Publication Date

May 29, 2006

U.S. Department of Health and Human Services

Examination of Texas Rider 37: A Medicaid "Money Follows the Person" Long-Term Care Initiative

Barbara A. Ormond, Anna S. Sommers and Kirsten J. Black

The Urban Institute

May 30, 2006

This report was prepared under contract # HHS-100-03-0011 between the U.S. Department of Health and Human Services (HHS), Office of Disability, Aging and Long-Term Care Policy (DALTCP) and The Urban Institute. 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, Linda Bergofsky, at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, SW, Washington, DC 20201. Her e-mail address is: Linda.Bergofsky@hhs.gov.

EXECUTIVE SUMMARY

INTRODUCTION: Overview of Rider; Study Objectives; Data and Methods

FINDINGS: Components of the Rider 37 Initiative; Rider Participants and Comparison Groups; Nursing Facility Industry Response to the Initiative; Service Costs and Program Expenditures for Medicaid Long-Term Care

SUMMARY

APPENDICES: APPENDIX 1: DATA AND METHODS; APPENDIX 2: TILE CLASSES

ENDNOTES

LIST OF TABLES: TABLE 1: Share of Texas Nursing Facilities with and without Transitions Under Riders 37 and 28 in Medicaid, Statewide and by Region; TABLE 2: Characteristics of Texas Nursing Facilities with and without Transitions Under Riders 37 and 28 in Medicaid; TABLE 3: Demographic Characteristics of Texas Medicaid Long-Term Care Recipients Rider Participants, Nursing Facility Residents, and CBA Clients; TABLE 4: Texas Medicaid Long-Term Care Recipients: Rider Participants, Nursing Facility Residents, and CBA Clients Distribution of Texas Index for Level of Effort Class; TABLE 5: Texas Medicaid Long-Term Care Recipients: Rider Participants, Nursing Facility Residents, and CBA Clients Percentage with Selected Diagnoses by Age Group; TABLE 6: Percentage of Texas Rider and CBA Medicaid Participants Receiving Assisted Living Services in Six Month Period After Entry into CBA Program; TABLE 7: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Urban/Rural Status; TABLE 8: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Assisted Living and Urban/Rural Status; TABLE 9: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Region: Recipients with No Assisted Living; TABLE 10: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Race/Ethnicity: Recipients with No Assisted Living; TABLE 11: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program: Texas Index for Level of Effort Class 211, All Recipients and Recipients with No Assisted Living; TABLE 12: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Texas Index for Level of Effort Class: Recipients with No Assisted Living; TABLE 13: Recipient Status at Six and Twelve Months after Entry into Texas Long-Term Care Medicaid CBA Program: Rider Participants and CBA Clients by Urban/Rural Residence; TABLE 14: Recipient Status at Six and Twelve Months after Entry into CBA Program by Texas Index for Level of Effort Class; TABLE 15: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients; TABLE 16: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients with Assisted Living; TABLE 17: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients with No Assisted Living; TABLE 18: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients by Texas Index for Level of Effort Class; TABLE 19: Factors Associated with a Possible Demand Response by Texas Rider 37/28 Medicaid Participants

EXECUTIVE SUMMARY

Background

The Texas state legislature included a rider to the biennial budget in its 2002-2003 General Appropriations Act intended to promote choice, independence, and community integration for Medicaid-funded nursing facility residents who expressed a desire to live in the community. Rider 37, as this provision was generally known, allowed Medicaid funds that were being spent for a person in a nursing facility to be transferred to the states community-based care budget when the individual elected to return to the community. The initiative was re-authorized in the 2004-2005 budget as Rider 28 and established in regulation in summer 2005 as the Money Follows the Person (MFP) funding policy. Unlike the Rider initiative, the MFP policy in its current form is now a full-fledged program with peripheral supports in place statewide, including relocation contractors, transition assistance funds, and interdisciplinary transition teams.

The Rider initiative in Texas is noted as an example that could serve as a model for other states with similar objectives. In this study, commissioned by the Office of the Assistant Secretary for Planning and Evaluation (ASPE) of the U.S. Department of Health and Human Services, we examine the experiences in the early years of the Rider initiative using information from site visits to four Texas communities in spring 2004 and analysis of Texas Medicaid long-term care data. As of December 31, 2005, a total of 10,156 people had moved to the community under the Rider initiative and the new MFP program.

Effectively, the Rider initiative provided an alternative eligibility route for the states various Medicaid community care programs. The largest Medicaid community care program in Texas, Community-Based Alternatives (CBA), serves older people and adults with physical disabilities who meet the criteria for nursing facility admission. It provides a comprehensive benefit package of long-term care services, including personal care services, nursing, minor home modifications, and therapies. All but a handful of Rider participants enter the CBA waiver program, so in this report we focus on that program.

Study Objectives

The research questions posed by ASPE fall into three broad areas: the transition process, participant characteristics, and service utilization and costs. The process issues studied include how the initiative works at the state and regional level, eligibility and enrollment processes, the scope of services and benefits, quality monitoring procedures, nursing facility participation in and reactions to the initiative, characteristics of Rider participants, as well as implementation issues and how they were resolved.

Data and Methods

The study included both a qualitative and a quantitative component. The qualitative component consisted of site visits to four counties in three Texas Department of Aging and Disability Services (DADS) service regions. Sites were chosen to reflect the diversity of regions and populations in Texas, and focused on areas with a large number of Rider transitions. Key informants included regional DADS officials and an array of provider and community stakeholders involved in assisting with or managing transitions under the Rider. Discussion topics included outreach and education about the Rider initiative for staff, nursing facility clients, and nursing facilities; the process to transition; procedures for monitoring clients after transition and assuring quality of services; administrative procedures and perceived burden; characteristics of participating clients and their facilities; service capacity in the community; and client satisfaction.

The quantitative component of the study used data from Texas Medicaid long-term care data files to explore how Medicaid recipients who opted to transition to the community through the Rider (Rider participants) compared to their non-Rider counterparts who remained in nursing facilities (nursing facility residents) or who received services through the CBA program (CBA clients). The Rider population studied consisted of 4,870 individuals who entered the CBA program directly after a nursing facility stay. The CBA population studied consisted of 15,895 non-Rider participants entering the CBA program beginning in September 2001 when the Rider took effect. The nursing facility resident sample consisted of 65,132 residents who did not use the Rider and who had at least one Medicaid nursing facility payment in October 2003. We compared these groups on an array of demographics, medical diagnoses, and expected level of care dependence. Within each of the groups and across characteristics, we looked at service use, average monthly cost of services, and individual status after entry into the CBA program. We also report on the proportion of facilities with Rider transitions, the frequency of transitions, and characteristics of facilities with and without transitions.

Eligibility for the CBA program is restricted to adults ages 21 and over. As an extension, this analysis is limited to the same age group. Therefore, findings in this report and policy implications extending from it cannot be generalized to children.

Findings: Overview

Texas initiated a MFP policy by authorizing the transfer of funds from one Medicaid account to another under a rider to its general appropriations bill. Program eligibility and procedures did not change, which allowed this far-reaching change to be accomplished relatively smoothly.

Texass experience with the Rider 37/28 initiative demonstrates that a MFP component can be incorporated relatively easily into an existing Medicaid long-term care program. This far-reaching change in the states approach to long-term care was accomplished through the authorization of an accounting mechanism allowing funds to be shifted provisionally from the nursing facility component of the Medicaid budget to the community care component. In Texas, because community care programs were in place prior to this accounting change, the only change that the Rider made to the Medicaid long-term care program was, in effect, to identify the nursing facility population as a separate eligibility category for community care programs and to make this eligibility group exempt from any caps on enrollment in these programs. The comprehensive nature of the CBA benefits package allowed a diverse population of nursing facility residents to move back to residential settings that were more integrated in the community. In Texas, the accounting provisions embodied in Riders 37 and 28 later became one component in a comprehensive MFP funding policy, which defined a Medicaid long-term care program and incorporated lessons from the Rider initiative about the transition process.

Findings: Process

Outreach

There is both formal and informal outreach to potential participants in the Rider. An array of stakeholders notifies nursing home residents about the Rider and identifies potential clients for transition to the community.

In all three regions where site visits took place, DADS staff caseworkers and social workers at nursing facilities serve key roles in identifying clients who are interested in or suitable for transition to the community. Local ombudsmen and advocates for the elderly and disabled are an important additional source of information; in some regions, they advocate for and assist residents transitions. As the initiative has matured, word of mouth has become a source of information for potential clients.

Steps to a Transition

Steps to a transition include determining eligibility, developing a service plan, and contracting for services in the community prior to the nursing home residents discharge. Eligibility for community programs is determined while the participant is still in the nursing facility, which minimizes gaps in services.

Individuals who are interested in leaving the nursing facility are assigned a DADS caseworker and work with a home health agency to develop an Individual Service Plan (ISP). The residents choice of agencies and input into the ISP is his or her chief opportunity to influence the care plan, as Texas waiver programs do not currently include provisions for consumer direction. The DADS caseworker then arranges for all of the necessary components of the plan to be put in place by the time the client leaves the nursing facility. Eligibility and service coordination tasks are conducted while the client resides in the nursing facility. DADS staff generally felt that the presence of informal supports greatly facilitated the transition process, but that clients without informal supports were able to transition. Estimates of the time from identification of an eligible resident to the day he or she moved back to the community ranged from one to two months up to as long as four to five months.

Issues Affecting a Smooth Transition to the Community

Informants identified three primary issues affecting a smooth transition to the community.

Payment for community-based services cannot be made prior to a clients discharge from the nursing facility, which sometimes complicates the coordination of services needed in place by the clients first day in the community.

An overarching issue affecting smooth transition to the community arises from the fact that the Rider is, in effect, an accounting mechanism that transfers funds from one program budget to another. Clients cannot receive services from a nursing facility and a community care program at the same time, but in order for a participants home to be made ready, home modifications need to be complete and adaptive aids and other supports in place when the client arrives. Changes had to be made in state policy to allow work such as home modifications to be contracted before the client leaves the nursing facility, even though payment for these contracts cannot be made until the clients first day in the community.

Respondents reported that the most difficult challenges to implementing transitions were finding suitable housing, identifying accessible transportation, and finding a physician in the community.

The most difficult challenges reported by site visit informants related to components of community care that are beyond the scope of the CBA program itself. Difficulty finding acceptable housing was identified as a pervasive problem that often delayed transition. The lack of accessible transportation in areas where housing was most affordable was problematic for placing participants. Finding a physician in the community was reported to be most difficult for long-term nursing facility residents or those new to the community, since these clients were least likely to have existing ties to local providers.

Some challenges, such as accessing transportation, are more acute in rural areas, and others, such as the risk of social isolation and finding physician services, are not specific to Rider participants or the CBA program. In general, community capacity for the provision of services to nursing facility residents moving back to the community likely contributes to some of the differences across regions in service use and in the number of transitions observed in this study.

Respondents reported finding reliable contractors for minor home modifications and maintaining quality personal attendant services as the most frequent challenges to a smooth transition.

Certain services presented challenges in implementation. DADS caseworkers and home health agency representatives reported that finding reliable contractors for minor home modifications were the biggest challenges. In addition, a large majority of Rider participants rely on personal attendant services, and these services are a repeated source of problems due to no-shows and high turnover rates among personnel. These problems are not limited to the CBA program and not attributable to the Rider, but need to be resolved if the program is to function smoothly.

Nursing Facility Response to the Rider Initiative

Facility participation is widespread. Transition volume at most facilities is low.

Facility participation in the initiative is widespread with transitions occurring in all regions and in both rural and urban areas, and in many different types of facilities. Transition volume at most facilities is fairly low. Over the three-year study period, a little more than one-third of participating facilities had only one or two transitions, while just under one-third had three to five transitions. The remaining third represented a range of Rider activity, including a few high volume facilities.

This study suggests that although the Rider is significant for individual participants, it is generally not a significant issue for nursing facilities generally. The total number of transitions across the state is small relative to the size of the nursing facility population, and the number from most individual facilities is small. Thus, the direct impact of the Rider initiative on nursing facility occupancy and the larger market for long-term care services, at the time of this study, was small. With the advent of a comprehensive MFP program in 2005 in Texas, the effect on nursing facilities can be expected to increase. The pace of transitions suggests that the changes will be incremental in nature, allowing the nursing facility market to adjust over time.

Quality Monitoring

Quality monitoring under the Rider initiative occurs through the community-based program in which they participate.

Formal quality monitoring under the Rider follows the procedures in place for the waiver programs that the Rider participants enter. Under the CBA, the ISP is reviewed by the caseworker quarterly for appropriateness. DADS is responsible for monitoring the home health agencies, which are, in turn, responsible for monitoring the quality of care provided by their employees. Non-DADS personnel, such as the ombudsman and Meals-on-Wheels volunteers, also assist in quality monitoring, either formally or informally. Day-to-day quality issues are handled through standard DADS complaint procedures, which allow a client to contact their DADS caseworker to report problems. Some advocates and nursing facility personnel expressed discomfort at the lower level of regulation and supervision of assisted living facilities, where a large minority of Rider participants live, compared with nursing facilities or home health agencies.

Findings: Rider Participants and Comparison Groups

Characteristics of Rider Participants

Although Rider participants represent a wide range of nursing facility residents, certain characteristics are associated with a greater likelihood of transition to the community under the Rider.

Rider participants represent a wide range of nursing facility residents in all regions of the state and in both rural and urban areas as well as a wide range of types of disabilities and medical conditions. This breadth of participation suggests that the program has provided opportunities to return to the community for most types of clients.

Rider participants are predominantly elderly and female, reflecting the characteristics of nursing facility residents in general. Hispanics are slightly over-represented in the Rider population, while non-Hispanic Blacks are slightly under-represented as compared to the general nursing facility population. The average length of stay in the nursing facility prior to transition is over seven months, and more than one-third of Rider participants statewide have resided in the facility for six months or more. About one-third of Rider participants are classified as clinically stable with the highest level of functioning and lowest number of limitations in activities of daily living allowed under nursing facility Medicaid eligibility requirements. However, even those needing the most medical resources, such as individuals with quadriplegia or in a coma, are represented among participants.

Younger residents and those in urban counties are more likely to use the Rider than older residents and those residing in rural counties. Those with the lowest care dependence are somewhat over-represented among Rider participants relative to the nursing facility population. The percentage of older Rider participants with diagnosed dementia is far lower compared to nursing facility residents. A higher proportion of Rider participants have a diagnoses of stroke compared to nursing facility residents.

Participant Status After Entry into the CBA Program

Participation in the CBA program is relatively stable among Rider participants. Over 70 percent are still receiving CBA services twelve months after entry into the program.

Over 80 percent of recipients in both the Rider and CBA populations remained in the CBA program in the sixth month after entry into CBA, though slightly fewer Rider participants than CBA clients were still in the program. The proportion remaining on CBA declines over time; nonetheless, 71 percent of Rider participants and 85 percent of CBA clients were still receiving CBA services through the twelfth month, which suggests that the CBA population overall is relatively stable. With a few exceptions and with consideration for omissions in the data, we found that the CBA program appears to be able to provide the opportunity for community living for a very wide range of recipients through the Rider, including those with mental health or behavioral issues and those requiring heavy care. Due to data limitations, no conclusions about hospitalization or mortality rates can be drawn from this study.

Overall, site visit informants evinced enthusiasm for the initiative. However, some informants noted that not all populations are equally well served. Informants identified clients with severe and persistent mental illness, clients with Alzheimers disease, and high-needs clients with physical disabilities as more difficult to transition. Respondents noted particularly that the lack of community systems to properly manage complex medication regimens for the severely mentally ill could be a precursor to return to a nursing facility for some participants.

Service Use

About one-third of Rider participants enter assisted living facilities (ALFs), which affect overall patterns of service use. Assisted living use varies widely across regions and participants level of care dependence.

About one-third of Rider participants enter ALFs, a rate five times higher than CBA clients. ALF use is twice as common in urban areas as in rural areas for Rider participants. Use of ALFs also varies by expected level of care need. About half of those with the lowest expected level of care needs are living in ALFs compared to 2 percent of those needing the highest level of care need. The share of both Rider participants and CBA clients with a nursing facility stay over the six-month period after CBA entry was low, regardless of assisted living facility use.

The difference in use of assisted living drives other observed differences in service use since the services included in the assisted package are not reimbursed separately. When service use is analyzed separately by assisted living status, there are few differences between the Rider and CBA groups in the proportion receiving each service.

Program Expenditures

Rider participants incurred costs in the CBA program that were about 10 percent higher than for non-Rider CBA clients, a difference driven in part by higher use of assisted living but also higher average costs for personal attendant services among those who did not use assisted living.

Despite differences in service mix between Rider participants and CBA clients, total average monthly program expenditures for all services appear to differ by about 10 percent--or $100 per client per month--with total expenditures of $1,143 per client per month for the Rider population compared with $1,043 for the CBA population. Analysis reveals higher average program expenditures for Rider participants across almost all subgroups, suggesting that no subgroup of Rider participants accounts for the observed difference. The difference in expenditures appears to be driven only in part by differences in the use of assisted living. The total average monthly program expenditures for ALF residents are higher than for non-ALF residents for both Rider participants and CBA clients, but a greater share of Rider participants enter assisted living. For recipients who do not use assisted living, Rider participants have a higher average cost for personal attendant services. In general, assisted living appears to be a nearly cost-neutral substitute for other CBA services.

INTRODUCTION

Overview of Rider 37

In part as a response to the 1999 Olmstead v L.C. decision, the Texas state legislature included a rider to the biennial budget in its 2002-2003 General Appropriations Act intended to promote choice, independence, and community integration for Medicaid-funded nursing facility residents who expressed a desire to live in the community. Rider 37, as this provision was generally known, allowed Medicaid funds from the states nursing facility budget to be transferred to the community-based care budget when an individual elects to make a transition to the community.¹ Any interested Medicaid nursing facility resident who met medical and functional eligibility criteria for one of the states community care programs was eligible, regardless of his or her length of stay in the facility under Medicaid. The inclusion of Rider 37 in the appropriations act provided a quick and efficient way for the state to add a money follows the person component to its Medicaid long-term care program. It did not require major restructuring of the Medicaid long-term care program, and officials expected it to be, at a minimum, budget-neutral. The Rider took effect in September 2001.²

As a rider rather than as regulation or a legislated program, this initiative was limited to the period covered by the biennial budget. In the next General Appropriations Act, the legislature included nursing facility transition language to the 2004-2005 biennial budget, this time as Rider 28. Experience with the initiative up to that time led to some modifications, which were incorporated in the Rider 28 language and the additional Riders 7b and 37.³ The modifications under Rider 7b loosened restrictions on the cost of services for an individual, allowing additional community services for current waiver clients up to 133% of the reimbursement rate that would have been paid for the same individual to receive comparable services in an institution over a six-month period.⁴ State officials report that, in practice, this restriction means that expenditures are capped at 133 percent of expenditures over the calculated service plan during a six-month timeframe, but that there is no direct relationship to costs incurred in the nursing facility.⁵ More significantly, the language of the new Rider 37 mandated that Rider 28 transitions to the community not increase the base number of appropriated community care waiver slots, so that once a transition client left the waiver program their waiver slot expired. Thus, the initiative was modified so that nursing facility transitions under the Rider would result in only a temporary expansion of the number of slots available under the various community-based service programs.

These riders (28, 7b, and the new 37) remained in effect through August 2005. In 2005, the 79^th Legislative Session enrolled, and the Governor signed into law, House bill 1867 codifying Rider 28 as the Money Follows the Person (MFP) funding policy. The MFP policy incorporates many peripheral supports statewide, including relocation contractors, transition assistance funds, and interdisciplinary transition teams. In this report, we refer to Texas MFP policy as the Rider initiative in order to distinguish the early period of implementation under Riders 37 and 28, the period of our study, from the more recent period under the 2005 MFP funding policy.

The Rider⁶ states that it is the intent of the legislature that as clients relocate from nursing facilities to community care services funds will be transferred from Nursing Facilities to Community Care Services to cover the cost of the shift in services.⁷ Reconciliation of the states long-term care budget takes place annually, and any unspent dollars for Rider participants under Community Care Services are transferred back to the Nursing Facilities fund.

Due to high demand for services, most of the Community Care programs maintain a list of state residents, called interest lists, who have expressed interest in participating in a program but for whom there are no available program slots.⁸ The largest waiver program in Texas, Community-Based Alternatives (CBA), serves older people and adults (age 21 and over) with physical disabilities who meet the criteria for nursing facility admission. It provides a comprehensive benefit package of long-term care services, including personal care services, nursing, minor home modifications, and therapies. As of March 2004, the program had a monthly average enrollment of 30,000 clients in FY 2003 and an estimated 64,000 people on the interest list.⁹ In comparison, by the end of February 2005, nearly 8,000 people had moved to the community under Riders 37 and 28.¹⁰

Study Objectives

The Office of the Assistant Secretary for Planning and Evaluation (ASPE) of the U.S. Department of Health and Human Services has an interest in better understanding alternative Medicaid financing models such as the Texas Rider initiative. Beginning in FY 2004 and in every budget since then, the Centers for Medicare and Medicaid Services included a five-year, $1.75 billion demonstration project, the Money Follows the Individual Rebalancing Initiative, to assist states in developing and implementing strategies to rebalance their Medicaid long-term care programs to incorporate a cost-effective mix of institutional and community options for care. The stated goal of the Rebalancing Initiative is ...to create a more equitable balance between institutional and community-based services spending; to increase the responsiveness and cost-effectiveness of the system; assist states to fulfill the Americans with Disabilities Act; and increase the amount of control individuals with disabilities are able to exert over service choices.¹¹ The Deficit Reduction Act of 2005 includes funding for state MFP demonstration projects. The Rider initiative in Texas is noted as an example of a unique state experience that could serve as a model for other states with similar objectives.¹²

ASPE commissioned this study of the Texas Rider initiative to learn about the early experience of the program and to help illuminate the strengths and limitations of the approach Texas has taken to implementing a MFP approach in serving a complex population.¹³ As part of this study, the Urban Institute undertook field research to better understand how the initiative works in varied settings in the state. This research was reported in an interim report to ASPE.¹⁴ A second interim report summarized findings from the Urban Institutes analysis of data provided by the Texas Department of Aging and Disability Services (DADS)¹⁵, a state agency overseen by the Texas Health and Human Services Commission.¹⁶

The research questions posed by ASPE fall into five broad areas:¹⁷

Components of the Rider 37 initiative--how the initiative works, eligibility and enrollment, services and benefits, quality monitoring;
Implementation issues and how they were resolved;
Characteristics of Rider 37 participants compared with non-participants in nursing facilities and in the community;
Utilization of program services; and
Costs associated with the Rider.

The next section of this report provides a brief overview of our research methods for both the field research and the quantitative analyses, including a brief description of the limitations of each of these analyses.¹⁸ We then present findings and observations from our qualitative study and quantitative analyses relevant to specific research questions. We conclude with a summary of our findings.

Data and Methods

The qualitative component of the research was based on key informant interviews at four sites in three DADS regions across the state during April and May 2004. Our selection criteria included counties with a high number of transitions, geographic spread across the state, and diversity in regional populations based on ethnicity and income. We included one predominantly rural county in order to look at potential differences in process between rural and urban areas. Based on these criteria, we chose to visit Ft. Worth in Region 3 (Tarrant County, north-central Texas), McAllen in Region 11 (Hidalgo County, on the Texas-Mexico border near the Gulf coast), and Abilene and the nearby rural town, Brownwood, in Region 2 (Taylor County and Brown County, west-central Texas).¹⁹

Discussions with key informants at each site were scheduled in advance of our visits. We met with regional DADS officials (caseworkers, supervisors, Medicaid eligibility staff, and quality control staff), nursing facility administrators, home health and other community providers, the long-term care ombudsman, representatives from the Area Agency on Aging, and local advocacy groups. For each category of key informant with whom we met, we developed a standard discussion guide.²⁰ All informants were promised anonymity to encourage them to express their thoughts freely.

In the quantitative analysis, we used data from Texas Medicaid long-term care data files to explore how Medicaid recipients who opted to transition to the community through the Rider (Rider participants) compared to their non-Rider counterparts who remained in nursing facilities (nursing facility residents) or who were already receiving services through the CBA program (CBA clients).²¹ The Rider population included 4,870 individuals, the nursing facility resident sample included 65,132 individuals, and the CBA sample included 15,895 individuals. Because there are only a small number of Rider participants under the age of 21 and because only adults are eligible for the CBA program, we limited our analysis to recipients who were age 21 and older. (More detailed methodology is provided as an appendix.)

Study Limitations

For the qualitative analysis, we intentionally chose sites where there had been large numbers of transitions from nursing facilities to community settings so that we could investigate how the Rider works on the ground and what had been the experiences of clients who had completed the transition process. However, it is likely that we would have learned a different story had we visited places where there had been few transitions. We might have learned, for example, what constitutes an absolute barrier to Rider implementation, such as resistance to transitions among local nursing facilities or inadequate capacity in the community for providing services to transitioned clients.

The quantitative analysis is limited in that the Texas Medicaid long-term care databases are designed to meet operational rather than research needs. Omission of data on hospitalizations and incomplete information on mortality limit the picture we could assemble on the status of clients who leave the nursing facility for the community. Our analysis is limited to adults age 21 and over. While all nursing facility residents are eligible to use the Rider irrespective of age at the time of our study, few children and adolescents actually had transitioned from nursing facilities using the Rider, as noted above. Therefore, our examination of the Rider and policy implications extending from it are not generalizeable to children. Additional limitations imposed by the data are noted where relevant to specific analyses.

FINDINGS

Components of the Rider 37 Initiative

The Rider does not define a program, but rather it opens another path to eligibility for existing Medicaid community care waiver programs. Site visit informants saw the new eligibility pathway as important because of the long interest list for most of these waiver programs. All expressed enthusiasm for the Rider initiative although most had suggestions for how to make it work better.

At its simplest, the Rider is an accounting mechanism, allowing the transfer of funds from one part of the Medicaid long-term care budget--nursing facility care--to another--community-based care. It builds on existing community care waiver programs, so its implementation is influenced by how these existing programs are functioning. All but a handful of Rider participants enter the CBA waiver program.²²

How the community care programs function in different areas of the state is influenced by community characteristics and local market characteristics. In addition, DADS regions have been allowed a fair amount of flexibility in how they implement the initiative so that the regional or local administrative structure and its working relationships with providers and advocates are also important. Therefore, we consider local socio-demographics, provider capacity, and regional implementation decisions in looking at how the Rider functions in a given community.

In this section, we describe the general procedures that are used under the Rider to move nursing facility residents to the community. We look at eligibility and outreach practices, the transition process, the role of relocation contractors, the services and benefits covered under the CBA program, and procedures in place for quality monitoring. Finally, we consider how this process differs for different types of clients and in rural versus urban areas.

Outreach and Eligibility

When the initiative was new, state and regional DADS officials took steps to inform nursing facilities and nursing facility residents about the program through various forms of outreach. The formal procedures for informing providers and clients included letters to individual nursing facilities outlining the new initiative and the associated procedures. Residents of nursing facilities in many regions also received a letter describing the initiative although, in some regions, such letters were not sent for fear that the resulting number of people seeking to leave nursing facilities would outstrip the regions ability to arrange transitions. In some regions, information about the program is included now in the Medicaid eligibility letter given to nursing facility clients. DADS staff reported that they had not received special training related to the Rider but that information about the initiative was included in their ongoing training.

In the three regions where site visits took place, DADS staff caseworkers and social workers at nursing facilities were reported to play an important role educating nursing facility residents about the initiative. Other organizations and personnel, such as ombudsmen and advocates for the elderly and for people with disabilities are an important additional source of information. In some regions, nursing facilities include a description of the Rider as part of discharge planning for residents at admission to the nursing facility. Over time, word of mouth has become an increasingly important source of information about the initiative for nursing facility residents and other possibly interested clients. No formal efforts are made to inform people on the interest list about the initiative but site visit informants reported that word of mouth has been effective in reaching this population.

Eligibility for entrance into community care programs under the terms of the Rider is limited to nursing facility residents who are eligible for Medicaid. For those potential Rider clients who have not been found Medicaid eligible already, there is a 45-day time limit within which the state must complete its Medicaid eligibility determination. If a nursing facility resident has already been deemed Medicaid eligible before applying for the waiver program, the process could be shorter than 30 days depending on the amount of time required to complete all arrangements for transition. Medical eligibility for Medicaid beneficiaries in nursing facilities is assessed every six months for the first two years of nursing facility care, but after two years, eligibility is assumed. In contrast, DADS reassesses medical and financial eligibility of Rider participants once a year for the duration of participation.²³

Steps to Transition

The first step in transition to the community is the identification and referral to DADS of people who meet the eligibility criteria and are interested in leaving the nursing facility. Nursing facility social workers and local ombudsmen are frequent sources of referrals to DADS for the Rider. Medicaid eligibility workers, who provide residents with information about Medicaid, including the Rider, are another source.

Once a resident has been identified as a potential candidate for the Rider, he or she is assessed for medical and financial eligibility for Medicaid and, if eligible, assigned a DADS caseworker. The caseworker gives the resident a list of participating home health agencies from which to choose and facilitates a meeting between the resident and the chosen agency. The home health agency must meet with the resident within 14 days to assess his or her needs and to develop an Individual Service Plan (ISP). The DADS caseworker arranges for all of the necessary components of the established ISP to be put in place by the time the client leaves the nursing facility. In addition to waiver program services, clients could apply for a one-time Transition to Life in the Community (TLC) grant of up to $2,500 for costs related to transition such as rental deposits, deposits on utility services, or basic kitchen supplies.

The state also has contracts with relocation specialists to assist in the transition process. The relocation specialist contracts began a one-year pilot project in May 2002 in selected areas of the state and were extended through November 2003. After that time, the contracts were expanded statewide. The relocation contractors were charged with outreach to potential transition clients and with assisting with the relocation of a specified number of clients. In addition, they coordinated the TLC grants. In September 2004, reimbursement for such costs was moved to the Medicaid waiver programs as Transition Assistance Services.

The residents choice of agencies and input into the ISP is his or her chief opportunity to influence the care plan. The waiver programs do not currently include provisions for consumer direction, such as a choice of workers or the employment of friends or family members as personal attendants.

There was a broad range of opinions among site visit informants on the role of family or other informal supports in the transition process. Home health agencies take into account informal supports in designing the ISP and noted that a client without informal supports would likely require a greater number of service hours and possibly more intensive care. Advocates contend that lack of family supports should not be a barrier to transition, that the waiver program should provide whatever hours are necessary and not penalize people without family. DADS staff generally felt that the presence of informal supports greatly facilitated the transition process.

Site visit informants said that the time from identification of an eligible resident to that residents transition day varied depending on the services needed, where the person was moving to, and the availability of informal supports such as family or friends to assist in the process. Estimates made by informants ranged from one to two months up to as long as four to five months.

Types of Clients Served

Rider participants represent a wide range of nursing facility residents in all regions of the state and in both rural and urban areas. State and regional staff, the advocacy community, and the relocation contractors all evinced great enthusiasm for the initiative and a commitment to meeting the needs of all clients. The ability of the initiative to assist in meeting the needs of a range of clients in many different settings is one of its strengths.

Site visit informants, nonetheless, reported that the Rider initiative and the associated community-based services do not serve all populations equally well. They noted that serving clients with severe and persistent mental illness or with Alzheimers disease has proved difficult. In the former, medication management has been an issue; in the latter, the perceived difficulty has derived from the changes in client behavior that often occur over the course of the illness.

Respondents also noted that some high-needs clients with physical disabilities can face difficulties in getting their needs met in the community on a long-term basis, and the procedures have been revised to address this issue. For example, once a client has chosen a home health agency, the agency cannot refuse to assess that client. The agency does, however, have the right to refuse to take someone on as a client if they are not certain of their ability to meet the clients needs. In that case, the client must choose another agency. The agencys right of refusal is a modification of original procedures, in response to agency assertions that the requirement to serve any client that chose them could lead them to accept clients whose needs they were ill-prepared to meet and with possibly serious implications for quality of care. Advocates, on the other hand, contend that loosening this requirement has made it harder for residents with higher intensity needs, such as quadriplegics or those who are ventilator-dependent, to find an agency to take them on as a client, skewing program participation toward those with lower care needs and setting up a barrier to transition for high-needs clients.

As will be seen in the discussion below of the medical conditions of Rider participants, people with a wide range of types of disabilities are being served through the Rider initiative. Therefore, it appears that the problem of differential access, if it exists, does not systematically eliminate opportunities to live in the community for the most severely disabled and mentally ill, but can create additional challenges. Nonetheless, differential access by different populations is an area that warrants monitoring so that if systematic difficulties are found steps can be taken to address the identified barriers to access.

At the time of this study, the Rider initiative had been in effect for about three years, so the long-term capacity for the initiative to support community-based living of high-needs populations has yet to be determined. As public support for community-based options increases, the market for long-term care services can be expected to respond through expansion and diversification of services to better serve all clients. The fact that the initiative has now moved from temporary status as a rider to full integration into the Medicaid program may serve to reduce market uncertainty about responding to the growing demand for community-based services.

Role of Relocation Contractors

In 2003, the state undertook a one-year pilot project to provide relocation services and increase community awareness of community-based care alternatives to nursing facility care in 33 counties. DADS expanded the relocation contracts statewide near the end of the study period, in spring 2004, which could ameliorate some identified access problems. Early in the contracts, there was some confusion about the role of the contractors. For example, in McAllen, DADS regional staff members were not widely aware of the existence of the relocation contract as part of the Rider initiative. However, when the contractor was mentioned by name, staff members were quick to say that they had worked with the contractor, particularly when they needed help with difficult cases, but under other, non-state funding.

Many of the relocation contractors have been working with people with disabilities under other, non-Rider funding and have substantial experience in promoting independent living for this population. Our discussions with local DADS staff and the relocation contractors left us with the impression that there had been an implicit division of tasks by type of client, with DADS working with elderly residents and the relocation contractors continuing their work with the younger disabled population. Thus, at the time of this study, the contractors appeared to be filling a niche within the system rather than the broader role envisioned in the pilot program. Nevertheless, the division of responsibilities seemed to be working well and the contractors role has the potential to evolve as they gain experience with the elderly population. DADS also has non-formal contacts with other private organizations, particularly advocacy groups, such as the Centers for Independent Living, United Cerebral Palsy, and the Mental Health Association, whose missions include monitoring access to services for their clients and so overlap to a certain extent with the goals of the Rider initiative.

There are limitations to what the relocation contracts can realistically be expected to achieve. In particular, site visit respondents feared that the small number of contractors spread across the state would limit their impact. The contractor serving McAllen was located in the next town, about 20 minutes away, making interaction with DADS regional staff easy. In contrast, in Abilene and Brownwood, advocates expressed doubts that a relocation contractor assigned to the entire region and based in Lubbock nearly 250 miles away would have much of an impact in rural areas. According to informants, a relocation specialist with such a large area to cover would have a difficult time making the necessary connections locally and identifying the necessary resources in all of the various rural areas to function effectively.

Covered Services and Other Benefits

Rider participants are entitled to the benefits that all Medicaid beneficiaries may receive (i.e., acute care, nursing facility care, and hospitalization). In addition, Rider participants are entitled to all authorized services covered under the community-based care program that they enter. In the case of the CBA program, the ISP sets the authorized level of services such as the number of hours of personal care attendant services, nursing services, specific therapies,²⁴ adaptive aids, and medical supplies. The ISP may also authorize minor home modifications, such as the widening of doors and installation of ramps to accommodate a wheelchair, and adaptive aids or durable medical equipment (DME), such as a hospital bed, to ensure their safety and independence in the community. Rider participants who enter the CBA program may also be authorized to receive adult foster care, respite care, and services in assisted living facilities (ALFs). At the time of our study, all Rider participants were entitled to an unlimited number of prescribed drugs. Since the implementation of Medicare Part D in January 2006, participants in the CBA program who are eligible for Medicare receive drug coverage through that program.

Issues Affecting a Smooth Transition to the Community

Moving someone who requires nursing facility level care from one setting to another involves the coordination of a large number of people, equipment, and supplies. It is, therefore, not surprising that informants identified issues, large and small, affecting a smooth transition to the community. An overarching issue arises from the fact that the Rider is an accounting mechanism that transfers funds from one program budget to another. Clients cannot receive services from a nursing facility and a community care program at the same time, but in order for a participants home to be ready for him or her to move into after leaving the nursing facility, home modifications need to be complete and adaptive aids in place when he or she arrives. In the early months of the Rider initiative, there were no provisions to contract for home modifications or purchase needed supplies before the participants transition day, making it difficult, if not impossible, to have the new living situation ready for occupancy. Changes have been made in the policy with respect to home modifications and DME, now allowing work to be contracted or orders placed before the client leaves the nursing facility. However, payment for these services and supplies still cannot be made by the CBA program until the client has been discharged from the facility.

Issues Related to CBA Services

Informants also identified issues with respect to certain services. Most often mentioned were personal attendant services and minor home modifications. Attendant services are a repeated source of problems; informants cited attendant no-shows and high turnover rates, although many described procedures in place to deal with the problem of no-shows. These problems are, of course, not limited to the CBA program and not attributable to the Rider but, nonetheless, are perceived as affecting the programs capacity to function smoothly for all participants.

Finding reliable and available contractors for minor home modifications was seen as a persistent problem, one that home health agencies did not feel well-equipped to address since such work has not been a standard part of home health services. Assuring the quality of the work was seen as difficult, and many agencies cited frequent problems with shoddy or incomplete work and delays by contractors. Agencies in different regions have taken different approaches to resolving the difficulties with home modifications. One agency had solved the problem by having a home modifications contractor on staff. In Abilene and Brownwood, local college students volunteer to paint houses and do yard work in the homes of transitioners, but most home modifications must meet ADA specifications and are not appropriate for volunteers.²⁵ Rural informants reported that finding contractors was particularly difficult and so volunteers in any capacity were an important part of the effort.

One informant suggested that one way to address the problems with minor home modifications would be to carve this benefit out of home health agencies responsibilities. It was felt that a carve-out to a designated network of contractors would help assure the availability of contractors when needed, would help develop an expertise in the types of modifications frequently needed by this population, and would allow the establishment of a standard quality control system for this service. In addition, monitoring the quality of modifications would be facilitated through the consolidation of the contracts for this work. Because minor home modifications should be completed before a client returns home, the cost of delays in completing such work include the longer duration of continuing nursing facility care prior to transition. Thus, it was suggested that both cost and quality considerations could be well served by a carve-out.

Issues Related to Non-CBA Services

The most difficult issues identified were those that concern components of caring for an individual in the community that were beyond the scope of the CBA program itself. Respondents identified housing, transportation, and a medical home (specifically, a supervising physician for home health services) as among these non-program issues that affect the implementation of the initiative.

Housing was identified by most informants as one of the most difficult issues to resolve. Once an ISP is established and receipt of services is authorized under a Medicaid community-based waiver program, the primary factor determining the length of time it takes to make the transition from nursing home to community is the ease or difficulty in finding a suitable home for the individual. The challenges include identifying a place that is safe given the individuals limitations and that accommodates their physical or mental disabilities, as well as identifying a source of transportation to and from, for example, health care providers, pharmacies, and grocery stores.

Other than family homes and ALFs, affordable housing that met the needs of people at risk of institutionalization appeared to be scarce in all regions. Advocates did not view housing located in areas without accessible transportation as acceptable. However, affordable housing was often outside of densely settled areas and less likely to be accessible to public transportation or within the service area of para-transit services targeted at people with disabilities.

The challenge of finding housing varied by region, with estimates of the time required ranging from one to two months up to as long as four to five months. In Ft. Worth, where moves to ALFs from nursing facilities were common, a smooth transition could take place in as little as 30 days from eligibility determination. In Abilene, site visit informants cited a shortage of available units in ALFs as a critical factor that delays transitions of nursing home residents to the community. Residents with a preference for assisted living or with no other option might have to wait months for a unit to become available. In Brownwood, where ALFs were not available, respondents said that moving in with family would be the only option for many nursing home residents interested in Rider participation.

The long-term provision of community-based medical services requires finding a physician in the community to accept the individual as a patient. While the physician at the nursing facility from which participants transition can prescribe home health services for up to 30 days, clients must find a physician in the community to provide ongoing care, including the prescription of continued home health services. Nursing facility social workers reported that residents with a brief stay in the facility often could return to a physician they had before admission, but long-term residents or those new to the community often required help identifying a physician.

Differences in Rural Areas

Many of the issues discussed here are different in rural than in urban areas; some are more acute, and some are specific to rural areas. Living independently in rural areas can pose special challenges for the elderly and disabled that must be addressed before they can leave a nursing facility and return home. In particular, the risk of social isolation can be a serious problem due to the lower population density in rural areas. Finding volunteers to check on the homebound is an ongoing task. Accessible transportation is critical for independence, but more difficult to arrange in rural areas, since rural public transportation systems are generally less well developed than urban systems, and medical transportation systems may be less well developed. Where public transportation exists, the service varies by county and some were said not to be well suited to the needs of peoples with disabilities. The labor issues in home health agencies noted above can be even greater in rural areas. Informants suggested that allowing family members to serve as attendants could alleviate this problem to some degree but that quality monitoring could then become trickier.

Basic services, such as grocery stores and pharmacies, are more dispersed geographically in rural areas posing particular problems for the less mobile. The problems of finding a physician in rural areas may be greater than in more densely settled areas. Access to specialty care may be hard to arrange since rural specialists are often shared among several towns in a broad service area. Some specialized rehabilitation services may not be available. Problems of provider access, however, affect rural areas generally and are not specific to the CBA program. Nonetheless, arranging necessary care for transitioners to rural areas may be more difficult because of provider shortages. Respondents noted that, despite all of these challenges, many clients who have lived in rural communities all their lives choose to return to their home and communities, and the value of having this choice should not be understated.

Nursing Facility Participation

Rider transitions are occurring in many different types of facilities and in all regions. This wide participation suggests that there is broad knowledge about the program across all types of facilities and that nursing facility residents in all regions are being given the opportunity to transition back to the community.

Statewide, 71 percent of nursing facilities with residents potentially eligible for the Rider had at least one resident who transitioned to the community using the Rider (table 1). The majority of facilities in each region have participated, though regional variation is evident.²⁶ While participation in the initiative has been broad, transition volume at most facilities is fairly low. About two-thirds of the 832 facilities with at least one transition had five or fewer transitions over the first three years that the Rider was in effect, and fewer than 10 percent of facilities had relatively high transition activity with more than ten transitions. Specifically, a little more than one-third of participating facilities had only one or two transitions, while just under one-third had three to five transitions. The remaining third represented a range of Rider activity. Fifteen facilities had 20 or more transitions, with one of these having 87 transitions. The patterns described statewide are observed in both urban and rural counties (table 2).

The size of the population that has opted to use the Rider--nearly 8,000 people since September 2001 and as of March 2004--is small relative to the size of the nursing facility population (approximately 63,000), as well as the population of adults receiving services through the CBA program (approximately 30,000) and the population that has expressed formal interest in the CBA program by placing themselves on the interest list (approximately 64,000). Thus, the direct impact of the Rider initiative on nursing facility occupancy and the larger market for long-term care services, at the time of this study, was small. With the advent of a comprehensive MFP program in 2005, the cumulative total participation from the Rider and the MFP program has climbed to nearly 11,000.²⁷ While the effect on nursing facilities can be expected to increase, the pace of transitions suggests that the changes will be incremental in nature, which would allow the nursing facility market to adjust over time.

The small number of participants using the Rider to access community-based services relative to the large nursing facility resident population reflects both the capacity for completing the process of transition and the demand for transitioning among nursing facility residents. On the one hand, the geographic and facility-level dispersion of the vast majority of Rider participants indicates that nursing facility residents who want to leave and have the resources in their community to do so likely will have this opportunity. On the other hand, our site visit study indicated that the process of assisting nursing facility residents with the move to a community setting is resource intensive with respect to state and regional administrative support services. In addition, the resources of participants, both psychological and familial, can be taxed by the transition. In all likelihood, these constraints affect the number of participants that the initiative can serve, and may also explain the low number of transitions from most facilities.

Quality Monitoring

Formal quality monitoring under the Rider follows the procedures of the waiver programs that Rider participants enter. Under CBA, DADS caseworkers contact clients by phone 30 days after they leave the nursing facility, and again by phone after six months. At one year, they visit the client in person. At the six and twelve-month interviews, a client satisfaction survey is administered to those still in the CBA program. The ISP is established for a year but reviewed by the caseworker quarterly for appropriateness.

DADS is responsible for monitoring the home health agencies, which are, in turn, responsible for monitoring the quality of care provided by their employees. At the beginning of the Rider initiative, home health agency contracts for the CBA waiver program were open-ended. By the time of our site visit, DADS had begun letting contracts for one year at time. DADS staff reported that the defined contract period makes it easier to deal with problem agencies, and there have been cases of contracts that have not been renewed. DADS staff monitors agency performance by sampling client records to monitor adherence to process regulations. Monitoring visits to the agencies are made in the tenth month of each contract. Client satisfaction interviews are conducted either by telephone or face-to-face annually for all CBA clients.²⁸

Non-DADS personnel also assist in quality monitoring, either formally or informally. Some nursing facility social workers make follow-up calls to residents after transition. Many make themselves available for phone calls from former residents post-transition, but they report receiving few calls. In a more formal process, volunteers in Abilene with the Meals-on-Wheels program provide an additional level of monitoring to rural CBA clients. They record any problems observed or reported by the clients they visited and fax the report to DADS regional staff and caseworkers for review and follow-up.

Most day-to-day issues with quality of service are handled through the DADS complaint procedures, which allow a client to contact their DADS caseworker to report problems. The caseworker is required to follow-up immediately with the service provider and assist in a resolution. Clients often use this means to report that an attendant has not shown up or to report specific problems with an attendant. Informants noted that complaints often arise out of disputes between attendants and clients over the attendants tasks and responsibilities. These cases can usually be resolved by the caseworker through client education on the attendants responsibilities.

A substantial minority of Rider participants live in ALFs. Some advocates and nursing facility personnel expressed discomfort at the lower level of regulation and supervision of these facilities compared with nursing facilities or home health agencies. Nursing facilities, particularly, expressed concern about the lack of a level playing field between community-based services and nursing facility care with respect to regulation, particularly in the area of quality assurance. Partly in response to such concerns, statewide monitoring of ALFs began in March 2004.²⁹

Quality monitoring processes are evolving as community-based care becomes more common. DADS staff noted that some areas of quality assurance were under review and that new procedures could be expected. Added oversight might improve quality but will almost certainly add costs. With greater opportunities for independence come greater opportunities for quality problems. In the new order of consumer choice in long-term care services, the state will have to decide how much responsibility it must assume for the quality of the services in the different places that clients choose to receive services.

Rider Participants and Comparison Groups

In order to help identify which types of residents are more likely to use the Rider to move to the community, we compared Rider participants with nursing facility residents. Large differences in the make-up of the Rider population as compared with the nursing facility population can point to groups for whom transition might be more difficult. Qualitative information gathered in the site visits helped illuminate possible mechanisms behind the differences thus identified.

In our comparisons between Rider participants and CBA clients, we focus on service use. Overall service use in each group is driven by differences in the composition of the groups but also by differences in service use between similar people within each group. We considered both of these factors here. Difference in service use also point to possible gaps in community-based services. Again, information from the site visits helps provide context for the quantitative findings.

Characteristics of Rider Participants

Demographics

Rider participants are predominantly elderly, with only one-third under the age of 65 (table 3). About two-thirds of Rider participants are female, reflecting the over-representation of women among the elderly. Just under one-third of Rider participants live in rural areas. About two-thirds of Rider participants are White and non-Hispanic in ethnic origin, about one-fifth are Hispanic, and one-tenth are non-Hispanic Black.

Nursing facility tenure

Analysis of Rider participants length of stay in a nursing facility suggests that the Rider initiative is reaching both long-term and short-term residents of nursing facilities. The average length of stay in the nursing facility prior to transition is over seven months with more than one-third of Rider participants statewide residing in the facility for six months or more. About one in ten Rider participants had a nursing facility stay of less than 40 days.

Expected level of care dependence

All recipients of long-term care services under Medicaid are classified based on the Texas Index for Level of Effort (TILE) that summarizes the clients care needs and medical stability and is the basis for nursing facility reimbursement under Medicaid. The TILE classes are based on medical diagnosis, limitations in activities of daily living, expected level of care needed, the presence of mental or behavioral conditions that affect staffing resources required, and receipt of rehabilitation services. Presence of family support or informal caregivers does not influence TILE determination.

TILE classes range from 211 to 201. A lower TILE number indicates a higher level of care or greater staffing resources needed to meet the clients needs and a higher reimbursement to nursing facilities.³⁰ For ease of analysis, we condensed some TILE classes with proximate payment rates, based on the assumption that grouping individuals by the resources required to support an individual would allow the most policy-relevant comparisons.³¹

There are Rider participants in all TILE classes (table 4). About one-third are in TILE 211, a clinically stable group with the highest level of functioning and lowest number of limitations (table 4). Another 8 percent are stable with the highest level of functioning, but have a mental/behavioral condition requiring additional daily staff intervention (TILE 210). People with moderate care needs (TILE classes 207-209) make up just over one-quarter of Rider participants. Almost one in five are classified as receiving rehabilitation services (TILE 202). Fewer than 10 percent are classified as having high care needs (TILEs 204-206), and fewer than 5 percent require heavy care (TILEs 201 and 203). Variation between urban and rural areas in the distribution of TILE classes suggests the rural participants are somewhat more care dependent and have a greater number of ADL limitations, although the differences are not large (data not shown).

Diagnoses of physical or mental illness

In addition to TILE classification, each recipient can be assigned up to five concurrent diagnoses per claim based on ICD-9 classification.³² We present information here on the proportion of each group that had a particular diagnosis as one of his or her five diagnoses (table 5). ICD-9 codes are grouped into their major classification of diagnoses, except for stroke, a condition that often requires rehabilitation.³³ Where applicable, residents are also assigned a diagnosis of a behavioral condition or mental illness. Specific groups of behavioral and mental illnesses were also identified based on their prevalence or potential policy relevance.³⁴

The diagnoses of physical illness among Rider participants, with only a couple of exceptions, vary little by age, and these variations are as would be expected in the population. Over 40 percent of Rider participants have a diagnosed mental illness. The most common mental illness diagnosis is depression among those younger than 65 (27 percent) and dementia among those over 65 (20 percent).

Characteristics of Rider Participants as Compared to Nursing Facility Residents

A comparison of the characteristics of Rider participants with those of nursing facility residents can help policy-makers understand which types of long-term care clients are more likely to access community care through the Rider initiative. Here we present notable differences between Rider participants and nursing facility residents in demographic characteristics, TILE class and physical and mental/behavioral diagnoses. We then consider differences across regions and between rural and urban areas.³⁵

Demographics

With a few exceptions, Rider participant characteristics reflect the characteristics of nursing facility residents, the population from which they transition (table 3). Nursing facility residents are generally older than Rider participants, and half as many are under the age of 65. Hispanics are slightly over-represented in the Rider population compared to the nursing facility population, while non-Hispanic Blacks are slightly under-represented.

TILE class

Rider participants are more likely than nursing facility residents to be in the lowest care need category (TILE 211) (table 4). A similar share (between 10 and 20 percent) of Rider participants and nursing home residents are receiving rehabilitation services (TILE 202), possibly reflecting a medical event that led to nursing facility admission. Somewhat surprisingly, the proportion of each group classed as high care (TILEs 204-206) is similar at about 7 percent. However, as would be expected, those classified as needing heavy care (TILEs 201 and 203) are found more often in nursing facilities (5 percent) than among Rider participants (3 percent).

Physical illness

Reported diagnoses support the association of medical events and nursing facility admission among Rider participants. Rider participants in both age categories (over or under age 65) are more likely than nursing home residents to have a diagnosis of stroke or cardiovascular disease other than stroke (table 5). Rider participants are also more likely to have a diagnosis of respiratory system diseases such as pneumonia, influenza, and bronchitis than all nursing facility residents. Rider participants and nursing facility residents in both age groups have similar diagnosis profiles although there are differences in the prevalence of conditions between the two groups.

Mental illness

The mental illness profile of Rider participants resembles that of nursing facility residents, but, while the pattern is similar, in all but one instance the prevalence of each diagnosis is greater among nursing facility residents than among Rider participants (table 5). The one exception is the higher prevalence of diagnosed depression among younger Rider participants than among younger nursing facility residents. The largest difference is seen in the percentage with a diagnosis of dementia. In both age groups, the percentage of nursing facility residents with diagnosed dementia is about 50 percent higher than the percentage of Rider participants, thus supporting the perception among site visit informants that these residents are more difficult to transition.

Geographic differences

Site visit informants in different regions offered different observations about the characteristics of those who transitioned. In Abilene and Brownwood, informants described the majority of the Rider participant population as elderly but without catastrophic illness, typically with a nursing facility stay of under six months. In Ft. Worth, where assisted living placements were much more common, informants reported relatively more middle-aged disabled clients with no family support. In McAllen, the DADS regional office reported that their clients were the elderly. These site visit comments are generally supported by the data discussed here.

In general, there are differences in the rural/urban distribution of Medicaid long-term care recipients within regions. For example, in Region 1, the majority of nursing facility residents (54 percent) live in rural areas while the majority of Rider participants (61 percent) live in urban areas. In Region 5, the trend is reversed with rural area residents disproportionately represented among Rider participants (76 percent) relative to nursing facility residents (59 percent). Overall, our analysis shows variation in the characteristics of recipients across regions and in rural versus urban counties, which suggests possible geographic differences in demand for services, community capacity for services, or differences in implementation of the Rider (data not shown).³⁶

Rider Participants Place of Residence in the Community

The variable indicating where Rider participants and other CBA clients lived once they entered the CBA program was not made available to us in the data set we used for the quantitative analysis. The service use data, however, did include information on the use of assisted living by Rider participants and CBA clients.³⁷ Therefore, here we cite DADS statistics on where Rider participants live followed by a discussion of trends in the use of ALFs based on our own analysis. The differences in ALF use are particularly important since, as will be discussed later, these differences are associated with differences in service use and costs between the two groups.

DADS statistics³⁸ for Rider participants show that 19 percent are living in the community alone, 34 percent in alternative residential care settings such as assisted living or group homes, 42 percent living with family, and the rest (about 4 percent) in adult foster care or in the community with other waiver program participants. These living arrangements vary markedly across the state, as can be seen in the statistics for the DADS service regions we visited. In Region 2 (headquartered in Abilene), clients who had moved to the community were almost evenly divided among living alone, alternative residential care, and living with family. In contrast, in Region 3 (Dallas/Ft. Worth), over half (54 percent) of Rider clients were in alternative residential care, and, in Region 11 (McAllen), 62 percent of Rider clients were living with family.

In our quantitative analysis, we found large differences between Rider participants and non-Rider CBA clients in the use of assisted living, with Rider participants more than five times likelier to enter an ALF than CBA clients (see table 6). About one-third (34 percent) of Rider participants statewide enter ALFs, a much higher percentage than CBA clients (6 percent). ALF use is twice as common in urban areas (40 percent) as in rural areas (19 percent) for Rider participants, with a similar pattern though at a much lower level for CBA clients (8 versus 4 percent, respectively).

Among Rider participants, use of ALFs also varies by TILE class. About half of those assigned the highest TILE categories (indicating the lowest expected level of care needs), are living in ALFs. In comparison, less than one-third (27 percent) of Rider participants with TILE 202 (rehabilitation) and less than one-quarter (22 percent) needing moderate care live in ALFs. Some Rider participants classified as needing heavy care live in ALFs, but this arrangement is relatively rare (2 percent).

The greater reliance on assisted living by Rider participants compared to CBA clients persists across other subgroups as well, including regions, age groups, and racial/ethnic groups. These comparative statistics do not allow us to say what influences a participants choice of living arrangement. The three regions with majority Hispanic populations, Regions 8, 10, and 11, include both the highest and the lowest prevalence of ALF use. Statewide, rural areas show lower use of ALFs than urban areas but the regions with the lowest proportion of rural residents (all under 10 percent) vary in the percentage of their Rider participants using assisted living from 29 percent to 52 percent. Regions that have a high share of rural residents (Regions 1, 2/9, 4, and 5) also show a range of ALF service use from 25 to 49 percent.

Differences by region might be driven by the supply of assisted living, which varies dramatically across the state. For example, the state reports over 100 Medicaid contracts for assisted living in Region 3, but only four such contracts in Region 11. Regional differences in ALF use among CBA clients are small, and the differences do not follow the regional distribution of use among Rider participants, suggesting that there are other important influences in addition to supply. Multivariate analysis would be required to sort out the effects of such factors as supply of ALF units, cultural preferences, and care needs.

Site visit informants suggested that some Rider participants choose assisted living in part because of the limited options available to individuals with little family support or high-cost ISPs that would have unacceptably high costs for CBA if they did not enter assisted living. Returning to the community living situation they had prior to entering the nursing facility might not be possible for many Rider participants, especially those who have had an extended facility stay and whose family home may no longer be available and/or whose informal support network may have dispersed.

Service Use of Rider Participants and Other CBA Clients

In this section, we present findings on the share of Rider participants and CBA clients that received various services available through the CBA program during the first six months after entering CBA first for all participants, and then by geographic area, racial/ethnic differences, and cognitive/functional status.³⁹

The data show that, statewide, Rider participants are less likely than non-Rider CBA clients to receive almost all services. The exceptions are nursing services, specific therapies (physical, occupational, and speech), and adult foster care. Nursing services are necessary to administer ISPs and nearly all clients (more than 95 percent) in both groups receive these services, while very few clients (under 2 percent) receive specific therapies and adult foster care. Nursing facility stays are roughly similar between CBA clients and Rider participants.⁴⁰ Only 4 percent of Rider participants and 5 percent of non-Rider CBA clients appear to have been admitted to a nursing facility during their first six months in CBA (table 7).

As noted above, Rider participants are much more likely than CBA clients to be residents of ALFs. Most Medicaid-covered or CBA-covered services including nursing services, adaptive aids, medical supplies, and therapies, are available to both ALF and non-ALF clients.⁴¹ However, some services, such as meal delivery and emergency response services (ERS), are included in the services provided by ALFs and are not billed separately for residents of ALFs. Other services, such as minor home modifications, personal assistance services (PAS), and adult foster care are not financed under CBA for clients in assisted living. Given the stark difference in ALF use between Rider and CBA clients, notable differences in the use of other services is expected in the full populations (table 7). Therefore, we look at service patterns separately for ALF and non-ALF residents (table 8).

When clients in assisted living are excluded, there are few differences between the Rider and CBA groups in the proportion receiving each service. (There are, however, differences between the two groups in the intensity of service use, as discussed in the section on expenditures.) For clients not residing in assisted living, PAS play a critical role in the ISP according to service claims data, with 88 percent of Rider participants and 91 percent of CBA clients using personal attendants. Adaptive aids and medical supplies are used by about half of the non-ALF clients in each population, and home modifications are provided to just over one-third of non-ALF Rider participants and CBA clients.

Geographic and racial/ethnic differences

Among clients not residing in ALFs, Rider participants and CBA clients in rural areas are more likely to receive some services than are those in urban areas. Although the differences are generally small, the pattern is consistent. Both Rider participants and CBA clients in rural areas are somewhat more likely to receive ERS, meal delivery, and adaptive aids. ERS and meal delivery services may be particularly important links to the community for clients in more rural areas where the broader social service networks may be less developed and clients may be at greater risk of isolation. The difference in adaptive aids may be related to the greater frailty of rural participants as evidenced by their TILE class assignment noted above.

There are also differences in service use among non-ALF residents across regions (table 9). For example, use of meal delivery service for both Rider participants and CBA clients in Region 2/9 (29 and 39 percent, respectively) and Region 4 (31 and 28 percent, respectively) is much higher than in other regions. Regional differences might reflect variation in the availability of community-based services. Caseworkers in Abilene (Region 2/9) described a large network that served the dual purpose of meal delivery and monitoring residents for personal safety on behalf of the regional DADS office. There is also wide regional variation in the receipt of minor home modifications, with use ranging from 23 percent of Rider participants and 20 percent of CBA clients in Region 3 to 55 percent of Rider participants and 60 percent of CBA clients in Region 11.

There are notable differences in service use among non-ALF residents between Hispanics and other groups. For example, Hispanics have lower rates of use for ERS and meal delivery compared to non-Hispanics of either race. This difference might be explained by a greater reliance on extended families for support among Hispanics receiving community-based services, as suggested in our site visits. Hispanics are more likely to receive minor home modifications, a difference that is less easily explained, but which is consistent with the high utilization of this service in Region 11, which has a high concentration of Hispanic residents (table 10).

Differences by functional and cognitive status

If the differences in service use between Rider participants and CBA clients were driven by differences in the use of ALF, we would expect to see the greatest leveling of differences in those TILE categories with the greatest difference in ALF use. The greatest difference in ALF use is seen in TILE category 211, which has a high prevalence of ALF use (table 11). For the most frequently used services (PAS, adaptive aids, medical supplies, minor home modifications, ERS, and meals), the differences in use between Rider participants and CBA clients in TILE 211 ranged from a high of 42 percentage points for PAS to a low of nine percentage points for medical supplies, when both ALF and non-ALF recipients are considered. When ALF recipients are excluded, these differences decline to seven percentage points for PAS and three percentage points for medical supplies. This finding suggests that although ALF use explains a large part of the observed differences in service use between Rider participants and CBA clients, other differences remain in service use patterns.

Service mix by TILE class for non-ALF residents generally demonstrates patterns consistent with expectations about the relative demand for services between groups of varying levels of care dependence (table 12). Differences in service use within a TILE class between Rider participants and CBA clients are likely driven by either a higher demand for service by one group based on greater frailty or a difference in the process by which services are assigned for Rider participants compared to CBA clients. Better data on frailty and mobility (e.g., specificity about ADL limitations) and family support analyzed in a multivariate framework, as well as further investigation at the local level on how services plans are established for each group, would contribute to a better understanding of the reasons for observed differences in service mix for these two populations. Our conversations with DADS officials suggest that assignment of TILE classes for nursing facility residents and, hence, for Rider participants, are driven by different payment incentives than for CBA clients,⁴² which could contribute to the observed differences.

Recipient Status after CBA Entry

Transition from a nursing facility is not the end goal of the Rider initiative. Rather, the initiative seeks to enable Medicaid long-term care recipients to live in the least restrictive setting possible, in accordance with the Olmstead decision. Therefore, it is important to understand how Rider participants fare once they leave the nursing facility. Here we consider measures of recipient status six and twelve months after leaving the nursing facility.⁴³ We compare Rider participants to CBA clients in order to determine how these two similarly frail populations fare after entering the CBA program. We look at the share of each group that remained in the CBA program, that was residing in a nursing facility, that were reported as deceased, and that had left the CBA program and whose status was unknown. We note important differences across functional and cognitive status groups represented by TILE classes.⁴⁴

Over 80 percent of both the Rider and CBA populations remained in the CBA program in the sixth month after entry into CBA. Program retention was slightly lower among Rider participants than among CBA clients (table 13). While the proportion remaining on CBA declined over time, 71 percent of Rider participants and 85 percent of CBA clients were still receiving CBA services in the twelfth month. Very few members of either group statewide were residing in a nursing facility in the twelfth month after entry into the CBA program. These findings suggest that the CBA population overall is relatively stable.

Mortality

By the twelfth month after entering the CBA program, 10 percent of Rider participants and less than 1 percent of CBA clients were reported to have deceased. Since a larger share of CBA clients was classified in less care-dependent TILE class relative to Rider participants, it is reasonable to expect a lower mortality rate among CBA clients. However, the share deceased represents a doubling of the share at six months for Rider participants, but no change for CBA clients, a plateau that is contrary to expectations and indicative of a possible data quality problem.⁴⁵

Similarly, the share of clients that appears to be no longer receiving any Medicaid service and therefore has unknown status by the twelfth month after program entry is 19 percent among Rider participants and 7 percent among CBA clients. Factors causing program attrition cannot be ascertained from the data. As with mortality reports, these shares represent a doubling of the share with unknown status at six months for Rider participants, but no change for CBA clients. State officials expressed concern that not all incidence of mortality has been captured in the database.⁴⁶ Thus, it is likely that some share with unknown status has deceased, and no firm conclusions about mortality rates should be drawn.

Comparisons across TILE classes

These patterns were consistent across regions and for Rider and CBA participants of different ages. As might be expected, the most notable differences were found by TILE class (table 14). Among Rider participants, those classified with high care needs (TILE 204-206) were the least likely to remain on CBA and most likely to have deceased by the end of the first year on CBA, with 66 percent of this TILE class still in CBA and 17 percent deceased. In addition, a high proportion (25 percent) of TILE 202 (rehabilitation) Rider participants had left the program with unknown status at twelve months. With these exceptions and with consideration for omissions in the data, we found that the CBA program appears to be able to provide the opportunity for community living for a very wide range of recipients through the Rider.

The experience of participants in two TILE classes is illuminating. First, although, as noted above, nursing facility residents in TILE class 210 (stable, with mental/ behavioral conditions) were less likely to have left the nursing facility under the Rider, Rider participants in this TILE class are about as likely as the average Rider participant to have remained in CBA at twelve months, 72 percent for TILE 210 (table 14) versus 71 percent of all TILE classes (table 13). This finding suggests that many recipients in this TILE class are suitable candidates for community-based care. Second, as we also noted above, nursing facility residents needing high care (TILE 204-206) were less likely to have left the nursing facility under the Rider and constitute a very small share of all Rider participants. Despite the initial challenge of finding suitable arrangements outside a nursing facility, the share of Rider participants in this TILE class remaining on CBA at the end of twelve months (74 percent) was comparable to retention for less care-dependent classes.

Other status findings

Information collected during the site visits provides perspectives from DADS caseworkers, ombudsmen, and nursing facility social workers who had witnessed first hand a variety of transitions. Overall, there was widespread agreement among informants in all of the sites visited that the majority of clients who had moved to the community were better off than they had been in the nursing facility. However, informants also characterized the circumstances of transitions for individuals as varying widely.

Although caseworkers and advocates cited the benefits of transitions to the community, such as improvement in clients mental health or outlook, not all assessments of the transition experience were positive. Nursing facility staff and some ombudsmen were more circumspect. Some recounted stories of specific clients who were not successful in transitioning. Lack of community systems to properly manage complex medication regimens for the severely mentally ill was cited as a problem and a precursor to return to a nursing facility for some residents who transitioned to ALFs or into other residential situations in the community. The risk of social isolation was also cited as a problem for the elderly and physically impaired that could be particularly acute for clients who live alone in sparsely settled rural areas. However, isolation alone did not necessarily appear to lead to a return to a nursing facility.

Provider opinions on the problems experienced in transitions frequently reflected a lack of confidence in other providers to meet the needs of clients. For example, nursing facilities were more likely to express concern that ALFs were accepting clients beyond their capability to serve, while advocates pointed to the individuals need for privacy and independence that could not be met in a nursing facility. Nursing facility administrators pointed to the much lower level of supervision and regulation for quality monitoring in community care settings and asserted that the resulting environment left open the possibility of quality of care problems going unrecognized once the resident had left the nursing facility. Confirmation of these perspective was beyond the scope of this project.

Nursing Facility Industry Response to the Initiative

There was some concern as the Rider initiative began that the nursing facility industry would not cooperate in an initiative that would likely reduce its share of the long-term care market, particularly in Texas, where nursing facility occupancy rates are relatively low (73 percent, on average, in 2003⁴⁷). The range of possible negative nursing facility reactions to the program include active resistance to allowing workers to contact nursing facility residents about participation, indifference reflected in a lack of cooperation but with no hostility to the initiative, and reluctant participation. The nature of our qualitative research made identification of these reactions unlikely (see limitations of the study, above), although the quantitative analysis shows broad nursing facility participation, suggesting that such negative reactions were not widespread.

We did identify two distinct positive reactions in our qualitative research. The first appears to contribute to the goal of the program to expand the range of long-term care options to include less restrictive settings backed up by nursing facility care. Specifically, some nursing facilities reported enthusiasm for the program and said that they had incorporated the program into their business plan. Their rationale was that if elderly or disabled community residents knew that a nursing facility stay could be temporary, they might be less reluctant to enter for rehabilitation or respite. Some nursing facilities reported marketing this aspect of their care in connection with the Rider.

The other type of positive reaction implies a broader interpretation of the aims of the initiative. One nursing facility reported arrangements with an ALF to admit their residents who had spent down and could no longer afford the facility fees, and then to secure their enrollment in the CBA program via the Rider so that they could return to the ALF with Medicaid financing. Neither regional staff nor the nursing facility involved thought that this bypass of the CBA waiting list, as they referred to it, was against either the letter or the spirit of the rules of the Rider. Indeed, they thought that this use of the Rider was consistent with the aims of the program to expand the use of community-based care and, more particularly, that it was in the best interests of the client.⁴⁸

While for individual participants, the existence of the Rider is significant, the Rider does not appear to be a significant issue for nursing facilities generally. The qualitative and quantitative data from this study do not suggest that there is active resistance by nursing facilities to the initiative. Furthermore, the total number of transitions across the state is small relative to the size of the nursing facility population, and the number from most individual facilities is small.

Service Costs and Program Expenditures for Medicaid Long-Term Care

In this section, we examine average service costs and associated program expenditures for Medicaid long-term care services provided through the CBA program for Rider participants and compare these costs and expenditures to those for non-Rider CBA clients.⁴⁹ Since we found that assisted living accounts for a large part of the difference in service mix between the two populations and since the CBA program does not cover certain services for recipients in ALFs, we examine costs separately for recipients who resided in an ALF at some point in the first six months and those who did not to determine if the subgroup with assisted living accounts for any observed difference in average costs. Because costs are likely to vary by the level of care needed, we look at costs by functional and cognitive status. Finally, we explore how overall Medicaid long-term care expenditures might be affected by changes in demand for Medicaid nursing facility services as a result of the Rider initiative.

Comparison of Long-Term Care Costs for Rider and CBA Clients

Despite differences in service mix described earlier, total average monthly program expenditures for all services appear to differ by only about $100 per client per month, with total expenditures of $1,143 per client per month for the Rider population compared with $1,043 for the CBA population (table 15). Average program expenditures are higher for Rider participants than for CBA clients across almost all subgroups (TILE class, rural/urban residence, age, and race/ethnicity), suggesting that no subgroup of Rider participants accounts for the difference in average program expenditures per recipient between Rider participants and CBA clients.

The primary reason that Rider participants are more costly on average than CBA clients rests with differences in the costs and service patterns of assisted living residents. Average monthly cost among assisted living residents is about $50 higher for Rider participants than for CBA clients. In addition, Rider participants are more than five times more likely to use assisted living. As a result, the average monthly program expenditure for assisted living is over six times greater for Rider participants compared with CBA clients ($348 versus $56, respectively).

Comparison of Long-Term Care Costs by Assisted Living Status

Most of the difference in expenditures between Rider participants and CBA clients can be attributed to the service restrictions for ALF residents. However, since total expenditures for Rider participants are only slightly higher than for CBA clients, it is plausible that assisted living is a nearly cost-neutral substitute for these restricted services. Differences in care needs as reflected in the different TILE profiles of the two groups and in the timing of entry into assisted living contribute much smaller shares to the difference in expenditures.⁵⁰

Table 16 presents average monthly per user service costs and program expenditures for recipients with ALF services in the first six months. The average monthly per user cost for ALF services among recipients with any ALF payments in the first six months after CBA entry is slightly higher for Rider participants ($1,152) than for CBA clients ($1,101). Among ALF residents, the average monthly program expenditures for all services was $1,241 for Rider participants and $1,195 for CBA clients. Table 17 shows cost and expenditure estimates for non-ALF residents. The average monthly program expenditures for all services is $1,089 for Rider participants and $1,034 for CBA clients. Almost all of the difference per recipient appears to be explained by the higher cost per PAS user.

In summary, ALF residents have higher average expenditures than non-ALF residents, and a larger share of Rider participants use assisted living than CBA clients. The primary factors explaining the higher total average monthly program expenditures for Rider participants compared to CBA clients are the greater reliance on assisted living among Rider participants, the higher cost of ALF services per user among Rider participants, and, for recipients who do not use assisted living, the greater intensity of PAS use among Rider participants as compared to CBA clients.

Comparison of Long-Term Care Costs by Functional and Cognitive Status

Some differences in expenditures among subgroups were noted. The differences were greatest by functional and cognitive status, and TILE classes. Table 18 shows average monthly program expenditures by service for each TILE group. As one would expect, recipients with fewer ADL limitations and lower care needs appear to have less intensive service needs and thus lower average costs in the community. Recipients in TILE 211 show the lowest average monthly expenditures ($973 for Rider participants and $909 for CBA clients), while recipients needing heavy care and with the most ADL limitations (TILEs 201 and 203) show the highest average monthly costs ($1,617 for Rider participants and $1,620 for CBA clients).

In most TILE classes, Rider participants have higher average monthly expenditures than CBA clients. CBA clients have higher average monthly expenditures for the TILE group with mental/behavioral conditions (TILE 210) and those with heavy care requirements (TILE 201, 203). (These TILE groups combined comprise less than 20 percent of each population.) The largest difference in expenditures between Rider participants and CBA clients is seen in the group in rehabilitation (TILE 202), with the average expenditure per Rider participant $267 more than the per CBA client. This group accounts for 17 percent of Rider participants but only 1 percent of CBA clients and, thus, contributes significantly to the overall higher average expenditures seen in the Rider group.

The differences across TILE groups is affected by whether the recipient resides in an ALF or not. Since a larger share of Rider participants than CBA clients resides in assisted living across all TILE groups, average monthly program expenditures for recipients who use ALFs are higher for Rider participants than for CBA clients. However, among recipients in TILE 211 who do not use assisted living, Rider participants have lower expenditures than CBA clients (data not shown). In the TILE 207-209 group, expenditures for recipients are higher for Rider participants than for CBA clients regardless of assisted living use (data not shown).

Change in Demand for Nursing Facility Admission in Response to the Rider

There has been some concern among policy-makers that the existence of the Rider could influence demand for Medicaid-financed long-term care. Specifically, there might be people who would use Medicaid-financed long-term care services if they were available in the community, but who do not want to live in a nursing facility. In the face of what has at times been a one to two year wait on the CBA interest list, some people might be willing to go into a nursing facility temporarily knowing that they would likely be able to leave using the Rider and receive services in the community. If this shift in demand were to occur, it would lead to higher overall Medicaid long-term care expenditures as a result of the increased use of nursing facility care, even if only for short stays. In addition, there would also be an increase in people on the CBA program with no commensurate reduction in the number of nursing facility residents.

We conducted an exploratory assessment of any evidence from the site visit and quantitative analysis that the demand for long-term care services has shifted as the result of the Rider. We developed hypotheses of how such a change in demand might occur based on reports by site visit informants and identified four indicators available in the data that might represent a demand effect. These indicators are: (1) a relatively short stay in the nursing facility prior to transition to the community, (2) a relatively low care dependence TILE class at nursing facility entry, (3) use of rehabilitation services, and (4) entry into an ALF within 40 days after CBA entry.

None of these indicators alone should be construed as evidence of a demand effect of the Rider. It is likely that each of the indicators simply reflects the variation in the needs of the nursing facility population from which the Rider population is drawn. However, in combination, these indicators might be suggestive of the possible influence of the Rider on demand for Medicaid-funded long-term care. Since there are other explanations for each of these indicators, it is likely that they would indicate an upper bound on this effect.

Our analysis revealed little evidence of an increase in nursing facility admissions as a conduit to CBA services attributable to implementation of the Rider. Depending on the combination of indicators used and the subgroup of Rider participants examined, we identified a small number of nursing facility admissions possibly attributable to the existence of the Rider, affecting from 1 to 8 percent of participants.

Table 19 shows the percentage of Rider clients meeting each of the four criteria statewide and for urban and rural counties. A rapid turnaround for a client with relatively low need might signal that the client entered the facility in order to gain access to the CBA program. However, in only 3 percent of Rider cases did a participant have an assigned TILE class of 211 and a nursing facility of less than 40 days. A rapid turnaround from community to nursing facility and back to community for a client who moves into assisted living might give credence to the suggestion that a Rider participant had entered a nursing facility to later gain Medicaid financing for services in assisted living. However, only 3 percent of participants both had a nursing facility stay of less than 40 days and entered an ALF within 30 days of CBA program entry.⁵¹ Finally, a rapid turnaround for a Rider participant who is receiving rehabilitation services may give credence to the suggestion that some patients discharged from the hospital to a nursing facility and then to CBA under the provisions of the Rider might otherwise have been discharged to the community without Medicaid funded services. The hospital would thus be an inadvertent access point for CBA via the Rider. However, a very small proportion of participants (about 1 percent) had both a short nursing facility stay and a TILE class of 202.

Since the combined criteria are not mutually exclusive, some Rider participants may meet more than one of the criteria. The last line of table 12 shows the percentage of Rider participants who meet any of the three criteria in combination with a stay less than 40 days. These numbers can be viewed as one measure of the upper bound for the effect of the Rider on demand for Medicaid-financed long-term care services. Statewide, the upper bound is 6 percent of Rider participants with small variation between participants in urban and rural counties (5 and 8 percent, respectively). The percentage varies across regions from a high of 10 percent in Region 2 to a low of 2 percent in Region 6 (data not shown).

The fact that most facilities have had only a few transitions over a 36-month period provides additional evidence that it is unlikely that there has been a redistribution in the demand for Medicaid long-term care as the result of nursing home behavior or changes in clients demand in response to the Rider. Nonetheless, we examined the few facilities with a high volume of transitions more closely. Of the 15 facilities with more than 20 transitions, six were in Region 3, which has one-fifth of all eligible nursing facilities but just over one-quarter of all transitions, and six were in Region 11, which has 6 percent of eligible nursing facilities but 13 percent of all transitions. There was one high-volume facility in each of Regions 4, 7, and 10. All of the 15 high-volume facilities are for-profit facilities, have 100 or more beds, and are dually certified for Medicaid and Medicare, and all but one are located in urban areas. High-volume facilities are more likely to be high occupancy; however, all occupancy levels are represented among high-volume facilities. The concentration of a large number of transitions among a small number of facilities primarily in two regions suggests that there could be some facility-specific or region-specific responses to the Rider initiative that have not developed into statewide trends.

These findings are at odds with our perception from the site visits that there was substantial activity in using the Rider mechanism to get around the CBA waiting list (see interim site visit report). As noted above, neither regional DADS employees nor advocates nor providers saw any perversion of the purpose of the Rider initiative in such activity. We believe, however, that the pathway to community-based services that was opened by the Rider was sufficiently new and, to those for whom the interest list had been a discouraging barrier to needed services, sufficiently exciting, that our respondents often enthusiastic descriptions of how this new access had helped certain of their clients may have served to overplay how widespread such practices are relative to what is suggested by the data.

Reactions to the possibility of the Rider serving unintended purposes in access to community-based services were very different in the three regions that we visited. In one region, intentional bypass of the interest list was openly espoused. In another, respondents volunteered that unintentional bypass was more likely. In the third, respondents were adamant that no one would enter a nursing facility in order to get community-based services. In a teleconference, state officials said that they had no reports of backfilling of nursing facility beds as a result of the Rider, and advocates said that nursing facility entry was so onerous that few people would be willing to do this.⁵²

SUMMARY

Rider Participant Characteristics

This study found that the Rider initiative allowed adults of all ages with a wide range of physical and mental conditions and expected levels of need for assistance to return to the community. Most Rider participants were stable medically with relatively low care needs, as are most nursing facility residents. However, even nursing facility residents with the highest levels of expected care needs are able to move back to the community. Residents with both long and short lengths of stay in the facility as well as Medicaid clients recovering from acute episodes, such as a stroke, used the Rider to meet their care needs in the community.

Despite the diversity of characteristics seen among Rider participants, the likelihood that any given nursing facility resident returned to the community varied with resident characteristics. The quantitative analysis showed that elderly nursing home residents with dementia, non-elderly residents needing additional supervision for behavioral problems, and residents requiring the highest level of care were less likely to transition using the Rider. Younger residents, those in urban counties, and Hispanics are more likely to use the Rider, as are nursing home residents recovering from a stroke. Site visit respondents emphasized that the presence of informal support greatly improved the chances that a nursing home resident could transition, and transition more quickly, but many insisted that the lack of a strong informal support network was not a barrier to returning to the community.

The greater prevalence of diagnosed mental illness among nursing facility residents as compared with Rider participants and CBA clients supports the contention by some site visit respondents that this population is less well-served by the Rider initiative than others. Nonetheless, once they have made the transition, residents with mental or behavioral conditions appear to be no less likely than other types of Rider participants to be able to remain in the community for at least twelve months. There are possible selection issues associated with this finding in that those most able to remain in the community, either because of personal characteristics or informal support systems, are those who leave the facility. Alternatively, as site visit respondents suggested, community capacity in the area of behavioral illness might be the binding constraint.

The data also show racial/ethnic differences in the use of the Rider. Site visits reports of a cultural bias among Hispanics against institutional care for family members among Hispanics were supported by the data. Hispanics were under-represented in the nursing facility population and over-represented in the Rider population, suggesting a group that avoids entering a nursing facility and readily leaves when given the opportunity. The Rider initiative appears to contribute to making the Medicaid long-term care program more culturally appropriate for this group.

Understanding which populations are most likely to use MFP options when available could help identify barriers that impede return to the community. A comparison of the policies and procedures in place in Texas with those in other states could help identify which program components are associated with ease of transition for various populations and so promote further improvements in MFP policy.

Challenges in the Transition Process

The process of identifying individuals who wish to transition from nursing homes into Medicaid long-term care programs in the community and planning the transition involves a number of critical steps, which all take place in advance of the residents discharge from the nursing facility. Texas facilitates the advance coordination of transition plans by funding coordination activities by DADS caseworkers with Title XIX funding. A critical distinction between the MFP process in Texas compared to other states, such as Ohio,⁵³ is Texass authorization of eligibility determination for community care programs, contracting for minor home modifications, and ordering DME prior to a clients return to the community. However, restrictions on payment for these services until the clients discharge from the nursing facility can still complicate the coordination of needed services by the clients first day in the community.

Several additional issues were identified by site visit respondents as contributing to delays in transition or complications in the placement of nursing facility residents in the community. Finding reliable contractors to make home modifications and maintaining adequate PAS were noted as ongoing problems, but did not appear to constitute absolute barriers to transition. The challenge of finding suitable housing, identifying accessible transportation services, and finding physicians in the community were often noted as issues that delayed or discouraged transition. In contrast, the supply of ALFs in some regions appeared to both expedite and encourage transition. The use of assisted living may be a symptom of an affordable housing constraint rather than a reflection of a preference for assisted living by the Rider population. Understanding how and why the choice for assisted living is made would be a useful area of further research.

Challenges in rural areas, such as transportation difficulties and the risk of social isolation, are not specific to Rider participants or the CBA program. Rather, they are a good example of the fact that, in general, community capacity for the provision of services contributes to some of the differences in the number of transitions and in service use across regions observed in the data and reported by site visit respondents.

Participant Status Over Time

Our limited analyses using of the Texas Medicaid long-term care data suggest that the majority of Rider participants across regions and of different ages and physical conditions remain on the CBA program after twelve months. Only a small minority of either Rider participants or other CBA clients returns to a nursing facility during this period, suggesting that the CBA program has been able to prevent nursing facility admissions for two groups with potentially different risks of admission. Reports by site visit informants suggest more variation in experiences after leaving nursing facilities than can be captured by the quantitative data. However, no data have been collected systematically in Texas, either through qualitative or quantitative methods, on client outcomes. More complete and reliable data on reasons for program termination, hospitalization, admission to nursing facility, and mortality would shed more light on outcomes for participants. Such outcome measures, and outcome measures specific to certain populations, such as individuals with mental illness, could assist in identifying areas for programmatic improvement over time.

Service Use

We found important differences in service use between Rider and non-Rider CBA clients. The primary difference is the markedly greater use of assisted living by Rider participants, a difference that appears to drive many, but not all, of the other observed differences in service mix and use. Multivariate analysis required to sort out the effects of a variety of influencing factors is beyond the scope of this study. However, both quantitative and qualitative data suggest factors that could influence the disproportionate representation of Rider participants in ALFs. Some of these factors, such as differences in the availability of informal support systems or in the need for care might also have contributed to the decision by these individuals to enter the nursing facility. Other factors include cultural preferences, availability of independent housing, and adequacy of transportation. Finally, some case managers and ombudsmen working to place nursing facility residents in the community saw assisted living as a step down from nursing facility care for residents who wanted to leave the nursing facility but who might not be ready to live more independently in the community.

Even controlling for the difference in the use of assisted living, some variation in service mix between Rider participants and CBA clients remains across regions, racial/ethnic groups, TILE classes, and age groups. The multivariate analysis needed to tease out the relative contributions of the various factors was beyond the scope of this study. However, understanding the factors contributing to these differences and monitoring trends in use over time could help predict the composition of demand for community-based services over the long term and could help pinpoint the cost drivers in community-based long-term care.

Program Expenditures

Our analysis of expenditures for Rider participants and CBA clients suggests that overall Rider participants have greater service intensity than CBA clients and, as a result, have average monthly expenditures that are about 10 percent higher.⁵⁴ This finding is not surprising since the distribution of TILE classes among Rider participants shows that they are generally more care dependent than CBA clients. However, while the difference in expenditures is not large, TILE class does not fully explain it.

Expenditures for residents of ALFs are slightly higher than for non-residents, and expenditures for Rider ALF residents are higher than for non-Rider ALF residents. Differences in service use between residents and non-residents of ALFs appear to reflect logical and likely appropriate substitutions between ALF services and the package of PAS, minor home modifications, and meals available to non-ALF residents. Among non-ALF residents, costs for Rider participants are higher than costs for CBA clients. This difference is due primarily to a greater intensity in the use of PAS among Rider participants. The expenditure difference was not associated with any subgroup of Rider participants. Thus, it appears that Rider participants are receiving a higher intensity of service across age groups, TILE classes, rural and urban counties, and race/ethnicity groups. This analysis could not, however, answer the pressing question of how expenditures for Rider participants would compare to the expenditures these clients would have incurred had they remained in the nursing facility.

Discussion

There are several questions that arise from this study. Most respondents identified housing as the constraining factor in the return to the community--a constraint that was alleviated to some extent by the use of assisted living. Communities with a less well-developed market in assisted living might be less able to transition all willing nursing facility residents. Although, housing for the elderly and disabled is beyond the scope of the Medicaid program, it affects the implementation of initiatives like the Rider that seek to help nursing facility residents return to the community. Demand for assisted living services might be a reflection of demand for accessible housing rather than demand for the full range of CBA program services.

Similarly, we heard reports that the CBA program, and the Rider as a means to entering that program, was in demand not for the home and community-based services that are at its core, but for the access to unlimited prescription drugs that these programs offer as a benefit. Some respondents suggested that better prescription drug coverage under Medicaid and Medicare could reduce demand for the CBA program to a certain extent. The newly created Medicare Part D may serve that function.

There are regional differences in the populations served and the mix of services. These differences likely reflect, at least in part, the flexibility that the DADS regions are allowed in implementation. The diversity of the DADS regions and the variations in implementation show the value of flexibility in adapting a state-level policy to local conditions and cultural norms. Implementation differences might be a response to local market conditions or cultural preferences. Alternatively, the differences might reflect gaps in services or culturally inappropriate approaches to service delivery among specific populations. Monitoring regional differences could provide one way to identify potential problem areas as well as potentially useful innovations.

The capacity of the community to respond to the needs of nursing facility residents returning to the community has been a recurring theme in this study. The Rider initiative is only one component of a broader rebalancing of Medicaid long-term care in Texas. The initiative has now been codified through regulation, giving it a permanence that should help convince community care providers to invest in increased capacity. A closer analysis of the service mix by region and county could help identify places where capacity is the constraint on additional transitions or on transitions by particular types of nursing facility residents. The challenge for states is to seek a balance between local flexibility and state support of local structural capacity in long-term care systems in order to achieve equal opportunity for independent living across different populations at risk of institutionalization.

APPENDIX 1: DATA AND METHODS

Qualitative Research Methods

The qualitative component of the research was based on key informant interviews at four sites in three DADS regions across the state during April and May 2004. Because of our interest in the transition process, we chose sites with a high number of transitions, based on data provided by DADS on the number of transitions by region and within each region by county. In addition, we looked for geographic spread of the chosen regions across the state. Within each region, we looked at county demographic characteristics including ethnic diversity, income, and home ownership (as a measure of wealth and stability of the population). Finally, we looked for a predominantly rural county with a relatively high number of transitions that was within a three-hour drive of one of the DADS regional headquarters in order to look at potential differences in process between rural and urban areas.

Based on these criteria, we chose to visit Ft. Worth in Region 3 (Tarrant County, north-central Texas), McAllen in Region 11 (Hidalgo County, on the Texas-Mexico border near the Gulf coast), and Abilene and the nearby rural town, Brownwood, in Region 2 (Taylor County and Brown County, west-central Texas).⁵⁵ Briefly, Tarrant County has a predominantly Caucasian population and, relative to the statewide average, fewer families with incomes below the federal poverty level, fewer individuals receiving Supplemental Security Income benefits, and a slightly larger proportion of African-American residents. In contrast, Hidalgo County has a strong majority of Hispanic residents, most of whom speak Spanish at home. Poverty is more widespread relative to the statewide average, and the population is older with a greater prevalence of disability. Taylor and Brown Counties are also predominantly Caucasian but with smaller Hispanic and African-American minorities than Tarrant or Hidalgo Counties. Brown County is predominantly rural, McAllen and Abilene are relatively small urban areas, and Ft. Worth is a large urban area, part of the so-called Metroplex of Dallas-Ft. Worth.

All of our discussions with key informants at each site were scheduled in advance of our visits. We met first with regional DADS officials, caseworkers, supervisors, Medicaid eligibility staff, and quality control staff. We visited nursing facilities identified by DADS as well as other nursing facilities chosen based on the number of transitions, the proportion of residents transitioned, and facility characteristics such as ownership status, history of deficiencies, and occupancy rate. We made an effort to visit at least one nursing facility with a high occupancy rate and one facility with a low occupancy rate (relative to other facilities with transitions) at each site. At each site, we met with home health providers and other community providers, such as ALFs and adult day care, chosen based on information gleaned from the literature and from preliminary conversations with advocates and DADS officials about the most important types of community providers in the region. In each area, we visited with the long-term care ombudsman and representatives from the Area Agency on Aging and met with local advocacy groups. Finally, when possible, we met with staff at the organizations that were beginning implementation of the new statewide relocation contracts.

For each category of key informant with whom we met, we developed a standard discussion guide based on a review of the literature and discussions with state officials and advocates and designed to elicit both knowledge and opinions from a wide range of people involved in the Rider initiative.⁵⁶ Discussion topics included outreach and education about the Rider initiative for staff, nursing facility clients, and nursing facilities; the process to transition; procedures for monitoring clients after transition and assuring quality of services; administrative procedures and perceived burden; characteristics of participating providers; characteristics of participating clients; services capacity in the community; and client satisfaction. All informants were promised anonymity to encourage them to express their thoughts freely.

Quantitative Research Methods

In the quantitative analysis, we used data from Texas Medicaid long-term care data files to explore how Medicaid recipients who opted to transition to the community through the Rider (Rider participants) compared to their non-Rider counterparts who remained in nursing facilities (nursing facility residents) or who were already receiving services through the CBA program (CBA clients).⁵⁷ We used a cross-section of adult Medicaid nursing facility residents as the nursing facility comparison group and all non-Rider CBA clients who entered the CBA program while the Rider was in effect as the CBA comparison group.

We compare Rider participants to Medicaid nursing facility residents to understand which types of residents are likely to transition to community care. The Medicaid nursing facility resident population in Texas represents the pool of people potentially eligible to use the Rider to gain access to Medicaid community-based waiver programs. We compare non-Rider CBA clients to Rider participants post-transition in order to understand how Medicaid recipients in community care who have not had a nursing facility stay immediately preceding entry into CBA differ from those who have. Non-Rider CBA clients represent Medicaid recipients who have been determined to be in need of the level of care required for nursing facility admission but who receive this care in the community without necessarily first going into a nursing facility. Because there are only a small handful of Rider participants under the age of 21 and because only adults are eligible for the CBA program, we limit our analysis to recipients who were age 21 and older.

Depending on the particular research question being addressed, we compare Rider participants to nursing facility residents and/or CBA clients based on their demographic characteristics, reported medical condition, expected level of need for care, and diagnosis. We also compared characteristics of the facilities that Rider participants are leaving and facilities with no transitions. We examined service mix and recipient status at six and twelve months after entry into the CBA program for Rider participants and CBA clients. Finally, we looked at costs associated with the provision of long-term care to the two community-based care groups. For all analyses, we considered estimates at the state level as well as across the ten DADS service administration regions and rural and urban counties. We also considered various characteristics of transitions to assess if and how the presence of the Rider initiative has affected demand for Medicaid long-term care services.

Comparison Groups

We developed three analytic samples corresponding to the three comparison groups using Medicaid long-term care data provided by Texas DADS.⁵⁸ We used Medicaid long-term care administrative data on all Rider participants beginning in September 2001, when the Rider option became available, and on all CBA clients who entered the program between September 2001 and March 2004.⁵⁹ We used Medicaid long-term care administrative data on a cross-section of all adults (age 21 or over) residing in nursing facilities in Texas in October 2003 as our nursing facility comparison group.⁶⁰

A Rider participant is here defined as a nursing facility resident who received authorization to enter the CBA program funded through Medicaid under the provisions of the Rider and has received CBA services in the community. We identified 5,033 individuals authorized to participate in the Rider initiative. Of these, 4,870 had a record of a payment or services under CBA and a confirmed nursing facility stay prior to receipt of CBA services and so were included in the Rider participant group. We restricted the nursing facility and CBA comparison groups to people who were not identified as members of the Rider population. For the CBA group, we excluded any clients who had no CBA payment data or only had a clinical pre-assessment payment for the CBA program, taking into account the possibility of lagged accounting of payment records. The nursing facility resident comparison group consisted of 65,132 individuals⁶¹ and the CBA comparison group consisted of 15,895 individuals.⁶²

Variables Used in Comparisons

Individual and facility level characteristics

For individual level analyses, we used data on age, gender, race and ethnicity, urban/rural county of residence, medical diagnosis, and length of stay in the facility prior to transition. In addition, we used the TILE class as an indicator of the expected level of care dependence.⁶³ For the nursing facility-level analysis, we looked at both the proportion of facilities from which a Rider transition occurred and the number of transitions per facility and compared facilities with and without transitions based on ownership status, bed certification (Medicaid-only versus mixed certification), size of facility (total licensed beds), rural/urban location, and occupancy rates.

Service use

We compared across samples the proportion of each sample that used any amount of each service in the first six months after entry into the CBA program. We focus on the first six months on the program for each person as the best representation of service use for all recipients since analysis of other time periods, specifically, 7-12 months and 13-18 months, are subject to greater selection effects due to attrition from the program. We excluded use of prescription drugs since all recipients are authorized to receive this benefit and because data on prescription drug claims were not available to us.

Measures of status after program entry

We looked at selected measures of status of Rider participants in comparison with those of CBA clients at six and twelve months after CBA entry. Our status measures are restricted to the data in the long-term care database, which are informative despite being limited in scope. These measures are whether, at the end of the period, the recipient was still on the CBA program, or in a nursing facility for other than respite care. In some cases, we found no evidence of continuing Medicaid long-term care service at the end of a period.⁶⁴ Because we had limited data on reasons for service termination, we classified these clients as having left CBA with unknown status.

Costs

Our cost analysis focused on the average monthly service cost of community-based long-term care for Rider participants and CBA clients over the first six months after entry into the CBA program. The analysis includes only long-term care costs incurred under the CBA program and excludes prescription drugs and all acute care costs, such as hospitalization and physician visits. Database structures did not allow us to calculate nursing facility costs with sufficient reliability to support research questions on the cost of nursing facility care, so analyses of nursing facility costs were not undertaken. We present long-term care costs under CBA as an average per person monthly payment rate, using data for the six months after entry into CBA, with allowances made for lagged billing.⁶⁵ The average payment per month is normalized to a standardized month of 30.46 days and is adjusted to 2003 reimbursement levels. We also computed average monthly program expenditures for Rider participants and CBA clients in the aggregate, which differs from our cost calculation in that it takes into account both the amount of the service used and the proportion of the group that used the service.

We looked at total costs and costs by type of service for the six most frequently used services. Medicaid does not finance certain CBA services for clients in an ALF since these services are generally included in the assisted living service package. Therefore, we calculated cost estimates separately for clients in assisted living and those not in assisted living to understand how costs and service use differ for these two groups of clients. A client who has any payment for an ALF during the six-month period is assumed to be in assisted living for the full period.

APPENDIX 2: TILE CLASSES

The Texas Index for Level of Effort, or TILE index, is an index used to determine the maximum daily payment rate to providers for a Medicaid recipient in Texas. The TILE levels or classes are based on diagnosis, information on limitations in activities of daily living (ADLs) converted to an ADL score (a lower score represents higher functional status), level of care intervention, the presence of mental or behavioral conditions that affects care, and receipt of rehabilitation services. A full description of each TILE is available at: http://www.dads.state.tx.us/handbooks/mhpm/res/apx06.pdf. A more detailed description of the determination process is available at: http://texinfo.library.unt.edu/texasregister/html/2001/Jul-06/adopted/1.....

Below is a brief description of each TILE level based on this information.

TILE 211 = Clinically stable and ADL score of 3 (lowest reimbursement level). An ADL score of 3 represents the highest level of functional status allowed when meeting Texas criteria for nursing facility admission. In addition, recipients in this class must not meet the criteria for TILE 202 or 210.

TILE 210 = Clinically stable with mental or behavioral condition affecting care level, and ADL score of 3. The following conditions qualify as meeting this mental/behavioral subgroup: incoherent/frequent disorientation requiring daily intervention; or disruptive or aggressive behavior requiring daily intervention.

TILE 209 = Clinically stable and ADL score of 4.

TILE 208 = Clinically unstable and ADL score of 3. To qualify as clinically unstable, the recipient must have one of the following conditions or be receiving one of the following treatments: recent amputation, seizures, dehydration, acute urinary tract infection, incontinence or a Foley catheter, oxygen administration, respiratory therapy, wound dressing for an open wound.

TILE 207 = Clinically stable and ADL score of 5-6.

TILE 206 = Clinically unstable and ADL score of 4-6.

TILE 205 = Clinically stable and ADL score of 7-9.

TILE 204 = Clinically unstable and ADL score 7-9.

TILE 202 = A recipient in this class must be receiving restorative nursing care as follow-up to rehabilitation therapy. The rehabilitation therapy must be physical or occupational therapy ordered by a physician and initiated due to a documented event. Therapy must be expected to result in significant, measurable, functional progress. In addition, the recipient must have an ADL score of at least 3.

TILE 201 and 203 = Requires heavy care, defined as having one of the following conditions or receiving one of the following treatments: coma, quadriplegia, stage III or IV decubitus with care required, non-oral administration of 60 percent or more of nourishment, daily oral or nasal suctioning, daily tracheotomy care or suctioning, excluding self-care. In addition, the recipient must have an ADL score of 6-7 (TILE 203) or 8-9 (TILE 201, highest reimbursement level).

Assignment of TILE Class for Each Population

When a Medicaid recipient enters a nursing facility, Medicaid determines the level of reimbursement through a Form 3652 Level of Care assessment. Based on this assessment, the recipient is assigned a TILE class, one of 11 payment categories. TILE classes are based on medical stability, medical diagnoses, and level of dependency, or the number of activities of daily living (ADLs) with which a recipient requires assistance.

Determination of TILE class is made when an individual first qualifies for Medicaid, and every six months thereafter. Assessments may also be conducted when providers observe changes in health status. Thus, each individuals payment record includes multiple TILE classes over time.

To compare the Rider sample to the two comparison samples, we used the TILE score closest to the Rider samples date of entry into CBA. We also examined the TILE scores of the Rider sample assigned at the time Medicaid eligibility was first determined (in the nursing facility), to determine if TILE scores were significantly different, or provided information on reasons for admission (i.e. rehab). Here the significant difference was that the sample had a higher proportion classified as rehabilitation (TILE 202) at admission and a lower proportion with TILE 211 relative to relative to the TILE classes at CBA entry.

Protocols used to search the data and obtain a TILE class for each individual.

Rider Sample

The goal of this analysis was to get a TILE score for each observation in the Rider Sample as near as possible to the Rider Start Date but before entry into the CBA program. For each observation a TILE score was taken from the Tex10b_LevelDataforCBAClnt database starting from the Rider Start Date and moving backwards in time. This exercise resulted in 695 missing cases (14 percent of the sample).

In order to reduce the number of missing cases, TILE scores were then taken from the same database from the Rider Start Date moving forward in time three days. Unfortunately, this method only reduced the number of missing cases by one observation. Next, given the likelihood that an observations TILE score did not change dramatically over short time periods, on the same database the nursing facility begin date was used going forward in time to extract a TILE score. This method also reduced the number of missing cases by only one observation. Finally, the Medicaid Long-Term Care Payment Database was used to extract a TILE score using the nursing facility begin date and moving forward in time. This method successfully reduced the number of missing cases by 444 observations. The final number of missing cases was 295 observations or approximately 5 percent of the sample.

CBA Sample

For the CBA Sample, the goal was to obtain a TILE score as near as possible to the CBA Start Date. Using the Tex10B_LevelDataforCBAClnt database, the CBA Start Date was used moving backwards in time to extract a TILE score. This method was quite successful resulting in only 923 missing cases (or nearly 6 percent of the sample).

In order to further reduce the number of missing cases, the same database was used going forward in time from the CBA Start Date, once again assuming that TILE scores do not change dramatically over short time periods. This method reduced the number of missing cases to 886 observations (or approximately 5.5 percent of the sample).

Nursing Facility Sample

The most recent available TILE score was extracted for the Nursing Facility Sample using the Medicaid Long-Term Care Payment Database. Only 460 cases were missing using this method (less than 1 percent of the sample).

ENDNOTES

In this report, we use the term transition to denote movement of a Medicaid recipient from a nursing facility to the CBA program under the provisions of the Rider. For example, a facility from which three residents have moved to the CBA program is said to have had three transitions. Similarly, a resident is said to have transitioned to the community.
The Texas budget cycle covers two years. The Texas fiscal year (FY) begins on September 1. Thus, the effective period of Rider 37 was September 1, 2001, (the beginning of Texas FY 2002) through August 31, 2003, (the end of FY 2003).
It is coincidental that this rider, Rider 37 to the 2004-2005 General Appropriations Act, has the same number as the Rider 37 to the 2002-2003 General Appropriations Act that established the nursing facility transition initiative.
SB 367 Interagency Task Force on Appropriate Care Setting for Persons with Disabilities. Promoting Independence Makes Good Cents. Texas Health and Human Services Commission. November 2003. Under Rider 37, the funds to be transferred to the community care budget were limited to either the average nursing facility costs or the actual cost for the individual client who was receiving services under the Rider initiative, whichever was higher.
Correspondence with Marc Gold, Department of Aging and Disability Services, May 16, 2005.
In this report, we refer to the initiative defined by Rider 37 and continued under Rider 28, as modified by Rider 7b and the new Rider 37, as the Rider except where the distinction among the provisions in the separate Riders is important.
Texas Health and Human Services Commission. House Bill 966 Report. October 2002. http://www.hhsc.state.tx.us/pubs/102502_HB966_InstCareStudy.html, Accessed August 25, 2004.
Eligibility for a Medicaid program is not determined at the time residents are placed on the interest list.
Texas Department of Human Services, LTC Education Services. Community Care Options. May 2004. p. 75. http://www.dhs.state.tx.us/providers/pi/Com_Care_Options_manual.pdf, Accessed August 26, 2004.
Personal communication with Marc Gold and Dan Anderson, DADS, May 16, 2005.
"Money Follows the Individual Rebalancing Initiative, http://www.cms.hhs.gov/newfreedom/0303mfir.pdf, Accessed July 14, 2004. See also The New Freedom Initiative: Presidents FY 2006 Budget and CMS Accomplishments, http://new.CMS.hhs.gov/NewFreedomInitiative/downloads/NFIAccomplishments..., Accessed February 15, 2006.
Promising Practices in Home and Community-Based Services, Texas--Appropriations Rider: Promoting Independence, Money Follows the Person, http://www.cms.hhs.gov/promisingpractices/tx-rider37.pdf, Accessed July 14, 2004.
Task Order Number 4, Contract number HHS-100-03-0011. "Examination of Rider 37: Texas Community Based Alternative." Effective September 19, 2003.
Barbara A. Ormond, Kirsten J. Black, Brigette M. Courtot, and Anna S. Sommers, An Examination of Rider 37: Texas Community Based Alternative, Site Visit Analysis, August 24, 2004.
The state of Texas health and human services organization underwent a major reorganization and consolidation in 2004. The programs referenced in this study were originally administered by the Texas Department of Human Services (TDHS), which no longer exists and so is considered a legacy agency. Parts of TDHS were subsumed under other agencies, including the Health and Human Services Commission (HHSC) and the Department of Aging and Disability Services (DADS). In this report, we make reference to the programs and activities of DADS, although at the time of implementation of the Rider, TDHS was the administering agency.
Anna S. Sommers, Barbara A. Ormond, Kirsten J. Black, and David DOrio, Examination of Rider 37: Texas Community Based Alternative, Analysis of State Data Systems, An Interim Report, Washington, DC: The Urban Institute, June 2005.
Task Order Request, Contract Number HHS-100-03-0011, An Examination of Rider 37: Texas Community Based Alternative (DA-03-003).
Detailed research methodologies are presented in the separate reports including detailed descriptions of all databases.
Selected demographic data for these sites are presented in Ormond et al. 2004.
Discussion topics included outreach and education about the Rider initiative for staff, nursing facility clients, and nursing facilities; the process to transition; procedures for monitoring clients after transition and assuring quality of services; administrative procedures and perceived burden; characteristics of participating providers; characteristics of participating clients; services capacity in the community; and client satisfaction.
For ease of exposition, we generally refer to non-Rider CBA clients simply as CBA clients and to CBA clients who entered the program via the Rider as Rider participants, recognizing that Rider participants are also CBA clients.
In addition to the CBA program, Rider participants have also entered the Community Living Assistance and Support Services (CLASS) waiver program for people with developmental disabilities, the Medically Dependent Childrens Program (MDCP) waiver for children, and the non-waiver Community Care program. Texas does not have an Independent Choices waiver.
Change in eligibility following reassessment can be appealed. Site visit informants noted that it is very rare that someone is dropped form the CBA program for medical ineligibility following reassessment.
Specific therapies include such services as speech or occupational therapy.
ADA = Americans with Disabilities Act.
Participation by region varied from a low of about 50 percent of facilities in Regions 1 and 6 to a high of almost 90 percent of facilities in Region 11. Not surprisingly, the three regions with the highest share of facilities participating are the regions with the highest number of transitions. Regions 3 and 11 each have six of the 15 facilities with more than 20 transitions. Of the two regions with the lowest participation, Region 6 is highly urbanized while Region 1 is predominately rural, and there are fewer large facilities in Region 1. Otherwise, the profile of facilities in these two regions is similar to the statewide profile, suggesting that facility characteristics explain only part of the differences across regions.
As of March 31, 2006, 10,711 people have transitioned under Riders 37, 28 and the MFP program. Personal communication with Marc Gold, April 25, 2006.
CBA case manager conducts a Form 2314 Client Satisfaction Interview with all CBA clients. These monitoring contacts have been required since the start of CBA in the mid 1990s although the current Form 2314 is more recent. The first monitoring contact at six months can be by telephone or face-to-face; the second monitoring contact, at twelve months, must be face-to-face.
Investigation of the details of the new ALF quality monitoring system was beyond the scope of this study.
TILE 211 has the lowest reimbursement rate and includes individuals who are clinically stable and meet the minimum requirements for nursing facility admission based on ADL limitations. (The lowest ADL score meeting the criteria for nursing facility admission is three.) TILE 201 has the highest reimbursement rate and indicates individuals requiring heavy care with the greatest number of limitations in ADLs. A detailed description of TILE classes and scoring can be found at http://www.dads.state.tx.us/handbooks/mhpm/res/apx06.pdf.
We grouped TILE classes as follows for analysis (n.b. these groupings are not used by Texas):
- TILE 211 (low care dependence) - clinically stable and ADL score of 3.
- TILE 210 (low care dependence with mental/behavioral care needs) - ADL score of 3 and clinically stable with mental or behavior illness or condition affecting care level.
- TILE 207-209 (moderate care dependence) - clinically stable with ADL score of 4-6, or clinically unstable with ADL score of 3.
- TILE 204-206 (high care dependence) - clinically stable with ADL score of 7-9, or clinically unstable with ADL score 4-9.
- TILE 202 (rehabilitation) - receiving restorative nursing care as follow-up to rehabilitation therapy with the expectation of significant functional progress.
- TILE 201, 203 (heavy care dependence) - requiring heavy care with ADL score of 6 or more, including recipients with quadriplegia or in a coma.
Diagnoses presented here represent standard groupings of ICD-9 codes used in recording the diagnoses. The first diagnosis listed, based on Form 3652-A instructions, is the primary diagnosis or reason that the client needs long-term care. Other diagnoses are current medical conditions with a direct bearing on the required treatment or nursing care. Because diagnoses may play a greater role in nursing facility admission than in community-based care, we use all diagnoses for analysis to provide greater comparability between Rider participants and CBA clients. The ICD-9 groupings are based on a system devised by Andy Kramer at the Colorado Research Group for nursing facility residents. The major classifications (e.g., diseases of the nervous system) are part of ICD-9 general classification system. Identification of specific conditions was reviewed by Robert Berenson, a physician on The Urban Institute staff.
Physical conditions and disease were grouped as follows:
- stroke;
- other cardiovascular (including hypertension, heart disease, and congestive heart failure);
- endocrine, nutritional and metabolic diseases, and immunity disorders (including diabetes, thyroid disorders, and nutritional deficiencies such as anemia);
- diseases of the nervous system and sense organs (including Parkinsons disease, multiple sclerosis, cerebral palsy, and epilepsy);
- injury and poisoning (including falls and adverse drug reactions);
- diseases of the respiratory system (including pneumonia, chronic bronchitis, and emphysema).
Mental illness and behavioral disorders were grouped as follows:
- dementia;
- depression;
- disorders related to drug and/or alcohol addiction;
- all other mental illness, predominantly forms of psychosis, schizophrenia, panic/anxiety disorders, and paranoia;
- no diagnosed mental or behavioral condition.
Detailed comparisons are found in the interim report on the quantitative analysis.
Full details on rural-urban differences are presented in the interim data analysis report.
According to the Texas Administrative Code Title 40, Chapter 82, the state defines an assisted living facility as an establishment that provides food, shelter, and personal care services to four or more unrelated persons in one or more facilities. The statutory rules state that the service philosophy of assisted living emphasizes personal dignity, autonomy, independence, and privacy and allows for resident's to age in place while receiving services according to level of need. In Texas, only licensed facilities may use the term assisted living. Assisted living services are covered by the state's Medicaid Home and Community-Based Services waiver program in licensed assisted living facilities (Mollica R. 2002. State Assisted Living Policy: 2002. National Academy of State Health Policy. Portland, Maine).
Texas Department of Human Services, Rider 28 Client Demographics, Data Effective Date: March 31, 2004.
Our analysis of service mix does not include estimation of the intensity of services received, only the share of recipients receiving each service. Data available on the number of service units received was not adequate to conduct such an analysis. Instead, we rely on an analysis of average monthly costs among users of each service to determine relative intensity of services between Rider participants and CBA clients. This cost analysis is presented below. We present findings here on service use in the first six months after program entry. We also considered other time periods but found little difference between these periods and the first six months. For ease of exposition, we here present findings for the first six months only.
This analysis of nursing facility stays during the six-month period includes stays of any length, excluding respite care.
Adaptive aids include durable medical equipment, such as beds and wheelchairs, and some aids not covered under traditional Medicaid, such as seeing-eye dogs. Medical supplies include perishable supplies, such as diapers. All clients receive unlimited prescription drug coverage. Information on this benefit is not available in our datasets.
TILE class determines a payment rate to nursing facilities for its residents, whereas services in CBA are paid based on fee-for-service and not based on the TILE.
We were limited by the available data to this measure of status.
The data used to assess recipient status is subject to important limitations and omissions. Detailed exposition on these omissions and the impact of various methods on results can be reviewed in Section VIII of Sommers et al. 2005.
A significant limitation of the data available to assess recipient status is that it is likely to underreport mortality, with some share of those with unknown status actually having died. In our assessment based on discussions with state officials familiar with the data and the populations served, the following circumstances are most likely to be associated with assignment of unknown status: (1) hospitalization, which is not included in the long-term care files; (2) death, which may not always be recorded in the data files; and (3) moved out of state. It is also possible that recipient health improved to a degree that left them medically ineligible for the CBA program, although our site visit informants suggested that termination for medical ineligibility rarely happens. With the available data, we have no means to assess the relative impact of each of these factors.
Personal communication with Marc Gold, Dan Anderson, and Gerardo Cantu of Texas DADS, May 16, 2005.
DADS, personal communication, December 16, 2003.
In the quantitative analysis, we attempted to quantify the extent of such behavior, as we discuss below.
We use data from the first six months after entry into the CBA program to estimate average monthly service costs for each Rider participant and CBA client by service. We then consider the cost per recipient for the complete set of services paid for by the CBA program for each population. This total cost calculation allows us to estimate the overall average monthly Medicaid CBA program expenditures for recipients in both groups in the first six months of service.
First, we found that Rider participants and CBA clients differ somewhat in the timing of their entry into assisted living in the six-month analysis period. Among Rider participants classed as ALF users, 95 percent had reported payments for assisted living in the first month after CBA entry, compared with 83 percent among CBA clients (data not shown). There are also differences in the share of each group continuing to reside in ALFs over the study period. The share with assisted living payments drops from 95 to 83 percent in the sixth month after CBA entry for Rider participants but remains stable at 83 percent among CBA clients (data not shown). The decline in ALF use over time among Rider participants supports the notion expressed by caseworkers in the site visits that assisted living could serve as a bridge from the nursing facility to more integrated community living. A longitudinal analysis of the living situation of Rider participants would be necessary to test this hypothesis.
Some of this 3 percent might have resided in an ALF prior to nursing facility admissions, but this information was not included in the available data.
Money Follows the Person: The Experience in Texas, ILRU teleconference, March 31, 2004.
Kasper, Judy and Molly OMalley, 2006. Nursing Home Transition Programs: Perspectives of Medicaid Care Planners. Kaiser Commission on Medicaid and the Uninsured, Washington, DC.
We did not have data on service intensity so we used expenditures by services as a proxy.
Selected demographic data for these sites are presented in Ormond et al. 2004.
A copy of the final discussion guide is included in Ormond et al. 2004.
For ease of exposition, we generally refer to non-Rider CBA clients simply as CBA clients and to CBA clients who entered the program via the Rider as Rider participants, recognizing that Rider participants are also CBA clients.
These three groups are described briefly here. A detailed account of how each sample was constructed is provided in Sommers et al. 2005. Both the Rider and the CBA groups described here constitute populations, and the group sizes are large. As a result, all but the smallest differences between groups will likely be statistically significant. Therefore, we recommend that policy-makers look for differences that are large enough to have policy relevance rather than those that meet the test of statistical significance. For this reason, we do not note statistically significant differences in this report.
We restrict the CBA client sample to this time period to minimize the effect of secular trends on observed differences between these two groups.
We chose October 2003 to represent current caseloads because it was toward the end of our study period but avoided holiday months.
Throughout this report, when we refer to the nursing facility population we are talking specifically about the Medicaid nursing facility population, unless otherwise indicated.
The size of each comparison group can vary depending on the analysis in question. For example, some analyses look at differences in service use after six and twelve months on the program and so include only those individuals who had been in the program for at least that length of time. Other analyses require similar adjustments in the comparison group.
The TILE classification system is described in Section III, and in more detail in Appendix 2. A description of each TILE is available at http://www.dads.state.tx.us/handbooks/mhpm/res/apx06.pdf.
There are many possible reasons for discontinuation of services. Hospitalizations are not recorded in the databases we used; therefore, if the client was hospitalized and died in the hospital, the death would not be reported in the long-term care Medicaid database. Alternatively, the client might have been found ineligible for Medicaid services at redetermination, might have become ineligible because his or her health improved and he or she no longer met the medical necessity criterion for program participation, or moved out of state, or did not make any required financial contribution, or behavioral issues made him or her unsuited for the program.
For recipients with less than six months of program participation, we averaged costs over the available months.

TABLES

TABLE 1: Share of Texas Nursing Facilities with and without Transitions Under Riders 37 and 28 in Medicaid, Statewide and by Region
	Share of Facilities with Transitions	Share of Facilities without Transitions
Statewide	71.1%	28.9%
Urban counties	68.1	32.1
Rural counties	76.2	23.8
Region 1: High Plains	50.6	49.4
Region 2 & 9: NW Texas	72.4	27.6
Region 3: Metroplex	75.7	24.3
Region 4: Upper East Texas	79.1	20.9
Region 5: SE Texas	74.7	25.3
Region 6: Gulf Coast	50.4	49.7
Region 7: Central Texas	67.9	32.1
Region 8: Upper South Texas	82.5	17.5
Region 10: Upper Rio Grande	81.3	18.8
Region 11: Lower South Texas	89.9	10.1
NUMBER OF FACILITIES	826	336
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care files. NOTE: Estimates are row percentages.

TABLE 2: Characteristics of Texas Nursing Facilities with and without Transitions Under Riders 37 and 28 in Medicaid
	Share of Facilities with at Least One Transition	Share of Facilities with No Transitions	Share of All Eligible Facilities¹
Number of Transitions
No transitions	--	100.0%	28.8
1-2 transitions	35.5%	--	25.2
3-5 transitions	30.7	--	21.8
6-10 transitions	20.1	--	14.3
11-20 transitions	12.0	--	8.6
More than 20 transitions	1.8	--	1.3
Ownership²
Non-profit	8.3	12.7	9.6
For-profit	89.6	59.5	81.0
Government	1.5	5.7	2.7
Other	0.6	22.1	6.7
Bed Capacity³
Medicaid only	7.1	30.1	11.0
Medicare and Medicaid	92.9	69.9	89.0
Facility Size³
Less than 50 beds	2.9	7.8	3.8
50-99 beds	30.8	46.4	33.5
100-149 beds	50.1	33.7	47.3
150 or more beds	16.2	12.1	15.5
Occupancy³
60%	23.6	30.1	24.7
60-69%	13.5	8.4	12.6
70-79%	17.5	18.1	17.6
80-89%	24.4	22.9	24.1
>90%	21.0	20.5	21.0
Urban	58.2	67.6	60.9
Rural	41.8	32.4	39.1
Region
Region 1: High Plains	4.9	12.2	7.0
Region 2 & 9: NW Texas	11.6	11.0	11.4
Region 3: Metroplex	21.8	17.3	20.5
Region 4: Upper East Texas	12.3	8.0	11.0
Region 5: SE Texas	6.7	5.7	6.4
Region 6: Gulf Coast	8.6	21.1	12.2
Region 7: Central Texas	13.3	15.5	13.9
Region 8: Upper South Texas	11.8	6.3	10.3
Region 10: Upper Rio Grande	1.6	0.9	1.4
Region 11: Lower South Texas	7.4	2.1	5.9
NUMBER OF FACILITIES	832	337	1,169
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care files. NOTE: Estimates are column percentages. A facility with any Medicaid certified beds is considered eligible. Excludes twelve facilities where ownership descriptions were unavailable. Excludes 186 facilities where capacity data were unavailable. Can include private beds.

TABLE 3: Demographic Characteristics of Texas Medicaid Long-Term Care Recipients Rider Participants, Nursing Facility Residents, and CBA Clients
	Number of Recipients	Mean Age	Under Age 65	Female	Race/Ethnicity				Resides in Rural County
	Number of Recipients	Mean Age	Under Age 65	Female	White, Non-Hispanic	Black, Non-Hispanic	Hispanic	Other	Resides in Rural County
Rider Participants¹
Statewide	4,870	70.2	30.0%	65.7%	67.0%	10.9%	18.8%	3.3%	31.0%
Urban counties	3,360	69.5	34.7	65.4	64.3	12.1	20.1	3.5	---
Rural counties	1,508	71.8	29.1	66.4	72.9	8.4	15.9	2.9	---
Facility Residents²
Statewide	65,132	78.5	14.7	70.7	69.4	14.2	14.0	2.4	31.8
Urban counties	44,393	77.7	16.2	70.6	65.1	16.4	15.7	2.8	---
Rural counties	20,737	80.1	11.5	71.1	78.5	9.6	10.2	1.6	---
CBA Clients³
Statewide	15,895	71.4	27.0	70.6	50.9	18.1	28.0	3.0	35.6
Urban counties	10,231	71.1	27.7	70.6	47.1	20.2	29.3	3.3	---
Rural counties	5,654	71.9	25.8	70.5	57.8	14.1	25.7	2.5	---
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All residents of nursing facilities who were Medicaid eligible at any point in October 2003. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 4: Texas Medicaid Long-Term Care Recipients: Rider Participants, Nursing Facility Residents, and CBA Clients Distribution of Texas Index for Level of Effort (TILE) Class
	TILE Class						Number of Recipients
	Less Care-Dependent			More Care-Dependent >
	Low Care(211)	Low Care with Behavioral(210)	Moderate Care(207-209)	Rehabilitation(202)	High Care(204-206)	Heavy Care(201, 203)
Statewide
Rider Participants¹	35.6%	8.0%	28.5%	16.9%	7.7%	3.3%	4,621
Facility Residents²	30.3	12.1	32.4	12.6	7.5	5.1	64,671
CBA Clients³	45.2	12.9	31.2	1.0	7.1	2.7	15,009
Urban Counties
Rider Participants	35.8	8.6	28.1	16.2	7.7	3.6	3,177
Facility Residents	28.3	12.1	33.6	13.1	7.3	5.8	44,079
CBA Clients	42.1	13.1	33.0	1.1	7.4	3.3	9,630
Rural Counties
Rider Participants	33.1	6.1	30.9	18.9	7.8	3.2	1,442
Facility Residents	34.7	12.3	29.9	11.7	7.9	3.6	20,592
CBA Clients	50.6	12.6	27.8	0.9	6.5	1.6	5,371
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. NOTES: The TILE class determines reimbursement rates to nursing homes and is based on level of care need and medical stability. TILE class was unavailable for 460 nursing facility cases, 249 Rider cases, and 886 CBA cases. TILE for Rider and CBA clients is the TILE at CBA entry; for nursing facility residents, it is the most recently assigned TILE. 1.All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All residents of nursing facilities who were Medicaid eligible at any point in October 2003. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 5: Texas Medicaid Long-Term Care Recipients: Rider Participants, Nursing Facility Residents, and CBA Clients Percentage with Selected Diagnoses by Age Group
	Selected Diagnoses											Number of Recipients
	Stroke	Other Cardio- vascular	Endocrine, Nutritional & Metabolic Diseases, & Immune Disorders	Diseases of the Nervous System & Sense Organs	Injury & Poisoning	Respiratory System Diseases	Any Diagnosis of Mental Illness:
	Stroke	Other Cardio- vascular		Diseases of the Nervous System & Sense Organs	Injury & Poisoning	Respiratory System Diseases	Dementia	Drug & Alcohol Addiction	Other Mental Illness	Depression	No Diagnosed Mental Illness
Under 65
Rider Participants¹	16.8%	61.8%	43.8%	24.8%	13.3%	16.4%	8.6%	2.7%	22.2%	26.9%	55.3%	1,605
Facility Residents²	13.2	51.9	30.6	26.3	16.5	11.6	12.8	3.3	26.0	21.5	53.9	9,560
CBA Clients³	11.5	55.5	41.1	29.3	8.3	18.8	3.6	0.4	13.8	16.5	70.3	4,188
65 and Older
Rider Participants	16.6	79.1	41.1	19.8	13.3	17.3	20.3	1.0	12.9	19.2	57.4	3,264
Facility Residents	11.7	68.2	30.8	24.2	17.0	11.6	31.9	0.9	17.9	20.4	48.7	55,572
CBA Clients	11.3	70.7	41.0	22.4	4.2	14.3	12.9	0.1	6.8	8.9	74.2	11,284
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files.NOTE: Rows may not sum to 100 because recipients may have more than one diagnosis. 1.All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All residents of nursing facilities who were Medicaid eligible at any point in October 2003. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 6: Percentage of Texas Rider and CBA Medicaid Participants Receiving Assisted Living Services in Six Month Period After Entry into CBA Program
	Rider Participants¹	CBA Clients²
Statewide	33.5%	6.3%
Urban counties	40.2	7.7
Rural counties	19.2	3.9
Region 1: High Plains	43.9	8.1
Region 2 & 9: NW Texas	34.3	5.4
Region 3: Metroplex	47.8	12.1
Region 4: Upper East Texas	48.5	8.9
Region 5: SE Texas	24.7	5.0
Region 6: Gulf Coast	29.4	5.3
Region 7: Central Texas	28.0	6.9
Region 8: Upper South Texas	17.5	5.6
Region 10: Upper Rio Grande	51.5	2.3
Region 11: Lower South Texas	3.8	0.3
Under age 65	36.2	5.2
Age 65+	32.2	6.8
White, non-Hispanic	41.3	10.6
Black, non-Hispanic	21.6	2.3
Hispanic (any race)	9.8	1.0
TILE
Low Care (211)	49.8	6.9
Low Care with Behavioral (210)	53.2	10.4
Moderate Care (207-209)	22.4	3.7
Rehabilitation (202)	27.2	10.2
High Care (204-206)	10.8	2.6
Heavy Care (201, 203)	2.0	0.0
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 7: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Urban/Rural Status
Type of Service	Statewide		Urban Counties		Rural Counties
Type of Service	Rider Participants¹	CBA Clients²	Rider Participants	CBA Clients	Rider Participants	CBA Clients
Assisted Living	33.5%	6.3%	40.2%	7.7%	19.2%	3.9%
Nursing Facility Stay	3.7	6.0	4.2	6.1	2.7	5.8
Nursing Services	95.8	96.3	96.2	95.9	95.1	97.0
Personal Assistance Services	59.7	86.3	53.0	84.9	73.9	88.9
Adaptive Aids	44.8	48.1	42.1	46.0	50.6	51.8
Medical Supplies	43.1	48.1	40.9	48.3	47.9	47.8
Minor Home Modifications	24.9	35.2	21.4	32.7	32.2	39.6
Emergency Response Services	20.8	28.7	16.3	26.4	30.2	32.8
Meals	12.1	19.6	9.0	17.0	18.7	24.2
Therapies (phys, occ, speech)	1.6	0.9	2.0	1.3	0.8	0.1
Respite Care	2.5	2.0	2.5	2.1	2.4	1.7
Adult Foster Care	1.4	0.5	1.6	0.7	0.9	0.3
NUMBER OF RECIPIENTS	2,998	15,611	2,038	10,016	958	5,586
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. NOTE: Columns do not sum to 100 percent since each person may receive multiple services. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 8: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Assisted Living and Urban/Rural Status
Type of Service	Statewide		Urban Counties		Rural Counties
Type of Service	Rider Participants¹	CBA Clients²	Rider Participants	CBA Clients	Rider Participants	CBA Clients
Nursing Facility Stay
Assisted living	3.9%	12.0%	4.4%	12.7%	1.6%	9.7%
No Assisted living	3.7	5.6	4.1	5.6	3.0	5.6
Nursing Services
Assisted living	95.6	95.7	96.2	94.8	92.9	98.6
No Assisted living	95.9	96.3	96.1	95.9	95.6	96.9
Personal Assistance Services
Assisted living	3.1	12.2	2.9	12.2	3.8	12.0
No Assisted living	88.2	91.3	86.7	90.9	90.6	92.0
Adaptive Aids
Assisted living	27.4	26.5	28.2	26.8	23.9	25.5
No Assisted living	53.6	49.5	51.4	47.6	57.0	52.8
Medical Supplies
Assisted living	29.3	33.0	28.6	30.7	32.6	41.2
No Assisted living	50.1	49.1	49.1	49.7	51.6	48.0
Minor Home Modifications
Assisted living	0.6	1.8	0.6	2.1	0.5	0.9
No Assisted living	37.1	37.4	35.4	35.2	39.7	41.1
Emergency Response Services
Assisted living	1.0	5.9	1.1	5.6	0.5	6.9
No Assisted living	30.7	30.2	26.6	28.1	37.2	33.9
Meals
Assisted living	1.0	3.5	1.0	3.0	1.1	5.6
No Assisted living	17.7	20.7	14.4	18.2	22.9	25.0
Therapies (phys, occ, speech)
Assisted living	1.3	1.4	1.6	1.8	0.0	0.0
No Assisted living	1.8	0.8	2.2	1.2	1.0	0.2
Respite Care
Assisted living	0.1	0.2	0.1	0.3	0.0	0.0
No Assisted living	3.7	2.1	4.1	2.3	3.0	1.8
Adult Foster Care
Assisted living	0.3	0.6	0.2	0.7	0.5	0.5
No Assisted living	1.9	0.5	2.5	0.7	1.0	0.3
NUMBER OF RECIPIENTS WITH ASSISTED LIVING	1,003	990	819	772	184	216
NUMBER OF RECIPIENTS WITH NO ASSISTED LIVING	1,995	14,621	1,219	9,244	774	5,370
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. NOTE: Columns do not sum to 100 percent since each person may receive multiple services. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 9: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Region: Recipients with No Assisted Living
	Nursing Facility Stay	Nursing	Personal Assistance	Adaptive Aids	Medical Supplies	Minor Home Modifications	Emergency Response	Meals	Therapies	Respite Care	Adult Foster Care	Number of Participants
Region 1: High Plans
Rider¹	9.4%	96.9%	76.6%	65.6%	51.6%	37.5%	26.6%	4.7%	4.7%	6.3%	7.8%	64
CBA²	7.2	96.8	85.8	55.0	45.2	38.2	42.5	7.2	3.5	2.9	1.7	649
Region 2 & 9: NW Texas
Rider	3.8	97.1	90.0	51.2	58.4	30.1	41.2	28.7	0.5	3.8	2.4	209
CBA	5.1	97.9	90.3	50.6	56.9	31.1	43.0	38.5	0.1	3.2	0.7	1,578
Region 3: Metroplex
Rider	3.8	96.4	80.7	49.1	46.2	23.1	29.5	20.0	3.1	1.4	1.4	420
CBA	6.1	96.2	89.3	50.9	53.2	19.5	38.1	26.2	2.5	0.8	0.7	2,673
Region 4: Upper East Texas
Rider	2.0	90.7	87.8	38.5	42.9	35.1	46.3	31.2	0.5	1.0	2.0	205
CBA	6.1	96.9	88.4	40.7	39.8	34.1	35.6	27.9	0.3	1.1	0.6	1,856
Region 5: SE Texas
Rider	2.3	95.4	90.8	53.4	42.8	31.3	33.6	19.9	0.8	3.8	0.0	131
CBA	5.7	95.4	91.5	41.6	38.2	23.7	35.3	29.2	0.1	3.8	0.2	1,021
Region 6: Gulf Coast
Rider	6.3	95.8	87.5	52.1	52.1	29.2	12.5	2.1	0.0	8.3	0.0	48
CBA	10.8	92.0	94.0	43.0	48.6	28.7	18.4	5.4	1.1	1.9	0.2	463
Region 7: Central TX
Rider	4.4	98.0	94.0	70.4	52.4	35.2	38.6	6.8	0.4	6.4	0.0	250
CBA	7.3	95.4	93.4	59.0	48.4	37.3	39.5	10.2	0.2	3.4	0.1	1,403
Region 8: Upper South TX
Rider	3.2	97.9	90.3	57.4	62.0	48.0	30.8	21.5	3.9	4.7	3.9	279
CBA	5.5	97.6	94.5	56.7	63.1	50.2	26.2	18.9	0.9	3.7	0.7	1,533
Region 10: Upper Rio Grande
Rider	3.1	100.0	59.4	12.5	31.3	31.3	28.3	3.1	0.0	3.1	18.8	32
CBA	1.9	98.6	85.3	11.4	52.9	22.8	20.5	13.3	0.2	1.2	2.8	429
Region 11: Lower South TX
Rider	3.4	94.4	94.4	56.1	46.8	55.2	14.9	10.4	1.1	3.9	0.3	355
CBA	3.7	95.5	94.4	53.0	44.5	60.2	10.0	11.3	0.1	1.3	0.0	3,009
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files.NOTE: Number initalicsrepresent regions with small numbers of participants, which might be less representative than regions with larger numbers of participants. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 10: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Race/Ethnicity: Recipients with No Assisted Living
Type of Service	White, Non-Hispanic		Black, Non-Hispanic		Hispanic (any race)
Type of Service	Rider Participants¹	CBA Clients²	Rider Participants	CBA Clients	Rider Participants	CBA Clients
Nursing Facility Stay	3.0%	7.0%	5.4%	5.2%	4.8%	3.8%
Nursing Services	95.7	96.3	97.1	95.9	96.2	96.6
Personal Assistance Services	86.4	89.4	91.4	93.8	91.3	92.9
Adaptive Aids	53.8	50.5	51.4	45.6	54.2	50.5
Medical Supplies	48.0	49.5	50.6	46.8	54.0	49.8
Minor Home Modifications	34.0	32.1	29.2	29.0	48.8	52.8
Emergency Response Services	36.2	36.1	30.0	35.0	18.9	17.7
Meals	19.1	21.2	22.2	27.2	11.7	16.1
Therapies (phys, occ, speech)	1.7	0.9	2.1	1.5	1.6	0.4
Respite Care	3.6	2.6	1.7	1.4	5.0	1.7
Adult Foster Care	1.7	0.6	2.1	0.5	2.2	0.4
NUMBER OF RECIPIENTS	1,995	7,094	243	2,755	504	4,337
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. NOTES: Columns do not sum to 100 percent since each person may receive multiple services. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 11: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program: Texas Index of Level of Effort (TILE) Class 211, All Recipients and Recipients with No Assisted Living
	TILE 211 (Low Care Dependent and Stable)
	All Recipients		Recipients with No Assisted Living
	Rider Participants¹	CBA Clients²	Rider Participants	CBA Clients
Assisted Living	49.8%	6.9%	--	--
Nursing Facility Stay	4.7	5.2	4.3%	4.7%
Nursing Services	95.7	96.8	96.1	96.7
Personal Assistance Services	43.4	85.2	83.9	90.5
Adaptive Aids	36.7	47.3	46.5	48.8
Medical Supplies	33.5	42.3	40.4	43.1
Minor Home Modifications	14.1	33.0	27.4	35.3
Emergency Response Services	18.7	35.4	36.3	37.5
Meals	11.5	23.0	22.2	24.4
Therapies (phys, occ, speech)	1.2	0.5	0.7	0.5
Respite Care	0.8	0.8	1.7	0.8
Adult Foster Care	1.0	0.4	2.0	0.4
NUMBER OF RECIPIENTS	1,075	6,697	540	6,233
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files.NOTE: Columns do not sum to 100 percent since each person may receive multiple services. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 12: Percentage of Texas Rider and CBA Medicaid Participants Receiving Selected Services in Six Month Period After Entry into CBA Program by Texas Index for Level of Effort (TILE) Class: Recipients with No Assisted Living
Type of Service	TILE Class
	Less Care-Dependent						More Care-Dependent >
	Low CareTILE 211		Low Care with BehavioralTILE 210		Moderate CareTILE 207-209		RehabilitationTILE 202		High CareTILE 204-206		Heavy CareTILE 201, 203
	Rider¹	CBA²	Rider	CBA	Rider	CBA	Rider	CBA	Rider	CBA	Rider	CBA
Nursing Facility Stay	4.3%	4.7%	0.0%	7.4%	4.2%	5.8%	2.6%	15.8%	3.4%	5.3%	3.0%	5.3%
Nursing Services	96.1	96.7	98.0	96.8	95.5	95.8	95.7	93.9	96.6	95.7	93.9	95.7
Personal Assistance Services	83.9	90.5	82.4	91.3	90.5	92.7	89.37	86.0	94.7	94.9	84.9	94.1
Adaptive Aids	46.5	48.8	44.1	50.5	58.5	50.7	58.6	46.5	57.8	51.0	51.5	44.5
Medical Supplies	40.4	43.1	40.2	50.1	56.0	53.6	49.1	47.4	59.7	62.8	52.5	60.5
Minor Home Modifications	27.4	35.3	32.4	40.6	41.2	39.9	42.9	32.5	44.2	40.0	39.4	31.2
Emergency Response Services	36.3	37.5	17.7	26.4	29.2	26.7	40.3	33.3	21.4	18.5	14.1	7.7
Meals	22.2	24.4	11.8	20.4	16.7	19.0	20.9	26.3	13.1	13.8	4.0	4.0
Therapies (phy, occ, speech)	0.7	0.5	0.0	0.5	1.6	0.8	2.0	3.5	1.5	1.5	10.1	3.2
Respite Care	1.7	0.8	2.9	2.3	4.1	2.9	3.4	1.8	4.9	4.2	8.1	6.9
Adult Foster Care	2.0	0.4	5.9	0.7	1.4	0.7	1.7	0.0	1.5	0.1	1.0	0.8
NUMBER OF RECIPIENTS	540	6,233	102	1,718	641	4,422	350	114	206	1,002	99	375
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files.NOTE: Columns do not sum to 100 percent since each person may receive multiple services. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 13: Recipient Status at Six and Twelve Months after Entry into Texas Long-Term Care Medicaid CBA Program: Rider Participants and CBA Clients by Urban/Rural Residence
	Recipient Status at Six Months					Recipient Status at Twelve Months
	Remains on CBA	Residing in Nursing Facility	Reported as Deceased	Left CBA/ Status Unknown	Number of Participants	Remains on CBA	Residing in Nursing Facility	Reported as Deceased	Left CBA/ Status Unknown	Number of Participants
Statewide
Rider Participants	83.5%	1.3%	5.3%	9.9%	2,998	70.6%	0.6%	9.7%	19.0%	1,872
CBA Clients	87.9	5.2	0.01	6.9	15,611	85.1	7.9	0.02	7.0	12,899
Urban Counties
Rider Participants	83.8	1.3	5.0	9.8	2,038	71.3	0.6	9.0	19.1	1,276
CBA Clients	87.3	5.4	0.02	7.3	10,016	84.7	8.0	0.02	7.3	8.177
Rural Counties
Rider Participants	83.0	1.4	5.9	9.8	958	69.2	0.8	11.1	18.9	594
CBA Clients	89.1	4.9	0.0	6.0	5,586	85.7	7.8	0.0	6.5	4,714
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care files.NOTES: Samples for six-month outcomes are restricted to Rider and CBA participants entering CBA prior to October 1, 2003, and for twelve-month outcomes are restricted to those entering prior to April 1, 2003.

TABLE 14: Recipient Status at Six and Twelve Months after Entry into CBA Program by Texas Index for Level of Effort (TILE) Class
	Recipient Status at Six Months					Recipient Status at Twelve Months
	Remains on CBA¹	Residing in Nursing Facility	Reported as Deceased	Left CBA/ Status Unknown	Number of Participants	Remains on CBA¹	Residing in Nursing Facility	Reported as Deceased	Left CBA/ Status Unknown	Number of Participants
Rider Participants
TILE Class
Low Care (211)	86.1%	1.3%	3.2%	9.5%	1,075	75.2%	0.6%	7.0%	17.2%	698
Low Care with Behavioral (210)	83.0	0.9	4.1	11.9	218	71.7	0.7	8.7	18.8	138
Moderate Care (207-209)	81.5	1.7	6.7	10.0	826	67.5	0.8	11.2	20.6	510
High Care (204-206)	77.9	1.7	10.4	10.0	231	65.7	0.7	17.1	15.7	140
Heavy Care (201, 203)	87.1	1.0	7.9	4.0	1.1	74.2	0.0	13.6	12.1	66
Rehabilitation (202)	83.0	0.8	5.0	11.2	481	66.7	0.0	8.7	24.7	288
CBA Clients
TILE Class
Low Care (211)	90.8	4.4	0.0	4.8	6,697	88.0	7.1	0.0	5.0	5,539
Low Care with Behavioral (210)	88.5	6.6	0.0	4.9	1,918	85.2	9.6	0.0	5.2	1,566
Moderate Care (207-209)	86.4	5.4	0.0	8.3	4,591	83.2	8.6	0.0	8.2	3,723
High Care (204-206)	80.2	5.4	0.0	14.4	1,029	78.1	6.7	0.0	15.1	846
Heavy Care (201, 203)	81.6	5.1	0.0	13.3	375	80.8	5.8	0.0	13.3	308
Rehabilitation (202)	74.0	15.0	0.0	11.0	127	72.2	15.5	0.0	12.4	97
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care files.NOTES: Italicsindicate a TILE Class with fewer than 100 recipients. Percentages based on this small group may be less reliable than for other groups. Samples for six-month outcomes are restricted to Rider and CBA participants entering CBA prior to October 1, 2003, and for twelve-month outcomes are restricted to these entering prior to April 1, 2003. Based on evidence of continued CBA service after six months.

TABLE 15: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients
Type of CBA Service	Average Monthly Service Cost per User of Each Service		Percentage of Recipients Using Each Service		Average Monthly Program Expenditure per Recipient
Type of CBA Service	Rider Participants¹	CBA Clients²	Rider Participants	CBA Clients	Rider Participants	CBA Clients
Assisted Living	$1,157	$1,100	33.5%	6.3%	$348	$56
Nursing Services	72	71	95.8	96.3	42	43
Personal Assistance Services	1,063	865	59.7	86.3	521	655
Adaptive Aids	566	606	44.8	48.0	56	63
Medical Supplies	119	116	43.2	48.1	26	34
Minor Home Modifications	2,013	1,926	24.9	35.2	123	166
Meals	96	90	12.1	19.6	9	15
Total Average Monthly Expenditure per Recipient
For seven primary services above					1,125	1,030
For all CBA services					1,143	1,043
NUMBER OF RECIPIENTS	2,998	15,611	2,998	15,611	Per month	Per month
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. NOTE: Payments were adjusted to 2003 reimbursement levels. Costs for recipients with less than six months of data are averaged over available months. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 16: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients with Assisted Living
Type of CBA Service	Average Monthly Service Cost per User of Each Service		Percentage of Recipients Using Each Service		Average Monthly Program Expenditure per Recipient
Type of CBA Service	Rider Participants¹	CBA Clients²	Rider Participants	CBA Clients	Rider Participants	CBA Clients
Recipients with Assisted Living
Assisted Living	$1,152	$1,101	100%	100%	$1,152	$1,202
Nursing Services	53	56	95.6	95.7	33	33
Personal Assistance Services	455	265	3.1	12.2	2	5
Adaptive Aids	548	568	27.4	26.5	34	31
Medical Supplies	120	115	29.3	33.0	18	22
Minor Home Modifications	495	1,002	0.6	1.9	0	1
Meals	77	40	1.0	3.5	0	0
Total Average Monthly Expenditure per Recipient
For seven primary services above					1,239	1,192
For all CBA services					1,241	1,195
NUMBER OF RECIPIENTS					1,003	988
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. NOTE: Payments were adjusted to 2003 reimbursement levels. Costs for recipients with less than six months of data are averaged over available months. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 17: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients with No Assisted Living
Type of CBA Service	Average Monthly Service Cost per User of Each Service		Percentage of Recipients Using Each Service		Average Monthly Program Expenditure per Recipient
Type of CBA Service	Rider Participants¹	CBA Clients²	Rider Participants	CBA Clients	Rider Participants	CBA Clients
Recipients with No Assisted Living
Assisted Living	$0	$0	0.0%	0.0%	--	--
Nursing Services	79	71	95.9	96.3	$46	$43
Personal Assistance Services	1,063	866	88.2	91.3	734	689
Adaptive Aids	584	607	53.5	49.5	66	65
Medical Supplies	119	116	50.1	49.1	30	34
Minor Home Modifications	2,009	1,927	37.0	37.4	176	174
Meals	95	90	17.7	20.7	12	15
Total Average Monthly Expenditure per Recipient
For seven primary services above					1,064	1,021
For all CBA services					1,089	1,034
NUMBER OF RECIPIENTS					1,995	14,621
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files. NOTE: Payments were adjusted to 2003 reimbursement levels. Costs for recipients with less than six months of data are averaged over available months. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 18: Average Monthly Costs for Each CBA Service and All CBA Services in Six Months After Entry into CBA Program: Rider Participants and CBA Clients by Texas Index for Level of Effort (TILE) Class
Type of CBA Service	TILE Class
	Less Care-Dependent						More Care-Dependent >
	Low CareTILE 211		Low Care with BehavioralTILE 210		Moderate CareTILE 207-209		RehabilitationTILE 202		High CareTILE 204-206		Heavy CareTILE 201, 203
	Rider¹	CBA²	Rider	CBA	Rider	CBA	Rider	CBA	Rider	CBA	Rider	CBA
Assisted Living	$515	$59	$512	$86	$220	$31	$308	$99	$120	$29	$32	$0
Nursing Services	37	39	36	41	39	38	38	36	41	45	151	180
Personal Assistance Services	290	554	276	617	659	762	635	613	852	884	1,029	1,097
Adaptive Aids	41	58	31	55	72	72	63	61	77	69	77	94
Medical Supplies	17	26	18	33	33	39	29	35	42	53	36	67
Minor Home Modifications	52	146	53	163	167	195	173	143	214	207	230	160
Meals	9	17	4	15	9	14	11	16	9	10	3	3
Total Average Monthly Expenditure per Recipient
For seven primary services above	960	898	930	1,010	1,200	1,152	1,257	1,004	1,356	1,297	1,557	1,600
For all CBA services	973	909	948	1,022	1,218	1,166	1,279	1,012	1,369	1,308	1,617	1,620
NUMBER OF RECIPIENTS	1,080	6,601	218	1,887	835	4,505	483	122	231	1,002	101	367
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care data files.NOTE: All payments were adjusted to 2003 reimbursement levels. Costs for recipients with less than six months of data are averaged over available months. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004. All CBA clients entering CBA program between September 1, 2001 and October 1, 2003.

TABLE 19: Factors Associated with a Possible Demand Response by Texas Rider 37/28 Medicaid Participants September 2001 - April 2004
	Percent of Rider Participants¹
	Statewide	Urban Counties	Rural Counties
Factor:
Nursing facility stay 40 days before CBA entry	10.1%	8.4%	13.9%
TILE 211 at admission to nursing facility	31.4	31.4	31.4
TILE 202 at admission to nursing facility	17.1	17.1	17.0
Entered assisted living facility after CBA entry	34.6	41.1	20.2
Combination of factors:
(1) Nursing facility stay 40 days and entered assisted living within 30 days	3.1	3.6	2.1
(2) TILE 202 at nursing facility admission and nursing facility stay 40 days	1.4	1.1	2.2
(3) TILE 211 at nursing facility admission and nursing facility stay 40 days	3.4	2.8	4.6
Any of (1), (2), or (3) above	6.0	5.2	7.6
SOURCE: Urban Institute analysis of Texas Medicaid Long-Term Care files. All Rider 37 and 28 clients who transitioned to the community between September 1, 2001 and April 30, 2004.

Files

TXltc.pdf (pdf, 1.6 MB)

Topics

Long-Term Services & Supports (LTSS) | Long-Term Care Financing

Location- & Geography-Based Data

State Data

Program

Medicaid

Examination of Texas Rider 37: A Medicaid "Money Follows the Person" Long-Term Care Initiative

Examination of Texas Rider 37: A Medicaid "Money Follows the Person" Long-Term Care Initiative

Barbara A. Ormond, Anna S. Sommers and Kirsten J. Black

The Urban Institute

May 30, 2006

TABLE OF CONTENTS

EXECUTIVE SUMMARY

Background

Study Objectives

Data and Methods

Findings: Overview

Findings: Process

Findings: Rider Participants and Comparison Groups

INTRODUCTION

Overview of Rider 37

Study Objectives

Data and Methods

FINDINGS

Components of the Rider 37 Initiative

Rider Participants and Comparison Groups

Nursing Facility Industry Response to the Initiative

Service Costs and Program Expenditures for Medicaid Long-Term Care

SUMMARY

APPENDIX 1: DATA AND METHODS

APPENDIX 2: TILE CLASSES

ENDNOTES

TABLES

ASPE.hhs.gov

Examination of Texas Rider 37: A Medicaid "Money Follows the Person" Long-Term Care Initiative

Examination of Texas Rider 37: A Medicaid "Money Follows the Person" Long-Term Care Initiative

Barbara A. Ormond, Anna S. Sommers and Kirsten J. Black

The Urban Institute

May 30, 2006

TABLE OF CONTENTS

Background

Study Objectives

Data and Methods

Findings: Overview

Findings: Process

Findings: Rider Participants and Comparison Groups

Overview of Rider 37

Study Objectives

Data and Methods

Components of the Rider 37 Initiative

Rider Participants and Comparison Groups

Nursing Facility Industry Response to the Initiative

Service Costs and Program Expenditures for Medicaid Long-Term Care

Connect with Us