of time in neighborhood, marital status, and number of children in the household. The three characteristics which were related significantly to attitudes (ATTSCALE) were age (r=.18, p=.002), ethnicity (binary variable, r=.13, p=.023), and sex (binary variable, r=.10, p=.038). Using these three variables together in multiple regression, age explained 7.8% of the variation in ATTSCALE, followed by ethnicity (3.4%) and sex (1.0%). Younger respondents, non-white respondents, and females had more positive attitudes. Altogether, these three variables explained 12.2% of the variation in our measure of attitudes toward people with mental retardation.
Changeable factors. The survey included nine true/false items about people with mental retardation. A scale of 0 to 9, based simply on the number of correct answers, was calculated for each respondent and was used as a general index of knowledge about retardation, people with retardation, and programs/services. Knowledge was found to be a strong predictor of attitudes (r=.46, p<.001). We ran a multiple regression in which the three unchangeable factors above were forced to enter the equation first, 'and together they accounted for 12.2% of the variation in the general attitudes measured by ATTSCALE; after them, knowledge accounted for an additional 18.2% of the variation.
There were two measures of contact with persons with mental retardation -- one measured overall frequency of contact, and the other item measured contact in given settings (e.g., school, work, neighborhood, shopping). Among the contact variables, only "contact in neighborhood" (a simple Yes or No item) predicted attitudes (r=.30, p<.001). Again using multiple regression, and entering the unchangeable characteristics of the respondents first, contact accounted for an additional 2.9% of the variance in ATTSCALE.
Both of these findings about the changeable variables suggested that general attitudes of citizens toward people with mental retardation were subject to change, either by increasing knowledge or by increasing contact.
Changes in Attitudes
We divided the sample into two groups: (1) those who knew, at the time of the second interview six months after CLA opening (Time 2), that the CLA had opened, and (2) those who did not know, at Time 2, that the CLA had opened. The analyses of interest concerned only those who were aware of the existence of the CLA.
Only 28% of the respondents in our sample were aware that a CLA had moved into their neighborhoods by Time 2.
Among these neighbors who were aware of the CLA, attitudes became significantly less positive from six months before to six months after, and then became significantly more positive again by 20 months after CLA opening. Attitudes were not significantly different between Time 1 and Time 3, which indicated simply that general attitudes had returned to their original levels after a temporary negative swing. This pattern is illustrated in Exhibit 8-4 on the next page. Statistically, the average ATTSCALE score six months be ore was higher than the average score six months after (paired t=1.80, (56), p<.001). The ATTSCALE scores six months after were significantly lower than at 20 months after (t=2.06, (56), p=.044). Scores at six months before and 20 months after were statistically indistinguishable.
Even by the time of the second interview, at six months after CLA opening, there was evidence that neighbors' attitudes were moving in a positive direction. Six months after opening, we asked the people who were aware of the CLA how they had felt when they had first heard about it, as well as how they felt about it at the time of the interview. The results are presented graphically in Exhibit 8-5 on the next page. Looking at the two bars on the right of the figure, the open bar represents the 7% of respondents who recalled being strongly in favor of the CLA when they first learned of its existence. The shaded bar represents the 20% who were strongly in favor "now." The entire figure shows the pattern of change -- there is a shift from the left to the right. Treating the responses as 5 point scales produces the same result from the statistical perspective. The average response "when first learned about CLA" was 3.09, and the average "now" was 3.64 (t=5.28, (132), p<.001).
| EXHIBIT 8-4. Neighbor Attitudes Before and After |
 |
| EXHIBIT 8-5. Self-Reported Attitude Changes |
 |
In sum, the respondents clearly recalled having been more opposed to having a CLA in their neighborhoods when they first heard about it than they were at the time of our second interview. This suggested that neighbor attitudes were already becoming more positive by the time of our six month post-opening interviews (and this was part of the rationale for the decision to conduct a third wave of interviews). This also suggested that, at one or two or three months after CLA opening, neighbor attitudes may have been considerably more negative than indicated by our six month results.
Contact in the Neighborhood
We explored whether contact with people with mental retardation in the neighborhood increased after the CLAs opened, and whether increased contact was associated with positive changes in attitudes. None of the contact variables showed a statistically significant increase, overall, for the sample of neighbors in this study. For the respondents who had said "No" in the first interview to "In the past six months, have you had any personal contact ... in the neighborhood," and who said "Yes" at 20 months after opening, we could discern no significant changes in general attitudes. This group of respondents followed the general pattern of temporary negative reactions followed by a return to the original level.
Real versus Hypothetical CLAs
We examined whether attitudes toward real CLAs were different from attitudes toward hypothetical CLAs. At the time of the third interview, about 20 months after the CLAs had opened, 68 of our respondents said they were aware of the CLAs' existence and 144 said they knew of no CLA in their neighborhoods (a few others did not answer this question). For the 68 who were aware, we then asked "How do you feel about that group home now?" and for the unaware 144, we asked "Imagine that a group home were located in your neighborhood...How would you feel?" These questions enabled us to compare the attitudes of people toward real CLAs to attitudes toward imaginary CLAs. The results are presented as the bar graph of Exhibit 8-6 on the next page. The striking aspect of this figure is the large difference on the right side; relatively few people (about 10%) were willing to express strong support for imaginary CLAs, perhaps partly because they were not sure what a CLA was or how it might affect the neighborhood. In contrast, the respondents who were aware of real CLAs that had been in their neighborhoods for an average of 20 months were more definite; 26% expressed strong support.
This study of neighbor attitudes was the first to interview neighbors before and after group homes/CLAs opened in their neighborhoods. We were most interested in the question of changes over time, that is, whether neighbors would become more accepting or rejecting toward people with mental retardation after a CLA moved into the community. We also wanted to find out whether any such changes were short term or long term. In addition, some important subsidiary issues included the relation between contact and attitudes and the differences between some peoples' feelings about imaginary CLAs in their communities and other peoples' feelings about real CLAs. Our findings on these issues may prove useful to policy makers and planners, as well as to those who are involved more directly with implementing and operating small community based residences.
In general, members of communities do seem to accept their neighbors with mental retardation. Six months after the opening of the CLAs in our study, only about 28% of neighbors were aware of it. The attitudes of these neighbors became less positive from pre-opening to six months post-opening, but, by 20 months post-opening, attitudes had returned to baseline levels. In other words, for the neighbors who were aware, negative reactions were only temporary.
| EXHIBIT 8-6. Feelings About Real and Imaginary CLA |
 |
From our baseline data, we could conclude that the neighborhoods which would be more accepting are those where the population is younger and includes more non-white people and more females. However, these three variables combined only predicted 12.2% of the variance in our attitudes scale, and, therefore, we would not recommend that they be given much weight in choosing sites for CLAs.
The responses to the questions about how much citizens would be "bothered" by various groups of new neighbors, and about how such neighbors might affect property values, are interesting. For people with mild mental retardation, respondents reported that they would be "bothered" very little, and they expected no major impacts on property values. For people with severe mental retardation, however, the responses were quite different. Possibly this was because of a public perception that severely retarded people are immediately recognizable in a neighborhood, and might therefore have a far greater impact on property values than people with mild retardation. Citizens on the average believed that small groups of people with severe mental retardation, of a different race, or with mental illness, all posed about the same level of threat to property values, and the threat was considerable. About the first and third groups, citizens were even willing to admit that they themselves would be bothered by such people; for people of a different race, however, respondents appeared to be saying "I don't mind, but most people do, so property values would be affected." The possibility of public hypocrisy in this regard should not be overlooked.
The questions about real and imaginary CLAs reveal that, for the average citizen, the idea of a CLA in the neighborhood is more threatening than the reality. In our study, people who were aware of CLAs were significantly more positive than were people who did not know of any CLAs in their neighborhoods.
The data on imaginary and real CLAs also suggests to us that, if there are no CLAs already in a neighborhood, then only about 10% to 20% of neighbors will be opposed to allowing one to open. If one is already in the neighborhood, then opposition will be even less, and strong support may be available. This finding might be related to a common tendency for community residences to accumulate in a given neighborhood -- the "CLA ghetto" phenomenon. The more there are, the easier it is to open a new one -- up to a certain point, at least.
The 10% to 20% figure implies that only a small proportion of citizens would be opposed to a CLA -- unless something is done to raise opposition to the CLA, especially before the CLA opens. One possible scenario -- one that Sigelman (1976) has found -- is that even a small number of community members who are strongly opposed can influence general community sentiment, to the extent that the CLA does not open. Sigelman reported that "Although only two neighbors initially opposed the [proposed hostel] program, community sentiment reversed due to the efforts of one intense critic, to the point that almost all of the people who had originally accepted the proposal signed a petition against it" (p. 28). Similarly, one of the eight sites we chose for our study of community attitudes did not open because of community opposition. It is worth repeating that our respondents who knew of the CLA reported feeling more favorable toward the CLA over time; but neighbors can not grow to accept or welcome CLAs if the CLAs are prevented from opening.
It also seems that unstructured contact with people with mental retardation has little measurable effect on attitudes. "Contact in the neighborhood," the only significant contact variable at baseline, predicted only 2.9% of the variance in attitudes beyond that predicted by the characteristics of the respondents themselves. Post-opening, an increase in contact in the neighborhood had no effect on attitudes, regardless of whether the respondent was aware of the CLA. Given the possible problems with giving community members "advance notice" (i.e., time to raise opposition and prevent the opening of the CLA), and given the possible positive effects of structured contact reported by other researchers (e.g., Aloia, Beaver, & Pettus, 1978; Ballard, Corman, Gottlieb, & Kaufman, 1977; Leyser & Gottlieb, 1980; Marlowe, 1979; Voeltz, 1980), structured contact, or structured contact plus education, after the CIA has opened, may influence attitudes favorably without allowing prevention of the opening of the CLA. We suggest this as an area for future research, and such research need not be exclusively quantitative -- for example, the case studies and anecdotal reports of Robert and Martha Perske (Perske & Perske, 1980) are also of tremendous value.
Our results suggest, to us, that future research should include further examination of the effects of structured contact with the CLA residents after the CLA has opened. We also interpret our results as supporting Sigelman's notion of a Machiavellian approach to the opening of CLAs. Neighbors do seem to become more favorable over time, and, as previously noted, attitudes cannot improve if the CLA never opens.
Aloia, G.F., Beaver, R.J., & Pettus, W.F. (1978). Increasing initial interactions among integrated EMR students and their nonretarded peers in a game-playing situation. American Journal of Mental Deficiency, 82, 573-579.
Baker, B.L., Seltzer, G.B., & Seltzer, M.M. (1974). As close as possible. Boston: Little, Brown, & Co.
Ballard, M., Corman, L., Gottlieb, J., & Kaufman, M.J. (1977). Improving the social status of mainstreamed retarded children. Journal of Educational Psychology, 69, 605-611.
Begab, M.J. (1969). The effect of differences in curricula and experiences on social work student attitudes and knowledge about mental retardation. Dissertation Abstracts, 29, 4111-4112.
Bradley, V.J., & Conroy, J.W. (1983). Third year comprehensive report of the Pennhurst Longitudinal Study. Philadelphia: Temple University Developmental Disabilities Center. [http://aspe.hhs.gov/daltcp/reports/3yrpenn.htm]
Budoff, M., Siperstein, G., & Conant, S. (1979). Childrens knowledge of mental retardation. Education and Training of the Mentally Retarded, 14, 277-281.
Gottwald, H. (1970). Public awareness about mental retardation. Arlington, VA: The Council for Exceptional Children.
Hollinger, C.S., & Jones, R.L. (1970). Community attitudes toward slow learners and mental retardates: Whats in a name? Mental Retardation, 8, 19-23.
Kish, L. (1965). Survey sampling. New York: John Wiley & Sons.
Lakin, C. (1978). Demographic studies of residential facilities for the mentally retarded: An historical overview of methodologies and findings, Project Report #3. Minneapolis: Developmental Disabilities Project on Residential Services and Community Adjustment, University of Minnesota.
Latimer, R. (1970). Current attitudes toward the mentally retarded. Mental Retardation, 8, 30-32.
Leyser, V., & Gottlieb, J. (1980). Improving the social status of rejected pupils. Exceptional Children, 46, 459-461.
Mamula, R.A., & Newman, N. (1973). Community placement of the mentally retarded: A handbook for community agencies and social work practitioners. Springfield, IL: Charles C. Thomas.
Marlowe, M. (1979). The games analysis intervention: A procedure to increase the peer acceptance and social adjustment of a retarded child. Education and Training of the Mentally Retarded, 14, 262-268.
O’Connor, G. (1976). Home is a good place: A national perspective of community residential facilities for developmentally disabled persons. Washington, DC: American Association on Mental Deficiency.
Perske, R., & Perske, M. (1980). New life in the neighborhood: How persons with retardation or other disabilities can help make a good community better. Nashville: Parthenon Press.
Sandler, A., & Robinson, R. (1981). Public attitudes and community acceptance of mentally retarded persons: A review. Education and Training of the Mentally Retarded, April, 97-103.
Sigelman, C.K. (1976). A Machiavelli for planners: Community attitudes and selection of a group home site. Mental Retardation, 14, 26-29.
Triandis, H.C. (1971). Attitude formation and change. New York: Wiley.
Voeltz, L.M. (1980). Children’s attitudes toward handicapped peers. American Journal of Mental Deficiency, 84, 455-464.
The purpose of this component of the Longitudinal Study was to identify and compare the costs of residential, day program, case management, habilitation, behavioral, medical, and other services in an institutional and in a community setting, and to explain the differences in program costs within and across these settings. This study area was intended to surmount many of the limitations found in previous studies by employing a more comprehensive and rigorous design, including more exacting cost-finding procedures in conjunction with more refined program performance measures (i.e., measures of program outputs and outcomes).
A limitation inherent in the study was that the unit costs of the Pennhurst Mental Retardation Center programs and of the community-based residential and day programs were not necessarily representative of state mental retardation center costs and community-based program costs in other parts of the state let alone other parts of the country. In fact, the institutional and community-based day programs serving Pennhurst class members were generally on the "resource rich" as opposed to "resource poor" end of the spectrum, and had at least the potential of directing more resources to their clients than did programs in most other parts of the state and country.
The Center for Residential and Community Services at the University of Minnesota, based on its 1982 mail survey of 279 state institutions and mental retardation units in public mental hospitals, reported an average facility per them nationwide in fiscal year 1981-82 of $92.85 or $33,890 per annum; in Pennsylvania, the average per them was reported to be $107.64 or $39,289 per annum. The per them rate at the Pennhurst Center was somewhat higher at $123.00 per day or $44,899 per year.
The mean cost per client of the sample (N=73) of community living arrangements (CLAs) serving Pennhurst class members in fiscal year 1981-82 was $33,237. This is higher than the mean per client costs of all community living arrangements in the Southeast Region of $22,951, much higher than the mean per capita costs of all Pennsylvania CLAs of $17,856, and a great deal higher than the estimated average for small community residential facilities in other parts of the country, $14,242.
The mean cost per capita of the sample (N=16) of community sheltered workshops and work activity centers serving Pennhurst class members was $7,800 in fiscal year 1980-81. This was more than twice the state average of $3,386 per year. The mean cost per capita of the sample (N=3) of adult day programs serving Pennhurst class members was even higher at $9,644 per year. The estimated cost per capita of work activity centers and adult day care programs in the United States was an estimated $3,928 in Fiscal Year 1980-81.
Given this limitation, the more generalizable and important findings may be those explaining differences in program costs rather than those indicating the absolute magnitude of these differences.
The cost study was confined to those programs at the Pennhurst Mental Retardation Center and in the five-county Pennhurst Service area (Bucks, Chester, Delaware, Montgomery, and Philadelphia) serving members of the plaintiff class. Plaintiff class members included persons at the Pennhurst Center or on the Pennhurst waiting list at the time the lawsuit was brought to the Center, and all those persons at risk of being admitted to the Center. The cost data covered the period July 1, 1981 -- through June 30, 1982.
The types of programs costed include residential, adult day (i.e., sheltered workshops, work activity centers, pre-vocational and adult day care), case management services, and specialized support services (specifically behavioral therapy, speech and hearing, and occupational therapy). Residential programs were narrowly defined to include services provided by residential care staff (including nurses) and their immediate supervisors. Some smaller community residential providers may have had only a single program, larger residential providers may have had a number of different programs. Each program consisted of one or more sites (apartments or home) and served clients reasonably homegneneous in terms of adaptive skills.
The study encompassed 42 living areas (residential programs) at the Pennhurst Center, and 102 residential programs in the community. The study covered four adult day programs at the Pennhurst Center and 35 adult day programs in the community. The study covered the specialized support programs in three of the five counties (n = 3) and at the Pennhurst Center (n = 1). The study covered the case management program at the Pennhurst Center (n = 1), the case management programs at each of the base service units in the five counties (n = 5), and the case management teams mandated by the court to serve members of the plaintiff class in each of the five counties (n = 5). Finally, the study examined the medical program and transportation program at the Pennhurst Center and the few medical programs and transportation programs in the surrounding communities for which cost information could be obtained.
Most of the program cost data, output data, and organizational data were obtained through reports on file at the Pennhurst Center and in each of the five surrounding county mental health and mental retardation (MH/MR) offices). Information on program cost data, output data, and other variables in part explaining variations in program costs were obtained in structured interviews conducted by the principal investigators with a purposive sample of 82 direct care staff at the Pennhurst Center, 17 community living arrangement staff, and 20 day program staff, with the mental retardation coordinator, case management supervisor, and Core team director in each of five counties, and with dozens of Pennhurst, county and state administrative staff.
Three types of unit cost measures were employed in comparing program costs: (1) cost per client day, (2) cost per hour of of direct care staff time, and (3) cost per hour of selected developmental services. The comparative analysis of the residential and day programs was structured to match (control for) the different types of programs in the community and in the Pennhurst settings in terms of the scope of services provided and types of clients served. The programs were classified into two groups according to the mean age of their clients: (1) adults age 22 and over, and (2) children age 21 and under. Each of these groups was divided into four subgroups according to mean adaptive behavior scores of their clients: (1) 0-25, (2) 26-50, (3) 51-80, and (4) 81-128. A ninth group included persons of all ages and adaptive levels but with overriding medical needs.
The client data necessary to make these classifications were obtained by Temple University using the Behavior Development Survey (BDS). The BDS is a short version of the AAMD Adaptive Behavior Scale. It was developed by the UCLA Neuropsychiatric Research Group at Lanterman State Hospital by selecting the most reliable and valid behavioral items from the full scale. The Temple Evaluation and Research Group extended the instrument by adding items covering client characteristics, family relationships, medical status, the individual habilitation plan, program goals, and type and amount of services delivered.
The community residential programs were also classified by facility type (i.e., those housing four to eight clients in group homes, those housing three or fewer clients in apartments, those housing three or fewer clients in small homes, and those where more or less than 24 hour supervision was required).
Explanations for differences found between the unit costs of programs at the Pennhurst Center and the unit costs of programs in the community were examined in terms of relative prices paid for program resources (e.g., staff, facility, etc.), level of resources employed (e.g., staff/client ratios), mix of resources employed (e.g., with or without nurses), and organizational variables (e.g., hours of service provided clients, the level of staff assistance provided clients, program size, and client mix). Regression analyses were used to explain the relative power of variables such as those cited above to explain unit cost differences among programs within Pennhurst and within the community.
The cost finding procedure was designed to capture as fully as possible the resources expended directly and indirectly, in the course of delivering services to clients. Generally, the total operating costs reported for programs were augmented by the costs of goods and services of benefit to the clients but not appearing on the books such as state and county charges for general administrative support, for the amortization of capital improvement costs, for insurance, and so forth. Deducted were those costs for goods and services of no perceptible benefit to clients such as research and development expenses, an litigation fees. Next were isolated those costs traceable directly to the residential day, and other client programs using the reports of expenditure by activity. Then the cost of the indirect activities were allocated among these residential, day, and other client programs in proportion to the direct costs of the programs in order to arrive at "loaded" costs for these programs.
Residential Programs
Comparison of residential program costs per client day at the Pennhurst Center and in the community. On average, community programs were found to cost less per client day than Pennhurst Center programs serving groups of like clients in terms of adaptive behavior, age, and medical need. Also, the community programs showed a much larger range in cost per client day, $19.64 to $252.66, than did the Pennhurst programs, $99.74 to $208.94.
Minimally supervised apartments were the least costly type of arrangement on average at $54.64 per day. Group home programs cost a little more per client day at $59.80 . Apartment programs and small home programs showed the highest average cost per client day at $74.84 and $121.93 respectively. In all cases, programs for children cost more, on average, per client day than like programs for adults.
Comparison of residential program costs per hour of direct service worker and supervisor time at the Pennhurst Center and in the community. Measured in terms of the cost per hour of direct staff time, a measure more indicative of staff level of effort than the cost per client day measure, the economic advantage of community residential programs over the Pennhurst Center programs increased dramatically. While the average community-based residential program cost 70% as much as the average Pennhurst program in terms of cost per client day, the average community-based residential program cost only 40% as much as the average Pennhurst program when measured in terms of cost per hour of direct staff time.
Comparison of residential program costs per hour of selected development services and nursing services at the Pennhurst Center and in the community. Along with nursing care, the following types of developmental services were selected to form this measure: cognitive and academic training, mobility training, sensorimotor training, social interaction training, recreation therapy, family life/sex education, community living skills training, dressing skills training, eating skills development, hygiene and grooming, supervised recreation, and supervised recreational trips. The average cost per hour of selected developmental/nursing service in the community residential programs was 42-93% of that in the Pennhurst residential programs, depending on the average adaptive skill levels of the program clients.
Factors explaining differences in the cost per client day in residential programs. Client age was the only client variable shown to have any significant predictive value (p=.05) in terms of the cost per client day in Pennhurst residential programs. It explained 40.4% of the variation in the cost per client day with programs serving older clients having higher per diems.
Client factors that explain differences in the cost per client day among residential programs set in the community. Unlike the Pennhurst Center, in the community programs, age was one of the two client variables not shown to have statistical significance as a predictor (p=.05) of the cost per client day -- the other was client maladaptive behavior scores. Alone, the mean adaptive behavior scores of residents accounted for 13% of the variation in cost per client day, and medical need accounted for 3.0% of the variation. Together, as part of the overall regression equation, these variables accounted for 23.8% of the variation in the cost per client day of community residential programs.
The limited ability to predict program cost per client day based on these client variables may indicate that to some extent clients are fit to program models as much if not more than program models are fit to clients. Some programs may be structured to provide intensive service, while others may be geared to provide less intensive service for the same types of clients. There is some support among our findings for this line of reasoning. The correlation was found to be negligible between the mean number of hours of service provided to clients per week and the mean adaptive behavior scores of these clients (r=0,03, n=155), mean maladaptive behavior scores of these clients (r=0.02, n=155), and mean medical needs of these clients (r=0.17, n=155). Similarly, there appeared to be no significant relationship between the level of assistance staff provided to program clients and the adaptive behaviors (r=0.08, n=16), maladaptive behaviors (r-0.18, n=16), and level of medical need (r=-.01, n=16) of these clients. Even the hours of direct staff time per client showed no significant relationship to the hours of developmental/nursing service that clients actually received.
Organizational factors that explain differences in the cost per client day among residential programs in the community. Six organizational variables were entered into the final regression. Surprisingly, differences in the average salaries paid to residential program workers and supervisors bore no significant relationship to the program cost per client day. In fact, the correlation between these variables was negative (r=-.031) (n=47).
In contrast, the "number of direct care staff per client" variable alone explained 47.6% of the variation in the program cost per client day. For each additional full-time direct care staff member per client, the added per them cost was $32.54 in Fiscal Year 1981-82. This is to be expected as direct staff costs amount to such a high percent of community residential programs. Together with the other organizational variables, this variable accounted for 62.6% of the variation in program costs per client day (f=.0000).
The remaining four organizational variables alone accounted for 32% of the variation in the cost per client day of community residential programs (f=0.001). The most statistically significant of these last four organizational variables as a predictor of the cost per client day of community residential programs was the type of facility (small home, apartments, group homes) which explained 10.2% of the variation in community residential program per diems.
The second organizational variable shown to be a statistically significant predictor of the cost per client day was the number of clients served by the program (i.e., program size). This variable was negatively correlated with program per diems and accounted for 11.9% of their variation. Interestingly, the findings, though not at all conclusive, suggest that given a particular type of facility, the greater the number of residents, the lower the cost; however, the marginal cost advantage associated with a larger number of residents may not hold beyond 10.
The generally higher cost per client day in smaller programs (one to four residents) likely reflects non-economies of scale. Non-economies of scale occur when residential programs are so small that fractions of inputs, particularly staff time and facilities, cannot be secured. It is difficult for example to rent one third of an apartment, or to hire a staff person at less than half time. On the other hand, to rent excess space, or to hire a full-time staff person where only a part-time staff person is required, leaves the program with excess capacity. Of course, too few staff or too little space are not programmatically acceptable alternatives.
“Whether the program was in its initial year of operation,” was the third statistically significant predictor accounting for 13.8% of the variation in community residential program per diems. At the start, the residential program sponsors must expend resources to organize, staff, equip and supply the residential facilities before arrangements can be made for clients to move in. These one-time costs and delays in reaching full occupancy combine to explain why the per client costs of new programs are extraordinarily high. Not surprisingly then, programs in their first year of operation reported costs per client significantly higher than the other programs.
According to this latter regression equation: in Fiscal Year 1981-82:
Programs housed in small homes could be expected to cost $21.91 per day more than the average community residential program.
For each additional resident, the program cost per client day could be expected to decrease by $3.36
A program in its first year of operation could be expected to cost an average of $8.73 more per day than programs in existence for more than one year.
The size of the provider measured in terms of the total number of residential slots across all programs, was not found to be a statistically significant predictor of the program costs per client day.
Factors that explain differences between the costs per client day of residential programs set at the Pennhurst Center and set in the community. There are a number of apparent explanations for the finding that the costs per client day of residential programs, on average, were higher in Fiscal Year 1981-82 at the Pennhurst Center than the costs per client day of comparable residential programs in the community. These explanations relate to the relative prices paid for program resources, to the levels of resources employed, to the types of resources employed, to the types of clients served, and to other less tangible differences.
Relative prices of resources. Personnel service costs include the costs of staff salaries and wages, fringe benefits, staff development, and other miscellaneous costs relating to personnel. Approximately 78% of the residential and day program expenditures at the Pennhurst Center, and 0 (n=122) of the expenditures of residential programs in the community were devoted to personnel services.
Pennhurst Center residential workers, super-visors, and nurses were paid an average of 30% more than their community counterparts. Fringe benefits including paid absences (e.g., holiday, vacation, sick leave, disability leave, and personal leave) for staff at the Pennhurst Center amounted to 63% of base salaries. Fringe benefits including paid absenc6s in community programs amounted to only 36.1% of base salaries, less than half the rate of Pennhurst employees.
If the salaries and fringe benefit levels of direct service staff positions in the Pennhurst residential programs were reduced to the salary and fringe benefit levels of the same positions in the community residential programs, the mean resident per them at Pennhurst would have been reduced by more than 9% from $128.08 to $116.22. Assuming that the average difference in the salaries of community and Pennhurst program staff of 30% is also reflective of the difference in the average salaries of community and Pennhurst support staff in Fiscal Year 1981-82, the mean resident per them would have been reduced by about 27% or approximately $35 per day to $91.00 per day -- a per them rate equivalent to that averaged in the community residential programs.
Level of resources employed. Measured in terms of the hours of direct staff time per client, community residential programs assigned slightly more direct staff resources on average to each client than did Pennhurst residential programs in Fiscal Year 1981-82. Community residential programs spent an average of 1,902 hours of direct staff time per client, 84 hours more than the 1,753 hours of direct staff time per client at the Pennhurst Center. This accounted for some of the difference in the average cost per hour of direct staff time in the Pennhurst residential programs and community residential programs.
Mix of resources employed. There are several explanations for the differences in program costs related to how resources are employed in providing services to clients at the Pennhurst Center and in the community, specifically:
The use of in-house versus out-of-house resources -- specialization of labor--In the community residential programs, the residential workers were expected not only to supervise and train residents, but oftentimes to act as the resident's guardian helping them to manage their personal funds and related affairs, to buy and prepare food, to help clean the residence, to do the laundry, and to help administer client medications. At the Pennhurst Center, the residential workers were not expected to perform these additional functions; they were accomplished by other specialists.
In private industry the specialization of function such as occurs at the Pennhurst Center can be cost advantageous. Specialized workers often demand less pay than workers who are more broadly skilled, and their rate of production can markedly increase as they become proficient in their area of specialization.
However, the opposite seemed to be the case in the delivery of residential services. Specialists at Pennhurst demanded higher salaries and fringe benefits than the generalists in the community residential programs, and the community program workers were able to make more productive use of their time through job expansion rather than job specialization.
Many of the same goods and services (e.g., recreation, security, library, religion) that were produced by the Pennhurst Center were bought by or on behalf of residents in community programs or were publicly available at no charge. The cost advantage here was to the community programs given that the costs of these goods and services was largely fixed, likely to be comparable to the costs at the Pennhurst Center and could be spread over a greater number of persons in the community than at the Pennhurst Center.
The medical model versus development model of care -- Simply put, the medical model tends to view the relationship between staff and residents in terms of doctor and patient and emphasizes diagnoses and prognoses, and treatment services. While the Pennhurst Center in Fiscal Year 1981-82 had moved in favor of the development model, the staffing pattern was still a vestige of the medical model with a significant complement of higher-paid doctors, nurses, and other medical support staff relative to community programs.
Day Programs
The per diem cost of the work activity center and sheltered workshop programs at the Pennhurst Center ($30.05) was only slightly more than the mean per them costs of work activity center and sheltered workshop programs set in the community ($27.99). The mean cost of the community adult day care programs, $37.75 per day, was equivalent to the cost of the adult day care programs at the Pennhurst Center, $37.29 per day.
The average cost per hour of direct staff time in the Pennhurst Center day programs was $19.48, 80% less than the average community day program cost of $24.54. For sheltered workshop and work activity center programs, the cost per hour of direct service worker and supervisor time at the Pennhurst Center, $19.77, was less than 70% of that of the community sheltered workshop and work activity center programs, $28.78. Conversely, in the case of adult day care programs, the cost per hour of direct service worker and supervisor time in the community ($9.73) was just over 50% of that at the Pennhurst Center ($18.72).
The average salary of Pennhurst day program service workers and supervisors in Fiscal Year 1981-82 was 28% higher than the salary enjoyed by their counterparts in the community programs (n=24), and the average fringe benefits rate (including paid absences) was double that of the community day programs (63% versus 36.1%). In addition, the community-based work activity centers/sheltered workshop programs were able to provide 532 (n-21) direct staff hours per client versus 340 at the Pennhurst work activity center/sheltered workshop program.
However, this difference was not manifest in the relative costs per hour of direct staff time. The cost per hour of direct staff time in community work activity centers/ sheltered workshops was $28.78, over 45% higher than the $19.77 cost per hour of direct staff time at the Pennhurst work activity center/sheltered workshop program. This appears to be due to the fact that unlike the Pennhurst Center program, many of the community work activity centers and sheltered workshops supported drivers, food service workers, business development, and other support staff. Indirect staff salaries and wages amounted to 36.5% of direct staff salaries and wages in the community work activity center and sheltered workshop programs; whereas, in the Pennhurst work activity center and sheltered workshop programs, indirect staff salaries and wages amounted to only 18.6% of direct staff salaries and wages.
Habilitative and Behavior Management Programs
The manner in which these services were made available differed at the Pennhurst Center and in the surrounding communi