U.S. Department of Health and Human ServicesU.S. Department of Health and Human Services
This report was prepared under contract #HHS-100-95-0046 between the U.S. Department of Health and Human Services (HHS), Office of Disability, Aging and Long-Term Care Policy (DALTCP) and the University of Maryland. For additional information about the study, you may visit the DALTCP home page at http://aspe.hhs.gov/_/office_specific/daltcp.cfm or contact the ASPE Project Officer, Pamela Doty, at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, SW, Washington, DC 20201. Her e-mail address is: Pamela.Doty@hhs.gov.
This report has benefited greatly from the thought-provoking comments of several people. In particular, we thank external reviewers Richard Browdie (Benjamin Rose), Nancy Eustis (Hubert H. Humphrey Institute of Public Affairs) and Robyn Stone (Institute for the Future of Aging Services). Several members of the Cash and Counseling Demonstration and Evaluation management team--Kevin Mahoney, Lori Simon-Rusinowitz, Marie Squillace, and members of the staff of the Centers for Medicare & Medicaid Services (CMS)--provided useful and insightful comments. We also appreciate comments and suggestions from the Florida Consumer Directed Care Plus program in the Department of Elder Affairs and the Agency for Health Care Administration.
The report would not have been possible without the contributions of several colleagues at Mathematica Policy Research, Inc. Nora Paxton, Licia Gaber, and Amy Zambrowski programmed the analysis, and Valerie Cheh provided thoughtful comments on an earlier draft. Patricia Ciaccio edited the report, and Jill Miller produced it.
The opinions presented here are those of the authors and do not necessarily reflect those of the funders (the Robert Wood Johnson Foundation and the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation), the Cash and Counseling National Program Office, CMS, or the demonstration states.
Consumer Direction of Medicaid Supportive Services. Roughly 1.4 million people with disabilities receive Medicaid-funded, noninstitutional supportive services each year. Home care agencies provide many of these services: under professional supervision, agency workers help beneficiaries with bathing, meal preparation, light housework, and other basic activities. "Consumer-directed care," in which Medicaid beneficiaries hire, train, supervise, and pay workers of their choice, is an alternative to the professional service model. Consumer direction increases beneficiaries' autonomy and control, but it also increases their responsibilities.
Cash and Counseling is a model of consumer-directed care that offers eligible Medicaid beneficiaries the opportunity to receive a monthly allowance to hire workers, including family members, and purchase other disability-related services and goods. Adult consumers can designate a representative, such as a family member or friend, to help them manage their care. Cash and Counseling also offers counseling and fiscal services to consumers and representatives. New Jersey, along with Arkansas and Florida, has tested the Cash and Counseling model as part of a three-state demonstration. Mathematica Policy Research, Inc. is the demonstration evaluator.
In New Jersey, the demonstration was open to adult Medicaid beneficiaries who were: (1) using state plan personal care assistance (PCA) or had been assessed as eligible for it, (2) not also participating in home and community-based waiver programs or a state-funded consumer-directed program, and (3) expected to require PCA for at least six months. The evaluation randomly assigned demonstration enrollees to participate in New Jersey's Personal Preference program (the treatment group) or to use PCA as usual (the control group).
Goals of This Report. This report describes the implementation of Personal Preference by synthesizing information from in-person discussions with program staff, a mail survey of program consultants, telephone interviews with consumers in the treatment group, and program records. It discusses the program's goals and features, the ways consumers managed their program responsibilities and took advantage of increased flexibility, and the degree to which consumers were satisfied with the program. (Other reports from the evaluation estimate the program's impacts on consumers, their caregivers, and public costs; describe the types of beneficiaries and workers that chose to participate in the demonstrations; and explain demonstration implementation and program operations in greater detail.)
The Personal Preference Intervention. The Personal Preference allowance was based on the value of beneficiaries' Medicaid PCA plans. At enrollment, consumers were eligible for monthly allowances of $1,062, on average. To receive the allowance, consumers or their representatives had to develop a written cash management plan that met the approval of the Personal Preference program. Consultants helped consumers develop their plans and monitored consumers' well-being. They were also available to advise consumers about recruiting workers and accessing community services. The fiscal agent was available to write checks for goods and services purchased with the allowance and to process payroll taxes and employment forms for consumers who hired workers. The program did not charge consumers directly for consulting services, but consumers did pay for some of the fiscal services they used. (The program paid for others.)
Industry Support. Personal Preference garnered the cooperation of personal care agencies, which it relied upon to identify prospective demonstration enrollees and the hours of care planned for them before and during its demonstration. The industry viewed consumer direction as inevitable and even beneficial for some people with disabilities. Moreover, the program director responded to industry concerns by, for example, discouraging consumers from hiring workers away from agencies.
Outreach and Enrollment. New Jersey initially planned to recruit 2,000 beneficiaries into its demonstration in 12 months, but it actually recruited 1,755 beneficiaries in 32 months (November 1999 to July 2002). To boost enrollment midway through the demonstration, Personal Preference made two major changes to its approach to outreach and enrollment. It originally delegated outreach and enrollment activities to a private, for-profit firm with which the state had an existing Medicaid contract. The enrollment contractor was to invite eligible beneficiaries to join the demonstration when they were assessed or semiannually reassessed for PCA. When enrollment rates consistently fell short of expectations and costs consistently exceeded them, the program hired state employees to conduct outreach and enrollment activities. It also separated the timing of enrollment from that of assessment.
Enrollment did not increase much after these changes, but the changes demonstrate the pros and cons of alternative approaches. The key advantage of having an existing contractor conduct outreach and enrollment was expediency--it took less start-up time than beginning contract procurement anew or recruiting and hiring new state employees. The advantage of linking outreach with PCA assessments was that the care plans developed from beneficiaries' assessments provided an up-to-date basis for calculating the allowances consumers would receive under Personal Preference. Conversely, the key advantage of hiring state employees was that program staff had more control over outreach procedures and could experiment with them. As long as enrollment rates lagged, the key advantage of separating outreach from assessment was that it enlarged the pool of potential enrollees that outreach workers could pursue at any time.
Consumer Characteristics. Despite its difficulties, New Jersey eventually recruited a fairly diverse population for its demonstration. The evaluation randomly assigned 871 beneficiaries to the treatment group--404 nonelderly adults and 467 elderly ones. Slightly more than half of these consumers were White, and slightly more than one-third were Hispanic. About four in ten consumers had graduated from high school.
Planning for, and Using, the Allowance. Six months after being assigned to Personal Preference, slightly more than half of all consumers had received the program allowance, and nearly one-quarter were still enrolled but had not received it. (Three percent of consumers were deceased at this time, and the other fifth had disenrolled from the program.) Getting started on the allowance was subject to many procedural delays. The program eventually reduced three sources of delay by: (1) assigning consumers to consulting agencies instead of offering them a choice of agencies; (2) consolidating caseloads across a smaller number of agencies; and (3) cutting, from 30 to 14 days, the notice it gave to personal care agencies to discontinue services to beneficiaries assigned to Personal Preference. Still, because of the number of people involved in the development, review, and approval of cash management plans, getting started on the allowance took a long time. Consumers and representatives, consultants, program staff, and fiscal agent staff all played roles.
Consumers who received the Personal Preference allowance took advantage of the opportunity to use it flexibly. Among those who were receiving the allowance at the time of the nine-month follow-up interview, 80 percent said they used the allowance to hire one or more workers. Nearly 75 percent of these consumers hired family members, and about 40 percent hired friends or neighbors. Most workers helped consumers with household and community tasks, personal care, and routine health care, and many provided assistance with transportation.
According to program records, consumers used about 80 percent of their monthly allowance to pay workers. Roughly five in ten consumers received up to 10 percent of the allowance as cash for incidental purchases, of types specified in their cash management plans. Slightly fewer than one in ten consumers used the allowance to buy assistive equipment during the month observed for this analysis.
Recruiting Workers. Recruiting workers was difficult for some consumers. One-quarter of all consumers said they tried to hire but could not. Nearly 30 percent of those who did hire said it was difficult, often because of a lack of interested or qualified candidates. Some consultants said they were uncertain about how much recruiting assistance the program expected them to provide to consumers, especially those who did not have a family member they wished to hire.
Consulting and Fiscal Services. Consultants reported that their most time-consuming Personal Preference duties were helping consumers develop cash management plans, performing administrative tasks, and advising consumers about payroll-related activities. Consultants believed their services were of value to consumers, and most consumers confirmed that consultants provided useful help.
All allowance recipients used the program's fiscal services--availability of these services seemed to be an important part of consumers' successful management of their fiscal responsibilities. Moreover, the program relied on the fiscal agent to prevent misuse of the allowance by double-checking the accuracy of consumers' cash management plans and verifying that check requests matched those plans.
Consumer Satisfaction. Nine months after being assigned to the Personal Preference program, 91 percent of consumers said they would "recommend the program to others who wanted more control over their personal care services." Among consumers who received the allowance, 82 percent said it had improved their life greatly or somewhat. Consumers who used their allowance to pay workers were uniformly satisfied with how workers performed their tasks and with their relationships with workers. Elderly and nonelderly consumers were equally satisfied with most aspects of paid workers' reliability, promptness, and disposition.
Disenrollment. Despite high levels of satisfaction, 22 percent of consumers chose to leave the Personal Preference program within a year of enrolling. Although there was no single overriding reason for voluntary disenrollment, consumers most commonly said they disenrolled because they believed it was or would be difficult to assume the responsibilities of an employer (34 percent) or changed their minds and were satisfied with their usual PCA services (30 percent). In addition, some consultants reported that some consumers enrolled in the program without fully understanding consumer direction, then disenrolled after they learned more about it. Nearly three-quarters of consumers who disenrolled or died did so without having received the program allowance.
Experiences of Different Types of Consumers. Multivariate models used to assess the experiences of different types of treatment group consumers suggested that, all else being equal, consumers who considered it very important, at baseline, to be able to pay family or friends for caregiving were more likely than other consumers to receive the monthly allowance and stay in the program. Treatment group consumers who lived alone were less likely than others to receive the monthly allowance and stay in the program. Hispanic treatment group consumers were less likely than non-Hispanic ones to receive the allowance and stay in the program, and black consumers were less likely than white consumers to receive the allowance. Among consumers who hired or tried to hire workers, those who were elderly were less likely than those who were not to say hiring was difficult. Age was not otherwise associated with program experiences.
Some policymakers have concerns about consumer direction of public funds. These include: (1) whether consumer direction should be available to all users of supportive services, (2) whether to allow family members to be paid for caregiving, (3) how to ensure consumer safety, (4) how to prevent the exploitation of workers, and (5) how to prevent the misuse of public funds. Personal Preference procedures addressed each of these concerns to some extent.
Assessing Suitability for Consumer Direction. New Jersey's policy was to not screen prospective enrollees on their suitability for consumer direction. Rather, the policy was to inform them of their responsibilities and rights under the program and let them decide whether to enroll and whether to select a representative. Consumers received PCA services as usual until they began receiving their program allowance, and they could disenroll from Personal Preference at any time and revert to usual services. Thus, Medicaid beneficiaries could try consumer direction without incurring great risk. A multivariate analysis suggested that New Jersey's decision to open the demonstration to all groups--including elderly adults, consumers with cognitive impairment, and those in need of large amounts of PCA--was sound.
Paying Family Members. While policymakers debate using public funds to pay family members, New Jersey allowed Personal Preference consumers to hire family members, including legally responsible spouses. The option to hire relatives probably was critical to the functioning of the program. Nearly three-quarters of consumers who hired workers hired family members (although only 2 percent hired a spouse). Although some consumers (27 percent) hired workers who were not family members, the proportion that did so successfully was considerably smaller than the proportion that tried.
Ensuring Consumer Safety. There was no evidence from consumers, consultants, or program staff that participation in Personal Preference led to any adverse effects on consumers' health and safety. Personal Preference monitored consumer safety and care quality primarily through consultants' contacts with consumers and representatives, which occurred by telephone and in consumers' homes. Moreover, while there was very little evidence or suspicion of consumer neglect or exploitation in Personal Preference, procedures existed for consultants and program staff to follow up if anything seemed amiss.
Preventing the Exploitation of Workers. Although Personal Preference workers had no formal mechanism to report grievances, worker abuse did not emerge as a serious problem in the program. More than half the consumers who used the allowance to pay workers, including family members, signed work agreements with them. Few Personal Preference consumers provided fringe benefits to their workers. Nearly all the workers were part-time, however, and part-time work rarely includes fringe benefits.
Preventing the Misuse of Public Funds. Misuse of the allowance was not a serious problem under Personal Preference, probably because the program took the potential for such a problem seriously. Appropriate use of the allowance was ensured primarily through program approval of the cash management plan and fiscal agent review to verify that expenditures were included in the plan.
The Cash and Counseling model proved administratively feasible and politically tenable in New Jersey during the evaluation period. Data from discussions with program staff, consultant questionnaires, and consumer surveys show that many consumers, who participated in Personal Preference voluntarily, ably managed their supportive services and found it rewarding to do so. In terms of retention and satisfaction, the program seemed equally attractive to elderly and nonelderly adults. New Jersey plans to continue offering Personal Preference as an option to eligible Medicaid beneficiaries.
Each year in the United States, about 1.4 million people with disabilities receive Medicaid supportive services benefits that help them live at home or in other community settings, instead of in institutions (Harrington and Kitchener 2003). Whether states offer such benefits as state plan personal care services (PCS) or home and community-based services (HCBS), they cover them in limited amounts and select the providers or vendors who can supply them. Often, case managers decide which benefits beneficiaries need, and nurses supervise home care workers. This system of service delivery has been criticized for over-medicalizing supportive services and for being too inflexible to effectively meet individual needs. Moreover, home care workers are in perennially short supply. Supply shortages worsen when the economy is strong, and they will likely deepen as the United States population ages and demands more supportive services.
As an alternative to traditional models of service delivery, states are increasingly offering Medicaid beneficiaries and their families opportunities to obtain supportive services directly from individual providers (O’Brien and Elias 2004; Velgouse and Dize 2000). This alternative has become known as “consumer-directed care,” because beneficiaries who use individual providers assume the employer’s role of hiring, managing, and (possibly) terminating their paid caregivers (Eustis 2000). Consumer-directed care is based on the premise that, because supportive services are “low tech” and nonmedical, they do not require the intervention of medical professionals. Rather, beneficiaries should be empowered to direct their own benefits as service consumers (Benjamin 2001; Stone 2001; Eustis 2000; Doty et al. 1996). In 1999, an estimated 139 publicly funded consumer-directed programs served adults or children with physical or developmental disabilities in the United States (Flanagan 2001).
From the perspective of many people affected by disabilities, consumer direction has the potential to meet individual needs better than traditional PCS or HCBS and to promote autonomy and independence. These two basic American values have been affirmed in recent years through policies such as President George W. Bush’s New Freedom Initiative. Consumer direction also could help address the shortage of home care workers by allowing people to pay family and friends for caregiving, thereby expanding the pool of potential workers. Finally, consumer direction could lower public costs by eliminating home care agency involvement in hiring, training, and supervising workers (Stone 2000; Eustis 2000).
Publicly funded consumer-directed programs also raise concerns. These include: (1) whether consumer direction should be available to all users of supportive services, (2) whether to allow family members to be paid for caregiving, (3) how to ensure consumer safety, (4) how to prevent the exploitation of workers, and (5) how to prevent the misuse of public funds (Benjamin 2001; Feinberg and Whitlach 2001; Kane and Kane 2001; Kapp 2000; Simon-Rusinowitz et al. forthcoming; Simon-Rusinowitz et al. 2000; Tilly et al. 2000; Doty et al. 1996).
Cash and Counseling, which is a fairly expansive model of consumer-directed care, provides a flexible monthly allowance that consumers may use to hire workers, as well as to purchase other services and goods they may need (within state guidelines). Adult consumers can designate a representative, such as a relative or friend, to manage, or help them manage, their care. Parents manage the care of consumers younger than 18. In addition, Cash and Counseling offers counseling and fiscal services to help consumers and representatives handle their program responsibilities. These tenets of Cash and Counseling--a flexible allowance, use of representatives, and availability of counseling and fiscal services--are meant to make consumer direction adaptable to consumers of all ages and abilities.
Arkansas, Florida, and New Jersey have each tested the Cash and Counseling model in their Medicaid systems as part of a three-state demonstration. The Robert Wood Johnson Foundation (RWJF) and the Office of the Assistant Secretary for Planning and Evaluation of the U.S. Department of Health and Human Services funded the demonstration. The Centers for Medicare & Medicaid Services (CMS) approved the demonstration under Section 1115 authority of the Social Security Act. The National Program Office for the evaluation, at Boston College and the University of Maryland, provided technical assistance to the states and oversaw the evaluation. Mathematica Policy Research, Inc. (MPR) is the demonstration evaluator.
The evaluation addresses four broad questions: (1) Who participated in the Cash and Counseling demonstration? (2) How were the demonstration programs implemented? (3) How did the programs affect consumers and their caregivers? and (4) How did the programs affect public costs? To estimate the programs’ effects on consumers, caregivers, and costs, the evaluation randomly assigned demonstration enrollees either to participate in Cash and Counseling (the treatment group) or to rely on PCS or HCBS as usual (the control group). With data from telephone interviews and Medicaid and Medicare claims, the evaluation compares the groups’ outcomes at designated follow-up intervals. The evaluation also is describing eligible beneficiaries’ reasons for agreeing or declining to participate in Cash and Counseling, and it is examining trends in the use of PCS and HCBS for indirect evidence that the demonstration affected the number of beneficiaries that used such services.
Research Questions. This report addresses the second broad evaluation question by describing the implementation of New Jersey’s Cash and Counseling demonstration program, Personal Preference. Unlike a companion report that describes demonstration design and program operations in greater detail (Phillips and Schneider 2003), this report focuses on program implementation as experienced by consumers and the program consultants who worked with them. The report considers:
Sources and Methods. This report draws on information and data from several sources:
New Jersey Site Visit.Researchers held in-person discussions with New Jersey state officials, Personal Preference staff members, officials of organizations representing the personal care industry in New Jersey, and staff members of organizations providing enrollment, consulting, and fiscal services under Personal Preference. (New Jersey used the term “consulting,” instead of “counseling,” in its demonstration.) The discussions were conducted in April 2001, about 18 months after the demonstration began random assignment.
Consultant Survey.Also about 18 months into the demonstration, MPR administered a mail survey to Personal Preference consultants. The survey questionnaire contained sections on consultants’ background, program caseload, uses and perceived misuses (if any) of the program allowance by consumers or representatives, and consultant activities. It also contained sections on whether the consultant had seen evidence of abuse of consumers by workers or representatives, recommended changes to consulting activities, and consultants’ overall assessment of the program. Most survey questions offered multiple-choice responses and asked consultants to circle all applicable responses or write in other responses. Questions eliciting consultants’ recommendations and overall program assessment were open-ended. Questionnaires were sent to all 50 consultants who had active Personal Preference caseloads when the survey was administered, and 37 consultants returned them.
Consumer Surveys.MPR conducted telephone interviews with consumers or knowledgeable proxy respondents immediately before consumers were randomly assigned to participate in Personal Preference, and six and nine months later. Each survey instrument covered a range of topics (listed in Table 1). Interviews were completed by 871 treatment group respondents at baseline, 783 at six months, and 747 at nine months.1 To obtain a complete picture of consumers’ Personal Preference experiences, we conducted follow-up interviews even if consumers had disenrolled from the program, were not receiving the monthly allowance, or had died (in which case we interviewed a proxy respondent).
Even among living consumers the use of proxy respondents was fairly widespread. For example, proxies completed 40 percent of baseline interviews (28 percent of interviews for nonelderly adults and 50 percent for elderly adults). Proxy respondents were asked to assess the opinions of consumers. Thus, during follow-up interviews, questions eliciting opinions were not asked if consumers were unable to form opinions (for example, because of a cognitive impairment) or if proxies did not feel comfortable assessing the consumer’s opinion. Questions about the consumer’s satisfaction and unmet needs were not asked if the proxy respondent was also a paid caregiver, because the proxy may have been unable to answer objectively.
Program Records. Personal Preference program records were available for the 871 consumers who were randomly assigned to participate in the program. The records included data on reasons for disenrollment and on receipt and use of the monthly allowance.
| TABLE 1: Topics Covered in MPR Consumer Surveys | ||
|---|---|---|
| Baseline Survey | Six-Month Follow-Up Survey | Nine-Month Follow-Up Survey |
| Household
composition and living arrangements Unpaid assistance Paid assistance, unmet needs, and satisfaction Use of HCBS Health and functioning Attitudes about consumer direction |
Program
participation and allowance receipt Allowance spending plan Use of the allowance Employer responsibilities Reasons for disenrollment |
Program participation and allowance receipt
Health and functioning Living arrangements Unpaid assistance Paid assistance Satisfaction with care and unmet needs Equipment, supplies, and modifications Use of allowance for equipment, supplies and modifications Receipt of community services and use of allowance Use of allowance to hire workers Allowance spending plan and employer responsibilities Reasons for disenrollment |
Survey and program data were analyzed primarily through an examination of frequency distributions, means, and cross-tabulations of constructed variables. Researchers also reviewed and coded open-ended responses to the consultant and consumer surveys. Logistic regression analysis was used to assess whether certain types of consumers fared better in the program (for example, by starting on the allowance and remaining in the program for at least a year). The regression models included a set of explanatory variables drawn from baseline interview and program records data.
Presentation and a Limitation. The body of this report consists of a narrative text and tables of selected descriptive statistics. The report’s appendix also contains many statistical tables. Some of these statistics are discussed in the report. For example, to enlighten the debate about the suitability of elderly adults for consumer direction, many of the appendix tables present statistics by consumer age group (18-64, and 65 or older). In addition, measures of satisfaction and unmet needs are presented by whether consumers responded to evaluation surveys themselves or through proxy respondents.
The report covers a period beginning in early 1996, when New Jersey submitted its demonstration proposal, and ending in July 2003, a year after the last demonstration enrollees had been randomly assigned for the evaluation. Nonetheless, the report is limited in that we conducted site visit discussions at only one point (April 2001), although Personal Preference of course continued to evolve, learn from experience, and make improvements. The report notes some programmatic changes that occurred after the site visit, but it was not possible to document them all.
In implementing a Cash and Counseling demonstration, New Jersey wished to test the feasibility of including a cash allowance model of consumer direction as an option for its state plan Medicaid Personal Care Assistance (PCA) program. From the outset, the state was particularly interested in learning whether consumers would use the allowance to purchase assistive equipment that the PCA program did not cover. As the demonstration unfolded, and a statewide shortage of personal care workers worsened, program staff also became interested in the potential of consumer direction to enlarge the supply of such workers. Although the state did not view savings as a goal of its demonstration, it did believe that Personal Preference might be more cost-effective than traditional PCA because it allowed consumers to purchase services in the free market. The federal government required that the demonstration be budget neutral.2
Adult Medicaid beneficiaries were eligible to enroll in the demonstration if they: (1) were using PCA or had been assessed as eligible for it, (2) were not also participating in HCBS waiver programs or a state-funded consumer-directed program, and (3) were expected to require PCA for at least six months. Recipients of both PCA and HCBS were excluded because authorization procedures differed for those services and consumers would have received assistance from Personal Preference consultants and HCBS case managers, which the program feared would cause confusion. It decided to include only beneficiaries who were expected to require PCA for at least six months because consumers would need several months to develop and implement a plan for spending the Personal Preference allowance.3 Except for this criterion, New Jersey relied on potential enrollees to decide whether they wanted to take on the responsibilities of consumer direction. It continued consumers’ usual PCA benefits until they developed and implemented their cash management plans, and it let them disenroll from Personal Preference at any time. Thus, the state ensured that Medicaid beneficiaries could try consumer direction without great risk.4
Key government and private sector stakeholders supported, or were actively involved in, the New Jersey demonstration. Within the state Department of Human Services (DHS), the Division of Medical Assistance and Health Services (DMAHS) prepared the demonstration proposal and applied for the required federal waivers. The Division of Disability Services (DDS) administered the Personal Preference program, and the executive director of DDS became the project director of Personal Preference. Although the New Jersey governor’s office was not directly involved in the demonstration, both the cabinet-level commissioner of DHS and the DMAHS director strongly supported it. The state board responsible for New Jersey’s Nurse Practice Act also viewed Personal Preference favorably. (The act limits the medical tasks personal care aides can perform, but the board did not believe the activities of workers hired by Personal Preference consumers should be similarly limited. Mindful of the shortage of personal care aides in the state, the board supported the demonstration because it tapped a different labor supply than that available to agencies.)
The state formed an interdepartmental work group to handle certain implementation issues. For example, staff from DMAHS’s Office of Information Services developed special software to track eligible beneficiaries, demonstration enrollees, and allowance recipients, and they worked with a contractor to identify allowance recipients on the state’s Medicaid Management Information System. The staff of a state-funded consumer-directed personal care program shared their experiences with Personal Preference staff.
DDS also involved advocacy organizations and providers of HCBS in the design of Personal Preference. Abiding by the wishes of the Alzheimer’s Association, the state allowed people with cognitive impairments to participate in the demonstration if family members, friends, or other representatives could help them. Otherwise, advocates for elderly people and nonelderly adults with disabilities had no major concerns about the Cash and Counseling model and supported the demonstration. The state’s personal care agencies, on the other hand, saw both pros and cons to the experiment in consumer direction.5 In general, the personal care industry believed some beneficiaries, especially adults who could work if they had help, needed a more flexible personal care program and would benefit from the Cash and Counseling model. The industry also welcomed the prospect of referring to Personal Preference those beneficiaries who were perpetually dissatisfied with agency services. On the other hand, the industry was concerned that consumers could abuse the allowance, family members hired as workers would exploit their situation by not providing agreed upon care, and workers would not be adequately trained. The industry’s major concern, however, was that consumers might hire its agency aides and reduce the agency labor force. To allay this concern, Personal Preference adopted a policy of discouraging consumers from hiring agency aides, and it informed allowance recipients of this policy in writing. Finally, industry representatives who took part in site visit discussions said that, because they knew and respected the director of Personal Preference, they were more willing to cooperate with the demonstration than they might have been otherwise.
To meet the needs of the demonstration evaluation, New Jersey set out to enroll 2,000 beneficiaries (half of whom would be randomly assigned to Personal Preference) into the demonstration in a year’s time. In the previous year, the state had provided PCA to an estimated 12,000 beneficiaries, so the enrollment target represented about 17 percent of eligible beneficiaries. When it became clear that New Jersey could not meet this target, the target was lowered to 1,755, and the enrollment period was extended to 32 months (November 1999 to July 2002).
From the outset of the demonstration, the program conducted community and beneficiary-level outreach activities. The director and assistant director of Personal Preference were responsible for community-level outreach. They made presentations to: (1) advocacy organizations, because their constituents were potential demonstration enrollees; (2) PCA providers, because the state would rely on them to identify potential demonstration enrollees; and (3) human services agencies, because the state planned to recruit them to provide consulting services to Personal Preference consumers. Community outreach was largely successful--the program garnered support and cooperation from advocacy organizations, PCA providers, and human services agencies. At the beneficiary level, however, outreach and enrollment proved challenging enough that the program employed two distinct approaches to these tasks.
Initially, the program contracted with a private, for-profit firm to handle outreach and enrollment, believing that using a contractor would be administratively practical and good for consumers. The selected firm had conducted Medicaid enrollment activities in New Jersey under an existing contract, which was easily amended to encompass Personal Preference. Amending an existing contract was faster than procuring a new contract or recruiting and hiring new state employees. More important, the contractor employed a large, multilingual staff. Compared with the small number of new state employees the program would have been able to hire to conduct enrollment, the contractor seemed to have greater capacity to reach and communicate with the demonstration’s geographically and ethnically diverse target population.
Initially, eligible beneficiaries were invited to enroll in the demonstration when they were assessed or reassessed for PCA. Timing the invitation in this way also seemed to have important advantages. First, the care plan developed from an assessment was needed to determine the amount of the Personal Preference allowance. If enrollment were timed to coincide with assessment, that care plan would remain in effect for about six months (barring a material change in the beneficiary’s condition or circumstances). Second, enrollment at assessment would spread the volume of enrollment-related work over a longer time, making the workload manageable for enrollment staff.
Outreach and enrollment activities initially included the following steps. Whenever personal care agencies conducted assessments of new beneficiaries or reassessments of continuing ones, they completed consumer data forms and sent them to Personal Preference.6 To identify and contact prospective demonstration enrollees, Personal Preference program staff verified beneficiaries’ eligibility against Medicaid records, entered their data into an electronic database, and mailed eligible beneficiaries an introductory letter and flyer. These materials briefly described Personal Preference and said that someone would telephone the beneficiary to schedule a home visit. Program staff then forwarded the prospective enrollee’s contact information to the enrollment contractor, who carried out the remaining activities.
At the enrollment contractor, staff members telephoned each new referral. Using a prepared script, the staff members asked if the beneficiary had received the introductory material, explained the demonstration, and tried to schedule a home visit. Staff scheduled visits when members of the beneficiary’s family would be present, because beneficiaries often sought family members’ advice about whether to participate in the demonstration and because family members might become representatives or paid workers under Personal Preference. During the home visit, enrollment field staff explained the program in detail and told beneficiaries what their monthly allowance would be if they were randomly assigned to the treatment group. Field staff also followed a prepared script, and some showed beneficiaries informational videotapes that had been made for the New Jersey demonstration. Depending on the beneficiary’s participation decision or inclination, field staff either helped them complete enrollment consent forms or tried to schedule a follow-up visit.
Although the enrollment contractor followed agreed upon procedures, it consistently failed to meet its target of 30 enrollees a week and consistently overspent its budget. Personal Preference took four steps to address these problems. It: (1) asked the contractor to spend less time pursuing reluctant or hard to reach beneficiaries; (2) referred new prospects directly to the enrollment contractor, without first mailing introductory materials; (3) began to allow beneficiaries to enroll in the demonstration without a home visit if they did not want one and were already knowledgeable about the demonstration; and (4) separated enrollment from assessment. In particular, it implemented the fourth step by encouraging personal care agencies to refer dissatisfied clients to Personal Preference at any time and by allowing the enrollment contractor to make presentations to groups of Medicaid beneficiaries without already knowing whether they were eligible for PCA.
These changes seemed to boost enrollment slightly, but temporarily. Midway through the demonstration, the director of Personal Preference took the major step of hiring state employees to assume outreach and enrollment responsibilities. In so doing, the program hoped to increase its control over outreach and enrollment, making it easier to try different approaches and to quickly discard those that failed.
Three state employees began work as full-time enrollment specialists in February 2001, after being trained by Personal Preference staff, the enrollment contractor, and a social marketing firm working under contract to the Cash and Counseling National Program Office.7 The enrollment specialists telephoned potential demonstration participants from the Personal Preference office one or two days a week. They spent the rest of the week making home visits and kept in touch with their supervisors by email. Six months after the state employees began work, however, they also had been unable to meet enrollment goals. They enrolled fewer than 50 beneficiaries a month, on average, far below the 70 needed to reach the revised evaluation target. (Appendix Table A.1 shows that half of all Personal Preference consumers enrolled during the first 14 months of the demonstration period; the other half enrolled during the last 18 months.)
Although demonstration enrollment was lower than expected, this may have been because expectations were too high, not because outreach fell short. Personal Preference enrolled about the same percentage of eligible PCA users as did the Arkansas demonstration program. Overall, about 8 percent of New Jersey’s eligible Medicaid beneficiaries enrolled in the demonstration, representing 6 percent of eligible elderly beneficiaries and 9 percent of eligible nonelderly beneficiaries (Foster et al. 2005). Although the program might have tried to boost enrollment in other ways (for example, in Arkansas, a letter from the governor seemed to attract beneficiaries), it is impossible to know how many more people might have enrolled if it had done so.
As noted earlier, New Jersey recruited human services agencies to provide consulting services under Personal Preference. The program was mindful of needing to serve a culturally diverse population and wished to give consumers a choice of agencies from which to receive consulting services. Therefore, the program initially signed memoranda of agreement with 34 agencies throughout the state. These included county boards of social services, Independent Living Centers, adult day care centers, private case management agencies, and Area Agencies on Aging.
Within a few months, however, it became clear that most consumers did not want to choose an agency, and few had enough experience or information on which to base a choice. Thereafter, the program began assigning consumers to agencies based on geographic area and the capacity of agencies to serve consumers effectively. Eventually, the program was assigning consumers to one of 12 agencies (down from 34). Most of these agencies had one or two staff members serving as consultants in addition to performing other agency responsibilities. Personal Preference paid agencies a lump sum per consumer to complete a cash management plan (initially $53, later $75) and an hourly fee thereafter for consulting (initially $18, later $26). It limited these payments to 19 hours (later 20 hours) per consumer per year.
New Jersey selected one organization, a for-profit human services firm, to provide fiscal services to all Personal Preference consumers. The fiscal agent earned fees from consumers and the state. Specifically, consumers were charged for such tasks as cutting checks (75 cents per check), stopping payment on checks ($28 per stoppage), and conducting criminal background checks ($15 to $60 per investigation). The state paid for other tasks, such as processing W-4 and other employment-related forms ($90 per set of forms).
Personal Preference based consumers’ allowances on their PCA care plans. These plans, prepared by Medicaid personal care agencies, indicated the number of weekday and weekend care hours the agency planned to deliver. (Special state authorization was required for more than 25 hours a week.) To determine Personal Preference allowances, the state calculated the amount it would have paid for agency services, then deducted 10 percent to cover the costs of consulting and part of the costs of fiscal services.8
Consumers could use the allowance only for the goods and services specified in their cash management plans. They could receive up to 10 percent of the monthly allowance as cash for incidental purchases if they specified the type of purchase in their plan (for example, care supplies or taxi fare). Likewise, consumers could save a portion of the allowance for one-time purchases identified in their plan (for example, bathroom modifications). Consumers could not use the allowance for food, entertainment equipment or supplies, or vacation or entertainment-related travel. The average monthly allowance at the time of consumers’ enrollment was $1,069 for nonelderly consumers and $1,056 for elderly ones.
The New Jersey Medicaid beneficiaries who enrolled in the demonstration--half of whom were randomly assigned to the treatment group to participate in Personal Preference--were a diverse group. Forty-six percent of treatment group consumers (404) were 18-64 years old, and 54 percent (467) were 65 or older. Slightly more than half these consumers described themselves as White, 38 percent as Black, 9 percent as some other race, and slightly more than one-third as Hispanic (regardless of race) (Table 2). Three consumers in four were female, and 41 percent had graduated from high school.
Approximately one-third of the consumers lived alone, but more than eight in ten received assistance from informal (unpaid) caregivers (Table 2). A substantial proportion of consumers said they lived in nonrural areas characterized by high crime or poor public transportation, where obtaining agency services or hiring individual providers might be difficult. Two-thirds of consumers needed help transferring and using the toilet, and 86 percent needed help bathing.
Nearly 60 percent of consumers said their functioning was worse at baseline than it had been during the previous year. About three-fourths of consumers said they needed more help with personal care than they were receiving at baseline. At that time, slightly less than half of all consumers (45 percent) had been receiving Medicaid PCA for at least six months.
Consumers in the two age groups differed notably in some respects. Elderly consumers were more likely than nonelderly ones to be female, be Hispanic, and have less than a high school education (Appendix Table A.2). Although consumers in both age groups were equally likely to need help transferring, bathing, and using the toilet at baseline, elderly consumers were more likely than nonelderly ones to say their functioning had worsened. Among elderly consumers, primary informal caregivers were most commonly their adult children (60 percent). In contrast, nonelderly consumers reported a greater variety of primary informal caregivers, including parents (25 percent), sons and daughters (19 percent), other relatives (17 percent), and nonrelatives (16 percent). Regardless of age group, few consumers reported that a spouse was their primary informal caregiver.
At the time of the baseline interview, more than nine in ten consumers said having a choice about the types of help they received was very important (Appendix Table A.3). More than eight in ten said having a choice about when caregivers came was very important. Three-fourths of consumers said the ability to pay family members was very important, and seven in ten said the same about paying friends.
| TABLE 2: Consumer Characteristics at Random Assignment | |
|---|---|
| Percentage | |
| AGE | |
| 18 to 39 | 15.3 |
| 40 to 64 | 31.1 |
| 65 to 79 | 32.3 |
| 80 or older | 21.4 |
| SELF-IDENTIFIED AS: | |
| White | 52.9 |
| Black | 38.0 |
| Other race | 9.1 |
| Hispanic (Regardless of Race) | 35.5 |
| FEMALE | 74.2 |
| GRADUATED HIGH SCHOOL | 41.2 |
| LIVES ALONE | 35.3 |
| HAS AT LEAST ONE INFORMAL CAREGIVER | 84.3 |
| LIVES IN AN AREA THAT IS: | |
| Rural | 10.5 |
| Nonrural but with high crime or poor public transportation | 44.6 |
| NOT INDEPENDENT IN: | |
| Transferring | 66.8 |
| Using toilet | 67.0 |
| Bathing | 86.3 |
| FUNCTIONING WORSE NOW THAN LAST YEAR | 58.7 |
| NEEDS MORE HELP WITH PERSONAL CAREa | 74.3 |
| PROXY RESPONDENT COMPLETED AT LEAST HALF OF BASELINE INTERVIEW | 40.0 |
| USED MEDICAID PCA FOR 6 MONTHS OR LONGER | 45.2 |
SOURCE: Personal Preference Program records and MPR consumer interviews
conducted by telephone immediately before consumers’ random assignment.
The table summarizes the characteristics of the 871 consumers randomly assigned
to participate in Personal Preference.
|
|
Consumers could begin using Personal Preference consulting services as soon as they were assigned to the program. Of the 37 consultants who completed the MPR questionnaire, 20 had been working for the Personal Preference program for more than a year when surveyed, the rest for less time (Appendix Table A.4). Each had an average caseload of six consumers when surveyed but reported having worked with an average of nine consumers altogether.
Consultants potentially had many responsibilities. During initial home visits, they helped consumers (or representatives) develop written plans for using the monthly allowance. Consultants reviewed the completed cash management plans and sent them to the Personal Preference program for formal approval. The program required consultants to speak with consumers by telephone at least monthly for the first six months after random assignment and to meet them in person quarterly, to monitor their well-being. Most consultants told consumers who had completed their cash management plans to call them if any questions arose while they were implementing their plan. For example, consumers could ask consultants for advice about recruiting and hiring workers and about making back-up arrangements. If consultants did not hear from newly enrolled consumers, they would call the consumers to fulfill the monthly requirement. If consumers needed to revise their cash management plans at any time, consultants helped them make revisions, reviewed the new plan, and forwarded it to the state program office for approval.
According to data from the consultant questionnaire, consultants devoted most of their time to a few tasks. At the time the questionnaire was administered, consultants spent only four hours a week, on average, on Personal Preference duties (Appendix Table A.5). However, it is likely that consultants later spent more time on program duties as demonstration enrollment continued, the number of consulting agencies decreased, and consultants’ caseloads grew.
Consultants reported that their most time-consuming tasks were: (1) helping consumers develop cash management plans; (2) performing administrative activities such as record keeping, updating case notes, and contacting other program staff; (3) advising consumers about payroll-related activities, such as setting wages and estimating payroll taxes; and (4) listening to or encouraging consumers. Most consultants believed that these services were of value to consumers. Of 37 consultants, 13 reported that at least one of their consumers required extensive monitoring (Appendix Table A.6). Consultants said the most common reasons for this were that consumers had difficulty completing paperwork (reported by ten consultants), had no experience as an employer (reported by seven consultants), or experienced frequent worker turnover (reported by seven consultants).
Many consumers were enrolled in Personal Preference for a long time before they began receiving their monthly allowance (Table 3). Others disenrolled without ever having received it. Three months after being assigned to the program, about 30 percent of all consumers had begun receiving the allowance, 61 percent were still enrolled but had not received the allowance, 7 percent had disenrolled, and 2 percent had died (Appendix Table A.7 and Table A.7a). Six months after assignment to Personal Preference, slightly more than half of all consumers had received the allowance, and nearly one-quarter were still enrolled but had not received it. Between the three and six-month points, the proportion of disenrolled consumers tripled (from 7 to 21 percent).
Initially, the program expected consumers to have completed a cash management plan and be receiving the allowance within 90 days of random assignment to the treatment group. (Consultants, not consumers, would have been held to this standard.) For reasons explained below, however, developing the plan often took much longer than 90 days, and the program did not enforce the 90-day standard. Senior program staff feared that doing so might lead consultants to develop the plan instead of helping the consumer do it. In effect, the program gave consumers as much time as they needed to make the transition to consumer direction. Of the 198 consumers who were still enrolled but had not received the allowance at six months, 22 percent (44 consumers) did receive the allowance before the end of the follow-up year (not shown). Their eventual success may affirm the program’s view that it was never too late for consumers to become active program participants. Some of these consumers may have been delayed by illness or by trying to recruit workers other than family members. Others may have needed more time to fully understand the program. For most consumers, however, not receiving the allowance within six months was tantamount to never receiving it, at least during the follow-up year. Of the 198 consumers mentioned above, 93 formally disenrolled, and 47 simply did not receive the allowance during the year. (The remaining 14 consumers died.)
| TABLE 3: Time from Random Assignment to Start of Monthly Allowance, by Age Group | |||
|---|---|---|---|
| Percentages | |||
| Overall | 18 to 64 | 65 or Older | |
| STARTED MONTHLY ALLOWANCE BY END OF MONTH: | |||
| 3 | 31.5 | 31.7 | 31.3 |
| 6 | 56.7 | 58.7 | 55.0 |
| 9 | 64.8 | 67.1 | 62.7 |
| 12 | 66.6 | 69.6 | 64.0 |
| SOURCE: Personal Preference Program records for the year following consumers’ random assignment. This table represents the 871 consumers randomly assigned to participate in Personal Preference. | |||
Allowance planning procedures, which affected the amount of time it took for the consumer to receive the allowance, were as follows. Consumers worked with their consultant to develop the cash management plan, which was to identify workers and other vendors, itemize desired goods and services in the amounts required, multiply these by unit or hourly costs, and account for applicable taxes. The consultant sent completed plans to the Personal Preference program office, where staff approved or denied the requested goods and services. The program office returned unacceptable plans to the consumer and forwarded approved plans to the fiscal agent. The fiscal agent double-checked all plans for accuracy and reviewed the forms in the proposed worker’s employment package for consistency with the plan. If the fiscal agent found problems with the plan or the employment forms, it returned the paperwork to the consumer. After all paperwork was approved and processed, the program office notified the consumer’s usual Medicaid personal care agency that it was to stop serving the consumer in 30 days’ time. Regardless of when the planning process was completed, however, the allowance would commence only on the first day of the following month, because of how New Jersey’s Medicaid Management Information System operated.
The program office took two steps that helped reduce the length of time it took to start the allowance. It reduced the notice it gave to personal care agencies from 30 to 14 days, and it began assigning consumers to consulting agencies, instead of asking them to choose an agency, as they entered the program. Still, completing a cash plan could be onerous. When the fiscal agent returned paperwork to consumers, those consumers who did not understand the problem or how to resolve it would have to contact their consultant. Some consultants complained that the fiscal agent returned paperwork without indicating the error. Even consultants sometimes could not identify the error and had to call the fiscal agent for an explanation. Both Personal Preference program staff and consultants reported frequent communication problems between consumers (or representatives) and consultants, state program staff, and fiscal agent staff. The number of people involved made it difficult for consumers to know whom to call about a particular type of problem. When problems arose, some consultants saved time by holding three-way calls among the consumer, the fiscal agent, and the consultant.
Use of Representatives. As noted earlier, Personal Preference consumers could designate an (unpaid) representative to manage, or help them manage, their program responsibilities. Representatives could help consumers decide how to spend the allowance (for example, whether to hire a worker, whom to hire, and how much to pay), supervise workers and monitor care, sign worker time sheets, and handle other program paperwork. No one could serve both as a consumer’s representative and as a paid worker.
During site visit interviews, Personal Preference consultants reported that up to two-thirds of elderly consumers named a representative, while the proportion was considerably smaller among nonelderly adults.9 Most consultants who completed questionnaires (30 of 37) said they worked with at least one consumer who used a representative (Appendix Table A.8). Representatives almost always were consumers’ family members or close friends. New Jersey’s ethnic and language diversity was a major reason for the widespread use of representatives. The program sometimes had to send consumers materials written in English only, and consumers may have depended on representatives to translate them. (Such mailings included brief notices that the materials were important and should be translated. The notices themselves were printed in 14 languages.) In addition, consumers with cognitive impairments or who already were receiving assistance with their affairs, such as help maintaining a checking account, were likely to use representatives. Program staff believed that about half the consumers who named representatives could have managed independently but felt too insecure to try.
Representatives’ decision-making roles varied considerably. Except when consumers were completely unable to communicate their preferences, consumers and representatives typically shared decisions. In some cases, they made decisions as a team. In others, the representative asked the consumer’s preference but then made the final decision. In still other cases, the consumer was the primary decision maker, but the representative served as a liaison to the program. Four consultants who completed questionnaires indicated that they “questioned the suitability” of a consumer’s representative, and one indicated observing “a serious divergence of wishes or interest” between a consumer and a representative. (Quotes indicate the wording of closed-ended questions, not responses. Respondents were not asked to elaborate.)
Use of, and Satisfaction with, Consulting and Fiscal Services. Like representatives, program consultants and fiscal agent staff helped consumers manage their program responsibilities. In spite of allowance delays, many consumers were pleased with the consulting and fiscal services they received. During six-month follow-up interviews, 72 percent of all consumers said a Personal Preference consultant helped them or their representative develop a cash management plan (Table 4 and Appendix Table A.9, Table A.9a and Table A.9b).10 More than nine in ten consumers who received help from consultants found it useful. They most commonly said consultants’ explanations of program rules were useful (reported by 76 percent of consumers), as was help clarifying goals, options, and priorities (reported by 40 percent of consumers) (Appendix Table A.10). By the time of their six-month interviews, 58 percent of consumers had received materials about recruiting workers, and 84 percent of them found the materials useful (Appendix Table A.9). Although a smaller percentage of consumers (42 percent) said their consultant advised them about recruitment (as opposed to simply giving them materials), the proportion that found the advice useful was high (92 percent) (Appendix Table A.9). As we describe later, some consultants who took part in site visit discussions said they were not sure how much recruitment assistance the program expected them to provide.
| TABLE 4: Use of, and Satisfaction with, Personal Preference Services | ||
|---|---|---|
| Percentage Reporting | Of Users, Percentage Finding It Useful | |
| Had Help with Cash Management Plan During First 6 Months | 71.7 | 93.7 |
| Received Advice About Recruiting During First 6 Months | 42.0 | 91.5 |
| Received Advice About Training Workers During First 6 Months | 33.9 | 86.5 |
| Used Fiscal Services During First 9 Months (if Received Allowance) | 97.0 | 92.4 |
| SOURCE: MPR consumer interviews, conducted by telephone 6 and 9 months after consumers’ random assignment. The table summarizes responses of 783 consumers who completed 6 month interviews and 747 consumers who completed 9 month interviews. | ||
The major fiscal services offered to consumers were: (1) check writing, and (2) preparing and filing tax returns for workers hired with the monthly allowance. Personal Preference allowed consumers to receive the allowance in cash and handle fiscal responsibilities themselves if they first passed a skills examination. This option garnered little interest, however; according to program staff, all consumers chose to use the services of the fiscal agent during the evaluation follow-up year. Of consumers who started receiving the cash allowance within nine months of random assignment, 97 percent said they used the program’s fiscal services, and 92 percent of them said the services were useful (Appendix Table A.9).11 These proportions varied little by age group.
Recruiting and Hiring Workers. Fifty-seven percent of all consumers reported that they had hired at least one worker with the allowance by the time of the nine-month follow-up interview, 25 percent had tried to hire but did not, and 18 percent had not tried (Table 5 and Appendix Table A.11). Most consumers who tried to hire family members were able to do so, but consumers had less success hiring other people they knew, such as friends, neighbors, church members, and agency workers.12 Consumers who tried to hire workers they did not already know (for example, by asking others for recommendations or posting ads) also were less successful than those who hired family.
| TABLE 5: Recruiting and Hiring Workers | |
|---|---|
| Percentage | |
| HIRING WORKERS WITH ALLOWANCE DURING FIRST 9 MONTHS | |
| Hired a worker | 57.2 |
| Tried to hire a worker, but did not | 24.7 |
| Did not try to hire a worker | 18.1 |
| ATTEMPTED RECRUITING METHODS, IF HIRED OR TRIED TO HIRE WORKERS | |
| Tried to
hire: Family member Friend, neighbor, or church member Home care agency worker |
73.8 40.0 28.9 |
| Asked family or friends to recommend worker | 32.1 |
| Posted or consulted advertisements | 7.5 |
| Contacted employment agency | 5.5 |
| SUCCESSFUL RECRUITING METHODS, IF HIRED WORKERS HIRED: | |
| Family member | 72.5 |
| Friend, neighbor, or church member | 23.6 |
| Home care agency worker | 13.3 |
| Through a recommendation | 11.4 |
| Through an advertisement | 5.3 |
| Through an employment agency | 0.9 |
| Through other means | 1.4 |
| SOURCE: MPR consumer interviews, conducted by telephone 6 and 9 months after consumers’ random assignment. The table summarizes responses of 815 consumers who responded to either or both interviews. | |
Although similar proportions of elderly and nonelderly consumers hired successfully, a larger proportion of nonelderly consumers than elderly ones tried to hire but did not (30 percent versus 20 percent; Appendix Table A.11). The difference seems related to whom nonelderly and elderly consumers tried to hire. Nonelderly consumers were less like than elderly ones to try to hire family (68 percent versus 79 percent), and they were more likely to try to hire friends and neighbors (48 percent versus 32 percent; top panel of Appendix Table A.12).
Of consumers who hired workers, 29 percent said they had difficulty doing so, and one-third of them said the difficulty was finding interested or qualified candidates (Appendix Table A.11). Consultants confirmed that it was difficult for some consumers to hire or keep workers. Eighteen consultants who completed a questionnaire said they worked with at least one consumer who had serious problems because their workers quit or were fired (Appendix Table A.13). Still, some consumers did have success with creative recruiting strategies. One consultant described a Chinese consumer who advertised in a Chinese-language newspaper and successfully recruited a worker who spoke the consumer’s language and prepared the consumer’s preferred foods. Another consumer hired a live-in aide through a nonMedicaid agency, after years of being dissatisfied with visiting aides from Medicaid PCA agencies.
In addition to recruiting and hiring workers, consumers had to decide whether and how to train them, how much to pay, whether to offer fringe benefits, and whether to describe such arrangements in a contract or written agreement. A sizable proportion of consumers or representatives who hired workers by the time of their nine-month interviews trained them in some way. Overall, 44 percent showed the worker how to perform tasks, and another 4 percent arranged for training outside the home (Appendix Table A.11). Only 11 percent of these consumers said that training workers was difficult. Consumers paid workers $9.84 an hour, on average. Fifteen percent of consumers said they provided fringe benefits, such as paid sick time, to their workers.13 Fifty-three percent of consumers who used their allowance to pay workers, including family members, signed contracts or work agreements with them. (Data on wages, fringe benefits, and contracts are not shown in tables.)
Neglect, Exploitation, and Abuse. The possibility that consumers could be exploited by workers or representatives, or vice versa, and the possibility that the Personal Preference allowance would be misused or squandered were major concerns for all involved in the program, as they were for the demonstration programs in Arkansas and Florida. At the same time, everyone directly involved in the demonstration realized that extensive control and oversight of consumers and their families were incompatible with the philosophy of consumer direction. Consumers had to be free to make their own choices, even if others disagreed with them. Personal Preference relied on consultants to be alert for evidence that consumers were being financially exploited or physically or verbally abused. It also established procedures for consultants to follow if they suspected anything was amiss. At the time of the New Jersey site visit and the consultant survey, neither neglect or exploitation, nor abuse of the allowance, seemed to be serious problems. On the questionnaire, one consultant out of 37 indicated seeing evidence of financial exploitation of one consumer by a worker, but did not provide any information about the case (Appendix Table A.14). No consultants who completed a questionnaire reported verbal or physical abuse of consumers by representatives or workers, but one reported seeing evidence of consumer self-neglect.
During site visit discussions, a consultant described one case in which a consumer seemed to have been subject to neglect. The problem was identified during the consultant’s initial home visit with the consumer--when the consultant found the consumer lying on a couch apparently comatose--before program participation had even begun. Following established procedure, the consultant immediately notified the state program office that the case needed investigation. For this case and others that caused concern, program staff then referred the case to a nurse employed by the state Medicaid program. The nurse visited the home to make an assessment, and program staff reviewed the nurse’s report. If the staff concluded that neglect or exploitation was likely, the case was referred to Adult Protective Services.
Personal Preference used to two methods to prevent misuse of the monthly allowance: (1) program approval of cash management plans; and (2) verification, by the program’s fiscal agent, that expenditures were authorized under the plan. No one who participated in site visit discussions--program staff, fiscal agent staff, program consultants, or stakeholders in the personal care industry--reported seeing evidence of material abuse of the allowance under Personal Preference. When a few consumers planned to use the allowance in ways not related to their personal care or independence, the program simply denied the requests.14 In addition, the fiscal agent was also required to provide consumers with monthly statements of their account credits, debits, and payments pending. The main purpose of the statements was to ensure that consumers knew their account balances; however, many consumers seemed not to understand that invoices pending were not reflected in the statement’s bottom line. Thus, the statements probably were not very useful in preventing misuse of the allowance (although they did allow consumers to identify any errors the fiscal agent made).
Consumers who used their allowance to hire workers determined how many to hire, what tasks they would perform, and when they would help. Among consumers who used the Personal Preference allowance to pay workers, about three-fourths hired one worker, another fifth hired two workers, and the remaining five percent hired three or more workers (Table 6 and Appendix Table A.16). Forty-four percent of consumers who hired had a paid worker who lived with them at the time of the nine-month interview. The number and type of workers (live-in or visiting) hired with the allowance varied little by age group.
| TABLE 6: Assistance from Paid Workers Among Consumers Who Hired with the Allowance, by Age Group | |||
|---|---|---|---|
| Overall | 18 to 64 | 65 or Older | |
| IN 2 WEEK PERIOD SHORTLY BEFORE INTERVIEW: | |||
| Had 1 worker | 75.0 | 71.4 | 78.3 |
| Had 2 or more workers | 25.0 | 28.7 | 21.7 |
| Had visiting worker(s) | 64.8 | 67.0 | 62.8 |
| Had live-in-worker(s) | 43.6 | 41.6 | 45.4 |
| WORKER HELPED WITH: | |||
| Housework or community choresa | 99.5 | 99.5 | 99.5 |
| Personal careb | 97.7 | 97.3 | 98.1 |
| Routine health carec | 91.8 | 90.3 | 93.2 |
| Transportationd | 66.8 | 70.8 | 63.3 |
| HOURS OF PAID CARE | |||
| 14 or fewer | 5.1 | 5.2 | 5.0 |
| 15 to 42 | 48.0 | 48.6 | 47.5 |
| 43 to 70 | 31.1 | 36.4 | 26.5 |
| 71 or more | 15.8 | 9.8 | 21.0 |
SOURCE: MPR
consumer interviews, conducted by telephone 9 months after consumers’
random assignment. The table summarizes the responses of 392 consumers who
hired with the allowance by the time of their interview and received paid
assistance during a 2 week period shortly before the interview. Of these
consumers, 11 had disenrolled from Personal Preference and were probably
reporting on help from agency workers.
|
|||
Although consumers in both age groups paid their workers to provide the same types of care, they differed in the amount of care they purchased. During the two-week period asked about in follow-up interviews, more than nine in ten consumers in both age groups said their paid workers helped them with housework or community chores, personal care, and routine health care (Table 6 and Appendix Table A.16). Two-thirds of consumers reported that their workers helped them with transportation (this percentage was somewhat higher for nonelderly consumers). During the same period, about half the consumers in both age groups paid their workers for 15 to 42 hours of the care they provided. Elderly consumers were more likely than their nonelderly counterparts to pay workers for substantially more hours. One-fifth of elderly consumers paid for 71 or more hours of care in two weeks, compared with one-tenth of nonelderly consumers.
According to fiscal agent records from month 8 of consumers’ program participation, more than eight in ten consumers used part of their allowance to pay workers, roughly five in ten received some of the allowance as cash for incidental purchases, and fewer than one in ten used the allowance to buy equipment (Appendix Table A.17).15 Both elderly and nonelderly consumers spent about 80 percent of the allowance paying workers (Appendix Table A.17a and Table A.17b). They received about 8 percent of the allowance as cash for incidental purchases that month. This amounted, in the month studied, to $29 received as cash by elderly consumers and $42 received as cash by nonelderly consumers.
Consultants’ reports about the contents of consumers’ cash management plans were consistent with the data from the fiscal agent. At least ten of the consultants who completed questionnaires reported the following plans for spending the allowance (other than paying workers): taxi fare or other transportation services, chore or homemaker services, laundry services, and ramps (Appendix Table A.18). When responding to the questionnaire, few consultants mentioned creative uses of the allowance (Appendix Table A.19). During site visit discussions however, some did mention creative equipment purchases. These included: (1) a portable support for a voice synthesizer so that the consumer could wear the synthesizer outside his home; (2) a scanner and talking computer that allowed a consumer to read mail and check worker time sheets; and (3) a fax machine so that a consumer with quadriplegia could send papers to doctors, insurance companies, and Personal Preference. While the Personal Preference Program staff was ultimately responsible for approving or denying the items in consumers’ cash plans, eight consultants mentioned that they denied consumers’ attempts to include cigarettes, food, or alcohol in their plans. For their part, 30 percent of consumers said program rules kept them from using their allowance to buy things that would have increased their independence. (They were not asked to provide examples.) However, only five consumers who voluntarily disenrolled cited program rules as the reason they did so.
Satisfaction with Personal Preference. Consumers were quite satisfied with the Personal Preference program. Of all respondents to the nine-month evaluation interview, 91 percent said they would recommend the program to others who wanted “more control over their personal care services” (Table 7). Among allowance recipients, 82 percent said the allowance had improved their quality of life “a great deal” or “somewhat.” Across age groups, elderly consumers were slightly more likely than nonelderly ones to report a great deal of improvement. When asked about the most important way the allowance improved their lives, consumers in both age groups most commonly cited the ability to choose their own caregivers (reported by 42 percent of elderly consumers and 33 percent of nonelderly ones), followed by the ability to obtain a higher quality of care than had been available previously (reported by 13 percent and 19 percent of elderly and nonelderly consumers, respectively) (Appendix Table A.20a and Table A.20b). About 12 percent of consumers in both age groups said the allowance helped them feel more independent, in control, or emotionally healthy.
Satisfaction with PCA. Nine months after random assignment, most consumers reported they were somewhat or very satisfied with their overall care arrangements and with specific aspects of their paid care (such as whether it was usually completed). Nonetheless, sizable proportions of consumers needed help, or more help, with some activities. Across age groups, consumers were largely, but not altogether, similar in their reports of satisfaction. Where differences appeared, elderly consumers were more sanguine than their nonelderly counterparts. For example, when asked to rate their satisfaction with their overall care arrangements, elderly consumers were more likely than nonelderly ones to be “very satisfied” and less likely to be “dissatisfied” (Appendix Table A.21a). Elderly consumers were also more likely to report that it would not be difficult to change their paid caregivers’ schedules if they needed to, and they were more likely to say paid caregivers never neglected them.
| TABLE 7: Satisfaction with Personal Preference, by Age Group | |||
|---|---|---|---|
| Percentages | |||
| Overall | 18 to 64 | 65 or Older | |
| WOULD RECOMMEND PERSONAL PREFERENCE TO OTHERS WANTING MORE CONTROL OVER THEIR PERSONAL CARE SERVICES | 91.1 | 90.9 | 91.2 |
| EFFECT OF MONTHLY ALLOWANCE ON QUALITY OF LIFE, AMONG RECIPIENTS | |||
| Improved a great deal | 57.2 | 54.2 | 60.0 |
| Improved somewhat | 24.7 | 26.7 | 22.9 |
| Stayed the same | 17.5 | 17.8 | 17.1 |
| Reduced somewhat | 0.4 | 0.9 | 0.0 |
| Reduced a great deal | 0.2 | 0.4 | 0.0 |
| SOURCE: MPR consumer interviews, administered by telephone 9 months after consumers’ random assignment. The table summarizes the responses of 747 consumers. | |||
Among consumers who hired with the allowance, the proportions reporting unmet needs for help with activities around the house or community, personal care, routine health care, and transportation ranged from about one-third to one-half (Appendix Table A.21a). In each instance, elderly adults were less likely than nonelderly ones to report unmet needs. The largest differences were in unmet needs for help with housework or community chores (reported by 40 percent of elderly adults and 54 percent of nonelderly ones) and for help with transportation (reported by 31 percent of elderly adults and 46 percent of nonelderly ones). Like the other demonstration states, however, New Jersey did not expect that consumer direction would eliminate all unmet needs, which may be impossible at any cost.
Because proxy respondents commonly completed evaluation interviews on consumers’ behalf, we compared their reports of consumers’ satisfaction with those of self-respondents (Appendix Table A.20b and Table A.21b). The two groups’ assessments were largely consistent with each other; however, proxy respondents were more likely than self-respondents to give very favorable ratings on broad measures of consumer satisfaction, such as overall satisfaction, the ability to get help with transportation, and the effects of the monthly allowance on quality of life. Proxy respondents were more likely than self-respondents to report that consumers had unmet needs for PCA at follow-up. This could be because proxies saw unmet needs where consumers did not or because consumers who needed proxies began with greater needs and still had them despite the program.16
Disenrollment. As in the other demonstration states, a substantial proportion of New Jersey consumers--33 percent--disenrolled from Personal Preference within a year of enrollment (Appendix Table A.22). Most did so voluntarily, according to program records, but others were disenrolled because they lost Medicaid or PCA eligibility, or because the program could not locate them. In addition, 6 percent of all consumers died. Disenrollment was not more common in one age group than the other, but more elderly than nonelderly consumers died during the follow-up year (7 percent versus 4 percent). Sixty-three percent of consumers who disenrolled from the program did so within six months of enrolling.
During six or nine month interviews, consumers who disenrolled voluntarily were asked why they had done so. The most commonly cited reasons pertained to employer responsibilities (reported by 34 percent of voluntary disenrollees) (Appendix Table A.22). Given that most consumers (73 percent) disenrolled without ever having received the program allowance, we infer that many of these disenrollees decided they did not wish to assume the responsibilities of an employer or could not find anyone to hire. Others who disenrolled said they changed their mind or were satisfied with their usual PCA services (30 percent), that they had problems with fiscal responsibilities (11 percent), or that the allowance was not enough (9 percent). Consumers’ reasons for choosing to disenroll differed somewhat by age group, even though their overall rates of disenrollment did not. Nonelderly consumers were most likely to disenroll because they believed it was or would be difficult to assume the responsibilities of an employer, which may reflect their difficulty in hiring nonrelatives. In contrast, elderly consumers were most likely to disenroll because they changed their minds or were satisfied with their usual PCA services.
Because demonstration enrollment was voluntary, Personal Preference presumably attracted Medicaid beneficiaries who wished to direct their own personal care. Nonetheless, participating in the program--developing a cash management plan, hiring workers, and purchasing other services and goods--may have required more effort than some consumers and representatives were willing to expend. Satisfaction with the program was high, but not universal. After being randomly assigned to the treatment group, what types of consumers found Personal Preference worthwhile and satisfying? One could speculate that consumers who felt ill served by New Jersey’s usual PCA program would be more willing than others to undertake the responsibilities of Personal Preference. This group might include consumers who found agency workers unreliable or too unlike them ethnically or culturally. Consumers with a strong desire to pay family or friends for caregiving might also be more motivated than others to fully participate in the program once enrolled. Conversely, one could speculate that consumer direction might be difficult for consumers with poor health or functioning, for those without hiring or supervisory experience, or for those without someone in mind to hire. Understanding the relationship between the characteristics, circumstances, and motivation of consumers and their probability of success at consumer direction could help program administrators hone their outreach efforts, identify possible shortcomings in program services, and dispel any prejudices about beneficiaries’ suitability for consumer direction.17
In this analysis, key indicators of consumers’ experiences with Personal Preference were regressed against a fairly comprehensive, but selected, set of characteristics measured during consumers’ baseline interviews. The outcomes were whether treatment group consumers:
The following discussion considers characteristics that were associated with outcomes at the .05 significance level. Estimated coefficients and p-values are found throughout Appendix Tables A.23 to A.27, as noted.
Several characteristics were associated with whether treatment group consumers received the program allowance within nine months of random assignment and remained in the program for nine months or one year. Consumers who considered it very important, at baseline, to be able to pay family or friends for caregiving were more likely than other consumers to receive the monthly allowance and stay in the program (Appendix Table A.23 and Table A.27). These consumers had a particular motivation for joining the demonstration, and they already had workers in mind to hire. Consumers who needed help getting in or out of bed and consumers who had unmet personal care needs were more likely than other consumers to stay in the program. These consumers, if they objected to the timing of agency services or to having agency workers help them with intimate tasks, may have found a better way to meet their needs under Personal Preference. Consumers whose primary informal caregivers were employed at baseline also were more likely than others to stay in the program. Employed caregivers may have found the inflexibility of agency services frustrating, and consumer direction may have provided the opportunity to hire workers who could be more accommodating.
In contrast, some treatment group consumers who did not receive the allowance or remain in the program may have been satisfied with their usual PCA services. In particular, consumers who had two or more paid caregivers in the week before baseline were less likely than consumers with no paid caregivers to receive the allowance or stay in the program (Appendix Table A.23 and Table A.27). If these consumers felt that their Medicaid PCA agency served them well, they may simply have decided not to switch to consumer direction.
Living alone seemed to be an obstacle to full program participation. Treatment group consumers who lived alone were less likely than consumers who lived with others to receive the monthly allowance and more likely to say hiring was difficult (Appendix Table A.23 and Table A.24). Live-in family members often serve as representatives or paid workers, and consumers without such nearby resources may have had difficulty participating in Personal Preference by themselves. In fact, consumers with two informal caregivers at baseline were more likely than consumers with none to receive the allowance and stay in the program for at least nine months (although the associations were significant only at the .10 level) (Appendix Table A.23 and Table A.27).
Hispanic treatment group consumers were less likely than nonHispanic ones to receive the Personal Preference allowance and stay in the program, and Black consumers were less likely than White consumers to receive the allowance (Appendix Table A.23 and Table A.27). NonWhite consumers were more likely than white consumers to say program rules were too restrictive (Appendix Table A.25). New Jersey’s population of PCA users is ethnically and racially diverse. The state tried to address this diversity in its marketing and informational materials, but it may have had difficulty providing consulting services to this population. For example, during baseline surveys, 36 percent of consumers described themselves as Hispanic, compared to 16 percent of consultants (or six of 37). Although some consumers may have liked the idea of hiring workers of their own ethnicity, communicating with others involved in the program may have been problematic if the consumer did not speak or read English.
All else being equal, treatment group consumers who joined the New Jersey demonstration relatively early in the enrollment period were less likely than later enrollees to have received their allowance promptly (Appendix Table A.23). Early enrollees were also more likely to find hiring difficult and less likely to be very satisfied with their overall care (Appendix Table A.24 and Table A.26). The better outcomes for later enrollees suggest that the program’s efforts to shorten or remove some allowance delays may have succeeded. In addition, the specially hired state employees may have enrolled a more select group of beneficiaries--ones who understood the program well or were more motivated to participate--than the group enrolled by the contractor initially retained for outreach and enrollment.
Among treatment group consumers who hired or tried to hire workers, a few characteristics (in addition to those already mentioned) were predictive of whether hiring was difficult (Appendix Table A.24). Consumers who lived in nonrural areas with crime or transportation problems were more likely than consumers who lived in nonrural areas without those problems to have difficulty hi