Florida's Medicaid AIDS waiver: an assessment of dimensions of quality.

Some State Medicaid agencies have implemented home and community-based waiver programs targeting acquired immunodeficiency syndrome (AIDS) patients. Under these initiatives, State Medicaid agencies can provide home and community-base services to persons with AIDS (PWA) as an alternative to more costly Medicaid-covered institutional care. This article evaluates quality of care under the Florida Medicaid waiver for PWA along two dimensions: program effectiveness and client satisfaction. Clients are generally satisfied with their case managers and the range and availability of services. Case managers appear to be well trained. Moreover, the probability of turnover is quite low, despite heavy caseloads and high mortality. The major difficulty faced by clients adn case managers relates to the process of becoming Medicaid eligible.

Introduction

In an attempt to meet the challenge of providing adequate health care services to an increasing number of AIDS patients under budget constraints, some State Medicaid agencies have implemented waiver programs which make available a targeted set of home and community-based services. These initiatives, authorized under section 2176 of the Omnibus Budget Reconciliation Act (OBRA) of 1981, enable State Medicaid agencies to provide home and community-based services to patients as an alternative to more costly Medicaid-covered institutional care (Miller, 1992). This policy option is regarded as a preferable, more humane mode of treatment, is potentially more cost effective, and may even yield improvements in quality of life for PWA if the home and community-based services are used in lieu of inpatient care (Jacobson, Lindsey, and Pascal, 1989).

Because most waivers targeting PWA were only recently implemented, empirical analyses as to the overall effects of such initiatives on access, quality, costs, and survival are rare. Other than our own published (Cowart, Mitchell, and Meyer, 1994) and unpublished (Mitchell and Anderson, 1994; Mitchell and Cowart, 1994) research, the only other published evidence to date on waivers for PWA is a descriptive study of the New Jersey waiver program by Merzel et al. (1992). They estimated the average monthly treatment costs per Medicaid enrollee in the New Jersey waiver program at $2,400. They compared their estimates with the $5,000 average monthly treatment costs for PWA projected by previous studies. Based on these comparisons, they concluded the costs of treating waiver enrollees in New Jersey was substantially lower than other treatment protocols (Bloom and Carliner, 1988; Hellinger, 1988, 1990, 1991, 1992; Andrews et al., 1991; Bennett et al., 1992; Hiatt et al., 1990; Scitovsky and Rice, 1987). None of these research efforts, however, examined whether the waiver program targeted at PWA could be deemed effective in terms of quality of care and client satisfaction. These measures must be considered in any evaluation of the costs and benefits of such a program.

In this article, we evaluate quality of care under the Florida Medicaid waiver program—Project AIDS Care. Evidence indicating that Project AIDS Care enrollees incur lower monthly Medicaid expenditures relative to non-enrollees has been documented elsewhere (Mitchell and Anderson, 1994). The two dimensions of quality we consider here are program effectiveness and client satisfaction. Data were collected by telephone interviews with clients or their informal caregivers, all case managers, and select providers in the Project AIDS Care program. Following Donabedian's (1980) classic conceptual framework for evaluating quality of care, we obtained data on structure and process indicators to evaluate the program and relied on client satisfaction to measure one dimension of outcome. Structure variables included the background, experience, and workload of case managers, provider availability, and adequacy of client records. Process variables included implementation of procedures related to obtaining Medicaid eligibility, waiver enrollment, provision of services, documentation, and training. We also obtained data on client or caregiver satisfaction with the choice and quality of services received. To our knowledge, this article is the first to document some dimensions of quality of care provided under Medicaid home and community-based waiver programs targeted to PWA.

Home and Community-Based Waiver Programs

Under section 2176 of OBRA 1981, State Medicaid agencies are permitted to offer an array of home and community-based services through a specific waiver program as an alternative to more expensive institutionalized care. A major advantage of the waiver is that specific Medicaid services can be targeted to selected patient groups and geographic locations rather than being delivered throughout the State. The aged, disabled, mentally retarded, and mentally ill qualified for waiver services under OBRA 1981. PWA could have been served under waivers which served the disabled as an alternative to nursing facility (NF) care. Persons with specific diseases or conditions (including PWA) became specifically eligible for waiver services under OBRA 1986 if they were inpatients in an NF or hospital. The Technical and Miscellaneous Revenue Act of 1988 extended waiver eligibility to persons with specific diseases or conditions (including PWA) who were not inpatients in an NF or hospital but did require NF- or hospital-level care.

Project AIDS Care became operational in January 1990, and is administered by the Project AIDS Care Division, a small staff of four professionals in the Medicaid Managed Care Office. The four staff members oversee and administer all aspects of the program, which includes 45 case-management agencies throughout the State. Under this waiver, the Medicaid program is able to pay for specific home and community-based services for disabled PWA who are deemed to be at risk of institutionalization in either a hospital or an NF. The waiver program is required to operate under specific criteria to conform to Federal guidelines. Among these criteria are: caps on the number of recipients enrolled, voluntary participation, and the provision of services under a written plan of care. Participants in Project AIDS Care are eligible to receive services traditionally covered by Medicaid as well as case management and an additional set of waiver services. The available waiver services include: homemaker services; specialized personal care to foster children; home-delivered meals; day health care; respite care; skilled services performed by a licensed nurse or respiratory therapist; chores, such as minor repairs and heavy cleaning; adaptive equipment; home modifications; consumable medical supplies; substance-abuse treatment; and education, support services, and mental health counseling.

Conceptual Framework

Donabedian (1980) developed the now-classic model for classifying information from which inferences about quality of care can be drawn. These measures are structure, process, and outcome. His definition includes satisfaction as an outcome measure: “Patient satisfaction may be considered to be one of the desired outcomes of care, even an element in health status itself. An expression of satisfaction or dissatisfaction is also the patient's judgment on the quality of care in all its aspects, but particularly as concerns the interpersonal process. It is futile to argue about the validity of patient satisfaction as a measure of quality. Whatever its strengths and limitations as an indicator of quality, information about patient satisfaction should be as indispensable to assessments of quality as to the design and management of health care systems” (Donabedian, 1990).

Subsequently, researchers further refined this model by expanding Donabedian's three classifications into four, the last being satisfaction. The justification for this extension was that patient or caregiver satisfaction interacts with all three original categories—structure, process, and outcome (Rosenberg, 1990). Aday and Shortell (1988) further define concepts involved in evaluation derived from Donabedian and others by separating the definition of quality into two attributes: (1) the correct protocol or approach adopted; and (2) the approach or protocol correctly implemented. Following Donobedian's conceptual model to measure quality of care as refined by Aday and Shortell, we obtained data on structure and process indicators to evaluate Project AIDS Care. We chose the classic model that includes satisfaction as one measure of outcome because it is suited to the chronic and debilitating nature of AIDS. Satisfaction (or dissatisfaction) seem more appropriate as an outcome measure for PWA when considered relative to other measures of quality, such as mortality or disability.

Methods

Previous studies based on surveys of PWA have identified methodological problems including confidentiality, cognitive defects secondary to AIDS, fatigue, and death prior to completion of the study (Fowler, Massagli, and Weissman, 1992). We adapted our data-collection protocol and procedures to minimize these effects. We selected computer-assisted telephone interviewing (CATI) because of high interview response rates, ability to assure confidentiality, flexibility in arranging convenient appointment times, and clients' choice of a setting for the interview. Further, telephone interviews are less fatiguing, allow for the coverage of a large number of questions in a relatively short time, and tend to be less expensive in time and personnel than other methods of data collection. Moreover, reliability controls are relatively easy to establish and enforce through good training of experienced interviewers and subsequent supervision during the interview process.

The Florida State University Survey Research Laboratory conducted the telephone interviews using CATI. This system minimizes the effects of both interviewer and data-entry error because it automatically routes the interviewer through the questionnaire. On closed-ended questions, all acceptable answers were programmed into the CATI survey instrument. Open-ended responses were recorded verbatim. Other standard procedures for telephone surveying were used: 8-12 hours of interviewer training, availability of bilingual interviewers, a maximum of 7 call-backs, and simultaneous computer access of data.

Sample Characteristics

Interviews were conducted with 3 cohorts: (1) a random sample of clients (and/or their informal caregivers) who were enrolled in Project AIDS Care from January through August 1992 and still active in the fall of 1992; (2) case managers employed by the 45 Project AIDS Care agencies in August 1992; and (3) approved referral providers for Project AIDS Care who were active during August 1992. An initial random sample of recipients (and/or their informal caregivers) (n = 600) was drawn from the active client list of 1992 enrollees. The list was reviewed to assure representation according to the geographic distribution of clients within the State. Since all high-density geographic areas were adequately represented, oversampling techniques were not necessary. To account for the high mortality rate in PWA, we asked the local agency case-manager supervisor to substitute a matched client for each client on the master list who was no longer with the agency either due to death or moving outside of the area. Substitute patients were matched by age, race, gender, and month of enrollment. Any bias arising from this substitution is apt to be minimal, because this procedure was employed only to retain a distribution of geographic and select personal characteristics identical to the original random sample.

Special procedures for confidentiality were instituted. The assurance of confidentiality was critical, and contributed to the high degree of participation among clients and caregivers. For all clients, the Project AIDS Care enrollment number, case-management agency, first name, and phone number were the only identifying information available to the researchers. The supervisory case managers contacted each client in the sample, gave a brief description of the purpose and content of the study, obtained verbal permission for participation in the study, determined the client's preference for interview site (home or case-manager's office), and scheduled a convenient time for the interview. The procedure for generating the sample was repeated until 125 responses were obtained. The final sample of 125, drawn from the original population of 2,500, has a sampling error of about 8 percent. We had to balance sampling error against the costs of conducting each interview, as funding constrained the number of interviews that could be conducted.

Case managers had phones installed in the homes of clients without them, since phone contact is essential if clients are to remain living independently. The telephone survey also included a second informed consent at the beginning of the interview. In addition, when clients preferred, were minors, or appeared to be cognitively impaired, informal caregivers or significant others were allowed to respond using an interview schedule that mirrored the client instrument. These strategies are in accordance with the suggestions for interviews discussed by Fowler, Massagli, and Weissman (1992) and Mishra et al. (1993).

The total population of supervisory case managers (n = 45) was interviewed. Staff case managers were also interviewed, when accessible, for a total of 118 interviews of supervisory and staff case managers. To enhance the response rate, the Project AIDS Care program administrator contacted case managers in advance to inform them of the survey. Research staff then telephoned case managers at a prearranged time. Case managers were asked for names and phone numbers of frequent referral providers, as well as the names of personnel where referral was unsuccessful. Telephone interviews were conducted with a select convenience sample of these referral providers (n = 27); all agreed to participate.

Instruments

Interview schedules were developed for use with case managers, select referral providers, and clients (or their informal caregivers). The interview schedules for use with case managers and select referral providers attempted to determine whether the approach was appropriate by obtaining their impressions of each step of the process to enroll and provide services for their clients. The interview schedule for use with clients or their caregivers attempted to determine if the program was operating in accordance with their expectations, as measured by their impressions of each step of the process, knowledge, and satisfaction. We obtained the specific content for the interview schedules from program staff, manuals, and policies and procedures. The variables for each instrument were derived from discussion with program staff, policies and procedures, and the provider handbook. The instruments were reviewed by the research team, program staff, and select case managers for content validity. Each instrument was pretested on a small group of respondents.

Following Donabedian's (1980) conceptual framework to measure quality of care, we obtained data on structure and process indicators to evaluate Project AIDS Care; the only outcome measure was patient satisfaction. Structural questions relate to the case-manager background, experience, and workload, provider availability, and adequacy of client records. Process questions include implementation of procedures related to obtaining Medicaid eligibility, enrollment in the Project AIDS Care program, services provision, documentation, and training. As noted, the one indicator of outcome considered was client satisfaction. Outcome questions include implementation of clients' freedom of choice, client-rights information, confidentiality, establishing effective client/case-manager relationships, and client or caregiver satisfaction with choice and quality of services received.

Results

Client/Caregiver Interviews

The sample of 125 clients (or their informal caregivers) provided information about their perceptions of the Project AIDS Care program. The respondents included 5 male caregivers, 17 female caregivers, 88 male clients, and 15 female clients. About 60 percent of the male clients interviewed were Caucasian and between 18-50 years of age.

Case management is the core service of the Project AIDS Care program. Both as a measure of the success of the program and as an outcome measure of quality of care, we were primarily interested in client/caregiver opinions about the effectiveness of the case managers. Virtually all clients or their caregivers (92.8 percent) reported that they had established an effective working relationship with their assigned case manager. Further, the majority felt their case managers were familiar with the Project AIDS Care program and the available referral resource services.

Clients received a range of available services under the State Medicaid plan and the Medicaid home and community-based services waiver (Table 1). The most frequently used Medicaid services included pharmacy (87.2 percent), physician (78.4 percent), home health (60 percent), and inpatient hospital care (50.4 percent). With respect to Project AIDS Care services, nearly all clients received case management Other services used frequently were: consumable medical supplies (62.4 percent), education and support (48.8 percent), and homemaker services (44 percent).

Table 1

Services Received by Clients in Florida's Project AIDS Care Program: 1992

Service	Number	Percent
Medicaid Services
Pharmacy	109	87.2
Physician	98	78.4
Home Health Care	75	60.0
Inpatient Hospital	63	50.4
Advanced Registered Nurse Practitioner	51	40.8
Optometry	42	33.6
Durable Medical Equipment	42	33.6
Personal Care	23	18.4
Nursing Facility	18	14.4
Obstetrics (Women Only)	4	3.2
Project AIDS Care Services
Case Management	122	97.6
Consumable Medical Supplies	78	62.4
Education and Support	61	48.8
Homemaker	55	44.0
Skilled Care	41	32.8
Chore Services	31	24.8
Home-Delivered Meals	26	20.8
Adaptive Equipment/Home Modification	23	18.4
Respite Care	19	15.2
Day Health Care	13	10.4
Special Personal Care/Foster Child	6	4.8
Home Drug Abuse Treatment	3	2.4

NOTES: n = 125. AIDS is acquired immunodeficiency syndrome.

SOURCE: Cowart, M.E., and Mitchell, J.M.: Telephone surveys conducted by the Florida State University Survey Research Laboratory, summer 1992.

We also wanted to elicit client/caregiver perceptions about familiarity with important aspects of the program, difficulty with enrollment into the program, and satisfaction with the mix and quality of services delivered (Table 2). Most felt that the services they received were appropriate to their needs (94.4 percent), they had a choice of services (74.4 percent), they participated in the development of a plan of care (75.2 percent), and they were familiar with the participant rights and responsibilities document (60.8 percent). Most (82.0 percent) recalled receiving the document during their first encounter with the case manager or upon becoming enrolled in the waiver program. A majority also indicated that client rights and responsibilities were explained to them by their case manager. Over one-half of the clients reported little difficulty with the enrollment and eligibility process for Project AIDS Care.

Table 2

Client Experience With Florida's Project AIDS Care Program: 1992

Measure of Satisfaction	Yes		No
	Number	Percent	Number	Percent
Appropriate Services	118	94.4	4	3.2
Choice of Enrollment in Project AIDS Care	103	82.4	10	8.0
Satisfied With Standard of Care	103	82.4	22	17.6
Read Plan of Care	96	76.8	20	16.0
Satisfied With Services	95	76.0	28	22.4
Participate in Plan of Care	94	75.2	27	21.6
Choice of Service Providers	93	74.4	25	20.0
Look at Patient Records	79	63.2	6	4.8
Familiarity With Participant Rights and Responsibilities	76	60.8	47	37.6
Did You Find the Following Somewhat or Very Difficult?
Obtaining Services	41	32.8	84	67.2
Eligibility Process	16	12.8	67	53.6
Enrollment Process in Project AIDS Care	9	7.2	79	63.2

NOTES: n = 125. AIDS is acquired immunodeficiency syndrome.

SOURCE: Cowart, M.E., and Mitchell, J.M.: Telephone surveys conducted by the Florida State University Survey Research Laboratory, summer 1992.

Forty-one clients (32.8 percent) reported that they had some difficulty in receiving provider services (Table 2). Service providers involved included physicians (n = 5), homemakers (n = 4), home-care workers (n = 3), respite (n = 3), skilled care (n = 3), consumable medical supplies (n = 3) and optometrists (n = 1). Twenty-two clients (17.6 percent) reported concerns about the quality of care received from their nurse practitioner (n = 4), physician (n = 3), home care services (n = 3), case management (n = 3), consumable medical supplies (n = 2), and chore services (n = 1). The nature of these difficulties were revealed in the open-ended comments by clients. Personality conflicts between the client and health care professionals (physicians, nurses) sometimes hampered the ability to feel satisfied with the care received. The ability of clients to receive services was also impeded if the home health care worker's transportation mode was malfunctioning.

Case-Manager Interviews

A major part of the quality assessment was conducted through telephone interviews with current case managers. One hundred eighteen case managers were successfully interviewed; this represents about 95 percent of the cohort of current case managers (Table 3). Virtually all case managers (97.5 percent) reported that their agency followed the Medicaid guidelines for the Project AIDS Care program. The cohort of case managers is predominantly female, between 25-54 years of age, and racially representative of the State population. The minimum education is a baccalaureate degree in one of the human-service disciplines (i.e., social work, nursing, or psychology). Most had previous experience in case management, and nearly all had some prior experience in human-services delivery. It should be noted that AIDS-waiver case managers appear well trained compared with case managers working in the aging and adult network in Florida. Only one-half of the case managers in the latter program had formal social work preparation at the baccalaureate level, and only one-fourth had 3 or more years of experience in their current position (Cowart, 1993).

Table 3

Characteristics of Florida's Project AIDS Care Case Managers: 1992

Characteristic	Number	Percent
Length of Time With Program
<1 Year	35	29.7
1-2 Years	80	67.8
Unknown	3	2.5
Race
African-American	25	21.2
Native American	1	0.8
Caucasian	92	78.0
Spanish Origin (n = 9)
Cuban	5	4.2
Puerto Rican	4	3.4
Other Hispanic	3	2.5
Age
23 Years	1	0.8
25-34 Years	38	32.2
35-44 Years	44	37.3
45-54 Years	23	19.5
55 Years or Over	10	8.5
Unknown	2	1.7
Gender
Female	81	68.6
Male	29	24.6
Unknown	8	6.8
Highest Education
12th Grade	2	1.7
Associate's	8	6.8
Bachelor's	77	73.7
Master's	26	22.0
Doctorate	3	2.5
Unknown	2	1.7
Years as Project AIDS Care Case Manager
1-2	14	11.9
3 or More	51	43.2
Previous Case-Management Experience	66	55.9
Previous Patient-Care Experience	108	91.5
Work Status
Full-Time (30-40 Hours/Week)	113	95.8
Part-Time (Under 30 Hours/Week)	5	4.2

NOTES: n = 118. AIDS is acquired immunodeficiency syndrome. The “unknown” category includes both “don't know” and refusal to respond.

SOURCE: Cowart, M.E., and Mitchell, J.M.: Telephone surveys conducted by the Florida State University Survey Research Laboratory, summer 1992.

Case-manager workload is an important consideration in the delivery of quality, accessible services. Program guidelines recommended 35 as the maximum number of clients to be served by each case manager. Case-manager workload was difficult to accurately measure, due to travel time to clients' homes and high mortality rates among the Project AIDS Care clients. We did not seek excessive details regarding the amount of travel required of case managers in order to care for their clients. Nonetheless, about one-fourth indicated they spent more than 25 percent of their time traveling, whereas just under 50 percent reported that their travel time amounted to less than 15 percent of each work week. Determining the exact number of clients in each case manager's caseload was difficult for two reasons. First, each case manager carried a number of clients awaiting Medicaid eligibility, a period which varied from 1 week to 2 months. Second, case managers lost clients from deaths and transfers out of the service area. For example, one case manager reported that seven of her clients died in 1 month. Since caseload numbers fluctuate, counting the 1-year total of unduplicated clients is not an accurate representation of each manager's caseload. Nonetheless, 70 percent of the case managers had 35 or fewer clients, 15 percent had between 36 and 50 clients, and the remaining 15 percent had more than 50 clients at any point in time.

As an unobtrusive measure of the quality of services delivered, we elicited information on case managers' perceptions of difficulty in carrying out the requirements of the Medicaid AIDS waiver. These questions attempted to target problematic areas of the program. The following areas were assessed: enrolling providers, screening, explaining the program to clients, obtaining Medicaid eligibility and disability determination, physician referral, home safety, and other requirements. The case managers carried out most functions related to the program requirements without difficulty. Exceptions included obtaining Medicaid eligibility (60.2 percent) and disability determination (55.9 percent), two functions outside the jurisdiction of the Project AIDS Care office. The most difficult function for the Project AIDS Care program was enrolling providers (39.0 percent).

Case managers generally felt that the process for enrolling clients into the program was less efficient. Most case managers reported no difficulty with conducting client prescreening, providing the client with information about Project AIDS Care participant rights and responsibilities, or providing the plan of care in a timely manner. Nevertheless, time to obtain Medicaid eligibility varied: 1-10 days (3.3 percent), 11-20 days (10.1 percent), 21-30 days (21.2 percent), 31-60 days (38.9 percent), or more than 60 days (11.7 percent). Length of time to determine Medicaid eligibility also varied by geographical region. Considering the severity of illness of AIDS clients, an extensive delay may be too long to avoid hospitalization or NF admission.

One of the requirements of the program was the provision of written documentation of client eligibility, assessment, and progress with services delivery. In general, case managers were attentive to initiating and maintaining client records in accordance with preestablished standards. They included information about client services (100.0 percent), direct conversations with clients (98.3 percent), direct conversations with providers (95.8 percent), needs assessment (97.5 percent), services provided (97.5 percent), and clients' choice of providers (94.5 percent). Direct conversations with supervisory case managers revealed that agency case managers worked together to improve the standard of their records through peer review, revision of forms, and chart organization. Record review by one of the authors during site visits to 10 agencies corroborated the responses reported by case managers. The case managers also indicated that maintenance of confidentiality was well established (97.5 percent) and that, when necessary, it was relatively easy to retrieve records of discharged clients (90.7 percent).

Because this is a relatively new program, virtually all case managers had been with the program for 2 years or less, although little turnover was noted (Table 3). More than one-half of the case-manager orientation was conducted on site, whereas nearly 47 percent of the instruction was delivered by the central office staff. Both aspects of training allow for program consistency across the State (Table 4). After “train-the-trainer” programming was instituted, orientation was provided on site, particularly in large agencies. Most case managers reported that their orientation and preservice training was somewhat or very adequate, whereas only 10 percent stated that their orientation and preservice experience was inadequate. About one-half of the case managers indicated that they had no immediate additional training needs. Those who indicated that additional training would be of value cited the following as areas of concern: services cost estimation; documentation, including writing care plans and problem recording; and skills in working with friends and family members of clients. From the responses, we infer that providing confidentiality and working directly with PWA were skills all but a few case managers felt they had mastered.

Table 4

Type and Quality of Orientation, Pre-Service Training, and Training Needs for Florida's Project AIDS Care Case Managers: 1992

Variable	Number	Percent
Type of Orientation (n = 96)
Central Office	44	45.8
On-Site	52	54.2
Adequacy of Orientation (n = 96)
Very Adequate	47	49.0
Somewhat Adequate	39	40.6
Inadequate/Adequate	7	7.3
Very Inadequate	3	3.1
Adequacy of Pre-Service Training (n = 58)
Very Adequate	25	43.1
Somewhat Adequate	26	44.8
Inadequate/Adequate	6	10.3
Very Inadequate	1	1.7
Additional Training Needs (n = 69)
No Additional Needs	49	41.5
Services Cost Estimation	39	33.1
Documentation	31	26.3
Problem-Oriented Medical Record	30	25.4
Writing Care Plans	29	24.6
Working With Family/Friends	29	24.6
Assess Level-of-Care Requirements	28	23.7
Client Assessment	20	20.2
Ethics	23	19.5
Working With Persons With AIDS	16	13.6
Confidentiality	10	8.5
Eligibility Process	7	5.9

NOTES: n = 118. AIDS is acquired immunodeficiency syndrome.

SOURCE: Cowart, M.E., and Mitchell, J.M.: Telephone surveys conducted by the Florida State University Survey Research Laboratory, summer 1992.

To deter institutionalization of PWA, a network of home and community-based services is essential. Rather than examine a central file listing available referral agencies and services, we queried the case managers about the availability of agencies and services under the program. Nearly all case managers reported they had ready access to State Medicaid plan and home and community-based waiver services; including physicians (80.0 percent), home care (94.2 percent), home-delivered meals (62.4 percent), adaptive equipment (74.1 percent), medical supplies (95.3 percent), education and support (85.9 percent), case management (92.9 percent), and pharmacy services (95.3 percent).

Another factor which hampered access to services appears to be the required administrative review by State staff when the cost of client services were expected to exceed $1,500 per month. This amount was established by the State office so that the program would be in accordance with the statutory requirement that the program demonstrate cost neutrality. By imposing this limit, the State office could monitor high-cost clients and attempt to arrange alternative sources of care for them. Reports of case managers substantiated this fact, as three-fourths indicated that they had clients who required more than $1,500 worth of services per month. Of the 71 case managers who indicated that they asked for exceptions to the $1,500 per month service limit, 16 had 1-2 clients, 43 had 3-10 clients, and 18 had more than 10 clients. This finding underscores the high costs associated with this disease.

Service-Provider Interviews

We asked each case manager for the name of an approved referral agency and contact person on their referral list, as well as the name of an agency and contact person which had refused to become a Project AIDS Care agency. Telephone interviews were conducted with 27 of these referral agencies to inquire about their perceptions of the program. Almost 63 percent of the reporting agencies were for-profit enterprises, another 22 percent were non-profit, and the remaining 15 percent were publicly owned. The provider respondents were geographically dispersed and offered the entire range of services available to Project AIDS Care clients (Table 5).

Table 5

Characteristics of Florida's Project AIDS Care Referral Providers: 1992

Characteristic	Number	Percent
Agency Ownership
Private For-Profit	17	62.9
Non-Profit	6	22.3
Public	4	14.8
Respondent Title
Administrator	3	11.1
Director	3	11.1
Other	21	77.8
Respondent Gender
Male	13	48.1
Female	14	51.9
Services Provided
Skilled Care	13	48.1
Personal Care	13	48.1
Homemaker	10	37.0
Respite Care	10	37.0
Consumable Medical Supplies	9	33.3
Education and Support	8	29.6
Adaptive Equipment/Home Modification	7	25.8
Case Management	6	22.2
Chore Services	5	18.5
Home-Delivered Meals	3	11.1
Day Health Care	3	11.1
Home-Based Drug-Abuse Treatment	3	11.1

NOTES: n = 27. AIDS is acquired immunodeficiency syndrome.

SOURCE: Cowart, M.E., and Mitchell, J.M.: Telephone surveys conducted by the Florida State University Survey Research Laboratory, summer 1992.

We asked the provider respondents for their opinions regarding the weaknesses of Project AIDS Care. All but two respondents were familiar with the Project AIDS Care program, and all were providers of services to the waiver program. One respondent had elected not to participate as a provider due to the level of reimbursement and quantity of paper work. We further asked the agency provider participants in Project AIDS Care about reimbursement levels, time lag between service delivery and reimbursement, required paper work, and the process of becoming eligible as a service provider. Less than one-fourth indicated that these factors affected their decision to accept Project AIDS Care clients (Table 6). Finally, we inquired about the range of clients the provider agencies served. Only one agency imposed restrictions on their clients, in that it did not render services to clients who resided outside the provider area.

Table 6

Florida's Project AIDS Care Referral Provider Opinions: 1992

Question	A Lot		Some		A Little		Not at All
	Number	Percent	Number	Percent	Number	Percent	Number	Percent
Does the level of reimbursement affect your decision to accept Project AIDS Care clients?	4	15	2	7	0	0	21	78
Does the amount of time elapsed between services delivered and date of payment affect your decision to accept Project AIDS Care clients?	2	7	3	11	0	0	20	74
To what extent does paperwork affect your decision to accept Project AIDS Care clients?	2	7	3	11	2	7	19	70
To what extent does the process of becoming eligible to be a Medicaid provider affect your decision to accept Project AIDS Care clients?	0	0	1	4	2	7	24	89

NOTES: n = 27. AIDS s acquired immunodeficiency syndrome.

SOURCE: Cowart, M.E., and Mitchell, J.M.: Telephone surveys conducted by the Florida State University Survey Research Laboratory, summer 1992.

Discussion

State Medicaid agencies are currently faced with the difficult challenge of providing adequate health care services to a rapidly increasing number of PWA. In an attempt to meet these needs in a cost-effective way, some States have enacted home and community-based waiver programs. Such waiver initiatives allow State Medicaid agencies to provide home and community-based services to PWA as an alternative to more expensive Medicaid-covered institutional care. Nonetheless, because most Medicaid waivers for PWA were only recently implemented, empirical analyses evaluating the overall effects of such initiatives on access, quality of care, costs, and survival are rare.

In this article, we evaluated quality under the Florida Medicaid AIDS waiver—Project AIDS Care—along two dimensions: program effectiveness and client satisfaction. The data were collected by telephone interviews with clients or their caregivers, all case managers, and select providers in the Project AIDS Care program. Following Donabedian's (1980) conceptual framework for quality of care, we obtained data on structure and process indicators to evaluate the program, and relied on client satisfaction to measure one dimension of outcome.

An individual is qualified to participate in Project AIDS Care if he/she is Medicaid eligible, disabled, has a clinical diagnosis of AIDS, and is likely to require institutionalization in either a hospital or an NF. As a consequence of these stringent eligibility criteria, the population of waiver participants is very sick and expected to die within 2 years. Because the entire population of participants has a short life expectancy, about 10 percent of those initially selected to be interviewed had already died. For these individuals, we substituted participants with similar characteristics (age, gender, race, month of enrollment). We contend that the sample would be significantly biased in terms of its sociodemographic composition if we had failed to substitute another participant for each individual who had already died. Given that this sample is drawn from a very sick population with a terminal disease, most respondents are expected to have poor outlooks on life. The loss of observations from the initial sample selection process cannot be avoided, since all persons are quite sick and will die at some point in the imminent future. This is a potential problem for all researchers who attempt to interview persons with terminal illnesses.

Despite the difficulties encountered in conducting interviews with PWA, the results based on the random sample of 125 clients indicate that most were generally satisfied with their case managers and the availability and type of services received. We found that case managers were well trained and, despite heavy workloads and high mortality rates, the probability of turnover was quite low. The major problem faced by clients and case managers related to difficulties in the process of becoming Medicaid eligible.

It should be noted that the perceptions of clients/surrogates as to the problems associated with enrollment and eligibility processes differ from the shortcomings reported by case managers. One reason for such differences in perceptions is that case managers may be more aware of the inefficiencies of the system, since they work directly with the enrollment and eligibility procedures on behalf of clients. A second explanation is that most case managers were predisposed to obtain access to temporary services for their clients during the time period when the client was awaiting eligibility to disability and other Medicaid services. Examples of other temporary services include local assistance with utilities or Ryan White funds, which can be used to obtain a variety of services for PWA.

Clearly, it would be worthwhile to compare the indicators of client satisfaction for PWA who are being served under a home and community-based waiver with those of PWA in institutional settings (i.e., NFs). Unfortunately, due to patient confidentiality requirements, it is not possible to obtain access to PWA who have been institutionalized. Nonetheless, although our results are only descriptive, they provide useful information for evaluating the success of the program. If a majority of the clients had indicated that they were dissatisfied with several dimensions of the program, this would be an indication that the program was not functioning to meet the needs of its participants. Given that no prior study has examined specific dimensions of quality of a home and community-based waiver program targeted at individuals with AIDS, descriptive analyses of this nature provide useful information regarding this component of the program. In conclusion, it appears that quality of care exists under the Florida Medicaid waiver for PWA in terms of the dimensions we considered.