Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries

The Louis de la Parte Florida Mental Health Institute Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries Huey J. Chen, Ph.D. ARNP Ren Chen, M.S. Shabnam Mehra, M.S. June 2005

This publication was produced by The Louis de la Parte Florida Mental Health Institute University of South Florida 13301 Bruce B. Downs Blvd. Tampa, FL 33612-3807 For more information, call 813-974-7409 or visit the Website: http://fmhi.usf.edu June, 2005 Louis de la Parte Florida Mental Health Institute Publication Agency for Health Care Administration (AHCA) series, 220-63, Tampa, Florida Recommended citation for the report: Chen, H. J., Chen, R., Mehra, S. (2005). Racial and ethnic disparities in health service use and perceived unmet health needs among Florida Medicaid beneficiaries. Tampa FL: Louis de la Parte Florida Mental Health Institute, University of South Florida. This document may be reproduced in whole or part without restriction as long as the Louis de la Parte Florida Mental Health Institute, University of South Florida are credited for the work. Submitted to the Florida Agency for Health Care Administration as a deliverable under contract M0505.

Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries Contents Executive Summary 1 Introduction 5 Background and Significance... 5 An Overview of Florida s Profile... 5 Access and Quality... 6 Specific Aims and Hypotheses... 7 Methods 8 Definitions... 9 Sample, Data Sources and Analytical Approaches... 10 Study Component I: Eligibility Status... 10 Study Component II: Mental/Physical Health Service Utilization Rates... 10 Study Component III: Unmet Physical/Mental Health Service Needs... 11 Results 12 Study Component 1... 12 Medicaid Eligibility Changes Among SSI Beneficiaries... 17 Summary of Component I Results... 21 Study Component II... 21 Mental Health Services... 22 Physical Health Services... 26 Summary of Component II Results... 30 Study Component III... 30 Physical and Mental Health Conditions... 32 Unmet Health/Mental Health Service Needs... 35 Unmet Medication Needs... 37 Barriers Related to Unmet Needs... 39 Satisfaction with Mental Health Services... 40 Satisfaction with Physical Services... 42 Summary of Component III Results... 43 Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries iii

Policy Implications and Recommendations 44 References 45 Appendix A: Change of Race/Ethnicity Distribution in Each AHCA Area 48 Area 1... 48 Area 2... 49 Area 3... 50 Area 4... 51 Area 5... 52 Area 6... 53 Area 7... 54 Area 8... 55 Area 9... 56 Area 10... 57 Area 11... 58 Appendix B: 2002 PSRDC Catcaid Documentation 59 Steps in Mental Health Catcaid Assignment... 59 Step 1 (All Mental Health Catcaids 01.00 20.50)... 59 Step 2 (All Mental Health Catcaids 01.00 20.50)... 60 Step 3 (Catcaids 01.00 03.50)... 60 Step 4 (Catcaids 04.00 18.00)... 60 Step 5 (Catcaids 20.00 20.50)... 60 Steps in Physical Health Catcaid Assignment... 61 Step 1 (All Physical Health Catcaids 51.00 99.99)... 61 Step 2 (Catcaids 63.00 66.00)... 61 Step 3 (Catcaids 51.00 62.00)... 61 Step 4 (Catcaids 98.00 99.00)... 61 CATCAID Table 1... 62 CATCAID Table 2... 64 iv Louis de la Parte Florida Mental Health Institute June 2005

List of Figures Figure 1 Conceptual Framework for Examining Medicaid Health Services Used by Different Racial/Ethnic Groups... 8 Figure 2 Race/Ethnicity Distribution of the Medicaid Population... 12 Figure 3 Medicaid New Enrollment Rates... 13 Figure 4 Medicaid Eligibility Loss Rates... 13 Figure 5 Duration of TANF Eligibility... 14 Figure 6 Eligibility Attrition for TANF Children... 16 Figure 7 Eligibility Attrition for TANF Adults... 16 Figure 8 SSI Eligibility Duration... 18 Figure 9 Attrition of Eligibility for SSI Children... 18 Figure 10 Attrition of Eligibility for SSI Adults... 19 Figure 11 Mental Health Services Used by SSI Adults in Area 1... 22 Figure 12 Mental Health Services Used by SSI Children in Area 1... 23 Figure 13 Mental Health Services Used by SSI Adults in Area 6... 23 Figure 14 Mental Health Services Used by SSI Children in Area 6... 24 Figure 15 Mental Health Services Used by SSI Adults in Areas Other than 1 and 6... 25 Figure 16 Mental Health Services Used by SSI Children in Areas Other than 1 and 6... 25 Figure 17 Physical Health Services Used by SSI Adults in Area 1... 26 Figure 18 Physical Health Services Used by SSI Children in Area 1... 27 Figure 19 Physical Health Services Used by SSI Adults in Area 6... 27 Figure 20 Physical Health Services Used by SSI Children in Area 6... 28 Figure 21 Physical Health Services Used by SSI Adults in Areas Other than 1 and 6... 29 Figure 22 Physical Health Services Used by SSI Children in Areas Other than 1 and 6... 29 Figure 23 Physical Health Condition of Adults... 32 Figure 24 Physical Health Condition of Children... 33 Figure 25 Mental Health Condition* of Adults... 33 Figure 26 Mental Health Condition* of Children... 34 Figure 27 Unmet Mental Health Needs of Adults... 35 Figure 28 Unmet Mental Health Needs of Children... 36 Figure 29 Unmet Physical Health Needs of Adults... 36 Figure 30 Unmet Physical Health Needs of Children... 37 Figure 31 Unmet Medication Needs of Adults... 38 Figure 32 Unmet Medication Needs of Children... 38 Figure 33 Adults Satisfaction with Mental Health Services by Race/ Ethnicity... 41 Figure 34 Child Caregivers Satisfaction with Mental Health Services... 41 Figure 35 Adults Satisfaction with Physical Health Services... 42 Figure 36 Child Caregivers Satisfaction with Physical Health Services... 43 Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries v

List of Tables Table 1 Likelihood of TANF Children Losing Medicaid Eligibility, Relative to White Children... 15 Table 2 Likelihood of TANF Adults Losing Medicaid Eligibility... 15 Table 3 Likelihood of Regaining Medicaid Eligibility by TANF Children... 17 Table 4 Likelihood of Regaining Medicaid Eligibility by TANF Adults... 17 Table 5 Likelihood of SSI Children Losing Medicaid Eligibility... 19 Table 6 Likelihood of SSI Adults Losing Medicaid Eligibility... 20 Table 7 Likelihood of SSI Children Regaining Medicaid Eligibility... 20 Table 8 Likelihood of SSI Adults Regaining Medicaid Eligibility... 20 Table 9 Consistency of Race/Ethnicity between Self-Reported and Administrative Data for Adults... 31 Table 10 Consistency of Race/Ethnicity between Self-Reported and Administrative Data for Children... 31 Table 11 Obstacles Obtaining Mental Health Services Reported by Adult SSI Beneficiaries... 39 Table 12 Obstacles Obtaining Mental Health Services Reported by Caregivers of Child SSI Beneficiaries... 40 vi Louis de la Parte Florida Mental Health Institute June 2005

Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries Executive Summary Both large- and small-scale research findings from the Agency for Healthcare Research and Quality s (2004) National Health Care Disparities Report, the Institute of Medicine s Unequal Treatment study (Smedley, Stith & Nelson, 2002) and a variety of specific research investigations (Mayberry, et al., 1999) have clearly identified and explained the important and un-ignorable issue of health disparity in our society. The Healthy People 2010 report (U.S. Department of Health and Human Services, 2000) and the President s New Freedom Commission on Mental Health call for the elimination of health disparities as one of their overarching goals (New Freedom Commission on Mental Health, 2003). Florida s Medicaid program provided health care related services to more than 10 percent of Floridians in the year 2000 over 2.2 million Medicaid beneficiaries, including low-income children and their families, the aged, blind, and the disabled (Alker & Portelli, 2004; Agency for Health Care Administration, 2004). With over 17 million residents, Florida is growing at a faster rate than the national average, and is experiencing a remarkable change in its minority populations, especially in the population of Hispanic or Latino origin. The comparison of demographic data to the outcomes of access and quality of health care might be expected to illustrate and explain differences in ethnic/racial experiences with Medicaid. The purpose of this study was to examine data pertaining to the pattern of eligibility status change, physical and mental health services use, and perceived unmet physical and mental health needs for potential racial and ethnic differences among Florida Medicaid beneficiaries. Methods The study employed analyses of secondary data, and included three study components: (a) five-year Medicaid eligibility data from July 1, 1999 through June 30, 2004 were examined for eligibility status change among Medicaid Temporary Aid for Needed Families (TANF) and Supplemental Security Income (SSI) beneficiaries across different racial or ethnic groups; (b) Medicaid claims data from fiscal year 2002 2003 were examined for physical health and mental health services used by Medicaid SSI beneficiaries of different racial/ethnic groups; (3) self-reported data collected by the researchers at the Louis de la Parte Florida Mental Health Institute between 1998 and 2004 for evaluation of the managed mental health care waiver programs were examined for unmet health service needs in relation to race/ethnicity. Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries

Key Findings Study Component I Eligibility Change The racial/ethnic distribution of the Medicaid population has changed over the fiveyear study period. The proportion of White beneficiaries decreased from 38.68 percent to 36.94 percent, and Black beneficiaries fell from 32.10 percent to 27.28 percent. On the other hand, in the same period Hispanic Medicaid beneficiaries increased from 18.37 percent to 21.43 percent, Asian beneficiaries increased from 0.45 percent to 0.6 percent, and unclassified Other beneficiaries went from 10.35 percent to 13.7 percent. These findings primarily resulted from greater new Medicaid enrollment rates by minorities other than Black. As a result, the racial/ethnic distribution of the Medicaid beneficiary population became more similar to the Florida census distribution (U.S. Census Bureau, 2004). Black and Hispanic children of both TANF and SSI beneficiaries were less likely to lose their benefits than White children did. Asian child TANF beneficiaries were also less likely to lose their benefit while no differences were found in Asian children in losing SSI benefits compared to White children. No differences were found in TANF or SSI benefit loss when children of Native Americans compared to White children but these findings were limited by the small number of Native American and Asian children in these programs. Adult Black TANF and SSI beneficiaries were less likely to lose Medicaid benefits than White adult Medicaid beneficiaries did. Both adult Hispanic TANF and SSI beneficiaries were more likely to lose their benefits compared to adult White beneficiaries. Asian TANF beneficiaries were also less likely to lose their Medicaid benefits while adult Asian SSI beneficiaries were more likely to lose their benefits than adult White SSI beneficiaries. Adult Native American SSI beneficiaries were also more likely to lose their benefits compared to adult White SSI beneficiaries. When the regaining of Medicaid benefits was examined, it was found that Black and Hispanic children were more likely to regain TANF and SSI benefits than White children. Children of Native American and Asian American were less likely to regain their TANF benefits, but not SSI benefits, compared to White children. Black adults were more likely to regain TANF and SSI benefits than White adult SSI beneficiaries. Hispanic and Asian adult beneficiaries were more likely to regain SSI benefits, but less likely to regain their TANF benefits compared to White adult beneficiaries. Study Component II Service Use Black and Hispanic Medicaid SSI beneficiaries, both children and adults, were less likely to use outpatient mental health services than White Medicaid SSI beneficiaries did. However, Black adult SSI beneficiaries in Area 1 were more likely to use emergency mental health services than White adult beneficiaries. Black SSI beneficiaries, both adults and children, were also less likely to use outpatient physical health services and tend to use more expensive treatment modalities, either emergency or inpatient physical care services equal or higher rates compared to White SSI beneficiaries. These findings appear to be coincided with finding of study Component III where no transportation was the most frequently reported barrier related to access services by caregivers of Black Medicaid child SSI beneficiaries. 2 Louis de la Parte Florida Mental Health Institute June 2005

Hispanic adult SSI beneficiaries were less likely to use outpatient physical health services than Whites. Hispanic adult SSI beneficiaries enrolled in a FFS plan and living in areas other than Area 1 and Area 6 were more likely to use inpatient physical health services compared to White adults. In contrast to mental health services use, Hispanic child SSI beneficiaries used all types of physical health services, including outpatient, emergency and inpatient services, at a similar rate in areas outside Area 6) or a higher rate than White children. These findings may be related to cultural issues which need further investigation. Study Component III Unmet Needs Inconsistency between administrative data and self-reported data was observed for race/ethnicity identification, especially among minorities other than Black. There was only a 33 percent to 39 percent consistency in identifying Asians or Hispanics, and less than 2 percent in identifying Native Americans. As predicted, non-hispanic Black child SSI beneficiaries were more likely to experience unmet physical and mental health needs than non-hispanic White Medicaid child SSI beneficiaries. Non-Hispanic Black adult SSI beneficiaries were also more likely to report having unmet mental health needs, but not unmet physical health needs Compared to non-hispanic White adult SSI beneficiaries. Non-Hispanic adult SSI beneficiaries also experienced more mental and behavioral health problems than White Medicaid SSI beneficiaries. However, there was no significant difference in satisfaction with mental health services between these two groups. Hispanic adults experienced more mental health symptoms than non-hispanic White adults. However, there was no difference in unmet mental health needs or satisfaction with mental health services compared to non-hispanic Whites. They were also less likely to report unmet physical health needs. In contrast, caregivers of Hispanic children indicated having more unmet mental and physical health needs for their SSI children compared to non-hispanic children. Caregivers of Hispanic children also reported difficulty obtaining needed medication for children at the same rate as caregivers of non-hispanic White children, a rate significantly higher than for caregivers of non-hispanic Black children. Policy Implications and Recommendations The Florida Medicaid population has changed demographically over last five years, with both White and Black populations proportionally decreased and Hispanic and Asian populations proportionally increased. Although the array of race/ethnicity variables is limited to simple categories, large discrepancies were found between administrative Medicaid ethnicity data and self-reported race/ethnicity identity, especially for minorities other than non-hispanic Blacks. For example, there is little-to-no consistency between self-reported and administrative data for Native Americans, and less than 40 percent consistency for Asian and Hispanic SSI beneficiaries. Meanwhile, in the category of Others in administrative data, over 50 percent were self-reported to be either Asian or Spanish. These discrepancies distort the true racial/ethnic composition of the Medicaid population. More accurate data collection on regular basis is recommended in order to have an accurate understanding of the diversity of the Medicaid population, which is needed in order to reduce the health disparity for the Florida Medicaid population. Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries

The lack of accurate data to measure and document progress was one of the obstacles to eliminating health disparity identified by McDonough and colleagues (2004). An increasing number of Medicaid beneficiaries are enrolling in some type of managed care plan. But in the present administrative of service use by Medicaid beneficiaries, we were unable to examine the entire population. Due to lack of data from managed care organizations in 9 out of the 11 AHCA areas we had to focus instead only on those beneficiaries who were enrolled in FFS plans in most of the State. The lack of managed care data in the study limits the generalizability of the study findings to Medicaid program as a whole. It is recommended that all health care organizations should report and collect accurate service utilization information, which can be used for quality improvement as well as policy development. Limited assistance to improve the quality of health care professionals was another obstacle to eliminating racial health disparity. Wanted doctor does not take Medicaid was the most frequently identified barrier related to the access of mental health services. Long waiting lists to see healthcare providers was also frequently reported to have been a problem. Why do Medicaid beneficiaries seek services from non-medicaid providers? Is it because the quality of Medicaid providers is in question? Or are there simply not enough providers for the needed services? It is recommended to study utilization disparities at the provider level, to further understand how services have been delivered and whether providers need more training to deliver quality health care for the Medicaid population. Furthermore, it is important to identify culturally competent best practice guidelines for Medicaid providers to reduce the disparities in Medicaid health services. In addition, No transportation was the most frequently identified barriers by caregivers of non-hispanic Black Medicaid child SSI beneficiaries. This may well be related to significantly higher rate of unmet physical/mental health needs of this population. Recommendation on public education about Medicaid services availability, such as Medicaid cab or public transportation may be helpful to caregivers in need of transportation help. Another alternative will be establishing community outreach public health approaches to offer home based care for this population to increase their access to health services. Louis de la Parte Florida Mental Health Institute June 2005

Introduction Background and Significance The purpose of this study is to examine potential racial and ethnic differences existing in Florida Medicaid beneficiaries (a) pattern of eligibility status change, (b) use of physical and mental health services, and (c) self-reported unmet physical and mental health needs. The realization of social interdependencies, cultural rearrangements, and core demographic trends has created the impetus for such an investigation. Both large- and small-scale research findings from the Agency for Healthcare Research and Quality s (2004) National Healthcare Disparities Report, the Institute of Medicine s Unequal Treatment study (Smedley, Stith and Nelson, 2002) and a variety of specific research investigations (Mayberry, et al., 1999) have clearly identified and explained the important and nonignorable issue of health disparity in our society. The Healthy People 2010 report (U.S. Department of Health and Human Services, 2000) and the President s New Freedom Commission on Mental Health call for the elimination of health disparities as one of their overarching goals (New Freedom Commission on Mental Health, 2003). Medicaid has provided financial resources for many uninsured. However, racial/ethnic disparity in service received, outcomes and quality of care continues to be prevalent. In comparing the difference in Medicaid pharmacy use between black and white dually eligible Medicare beneficiaries, Schore and colleagues (2003) found that black beneficiaries have significantly fewer prescriptions filled. Similar findings were observed among Medicaid children as African American youth had a lower rate of treatment with psychopharmacological agents compared to White Medicaid children (Zito, Safer, dosreis, & Riddle, 1998). Among elder Medicaid beneficiaries with depression, non- Hipanic blacks were also less likely to receive treatment compared to White (Strothers III, Rust, Minor, Fresh, Druss, & Satcher, 2005). This study was designed to explore any racial/ethnic disparity exist in Florida Medicaid health care system and barriers related to disparity and enable Florida s Agency for Health Care Administration (AHCA) staff to review current policies, funding streams, management structures, and practices along all levels of the Medicaid system. As Florida s racial and ethnic populations continue to shift and grow over time, changes to the Medicaid system will be needed. An Overview of Florida s Profile Many of the same health care disparities found at the national level exist within Florida (Florida Department of Health, 2003). Some examples include: Non-elderly Hispanics (54 percent) and African-Americans (51 percent) were more likely to lack employer coverage for health insurance than non-elderly Whites (29 percent) during 2000-2001; The overall death rate of African-Americans exceeds that of Whites at the same average age (a 36 percent disparity in 2000); There are considerable disparities across a wide range of disease-specific morbidities, such as cancer, cardiovascular disease, diabetes, and HIV/AIDS. Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries 5

Specific concerns about mental health disparity (Mental Health: A Report of the Surgeon General, 1999): Minorities have less access to, and availability of, mental health services. Minorities are less likely to receive needed mental health services Minorities in treatment often receive less quality of mental health care. Minorities are underrepresented in mental health research. Section 381.7351 of the Florida Statutes, section 381.735 (passed June 8, 2000) initiated the Reducing Racial and Ethnic Health Disparities Closing the Gap grant program, which provides grants to local counties and organizations for community- and neighborhood-based organizations to improve health outcomes for racial and ethnic populations in seven target areas, i.e., cancer, cardiovascular disease, diabetes, adult and child immunizations, HIV/AIDS, maternal and infant mortality and oral health care (Florida Department of Health, 2003). With over 17 million residents, Florida is growing at a faster rate than the national average, and different rates of population increase across ethnic groups is resulting in a changing population composition. For example, persons of Hispanic or Latino origin constitute 16.8 percent of the Florida population, compared to 12.5 percent in the national distribution (U.S. Census Bureau, 2004). The growth of this population is especially notable and it has become the largest minority group in Florida (Florida Office of Economic and Demographic Research, 2003). White Caucasians are 65.4 percent of the Florida population, vs. 69.1 percent nationally, and are declining in their proportion of the population. In 2000, Florida s Medicaid program provided health-care related services to 2.2 million Medicaid beneficiaries more than 10 percent of all Floridians including lowincome children and their families, the aged, blind, and the disabled (Alker & Portelli, 2004; Agency for Health Care Administration, 2004). The comparison of demographic data to outcomes involving access and quality of care can illustrate the reasons behind differences in ethnic/racial experiences with Florida Medicaid. Access and Quality Although access is often considered a component of quality, it is delineated separately here for the purpose of our investigation. There are two types of access that can affect Medicaid beneficiaries: qualified and actual. Qualified access implies services being available on paper, whereas actual access addresses whether those services are obtained. A recent analysis of the differing racial/ethnic perceptions of the health care system found that demographics, sources of care, and patient-physician communication explain most of the differences between population groups, but health care system-wide bias and cultural competence are not fully explained by such factors (Johnson, Saha, Arbelaez, Beach, & Cooper, 2004). Language barriers are found to significantly affect racial and ethnic minorities perceptions of the health system, service utilization and health outcomes (Perez-Stable, Naapoles-Spring, & Miramontes, 1997; Weech-Maldonado, et al., 2003; Weech-Maldonado, et al., 2004). A study examining the racial/ethnic disparities in the utilization of preventive services among Minnesota Health Care Program beneficiaries (e.g., Medicaid and Minnesota Care) found that language and cultural misunderstandings Louis de la Parte Florida Mental Health Institute June 2005

were barriers for African American, Latino, Hmong, and Somali beneficiaries (Minnesota Department of Human Services, 2003). Empirical outcomes research comparing strategies and actions to eliminate these disparities is lacking (i.e., best practices ), and instead there are only broadly defined promising practices, such as cultural competency (Brach & Fraser, 2000), prevention (Collins, Hughs, Doty & et al., 2000), and community and patient feedback (Betancourt, Green & Carrillo, 2002). The reasons for this void have been convincingly argued to include that the barriers to health care of equal quality are not well-described, and that there are problems with study design and methodology (Cooper, Hill & Powe, 2002). Specific Aims and Hypotheses Since the implementation of the Florida Medicaid waiver program in 1996, researchers at the Louis de la Parte Florida Mental Health Institute (FMHI) have conducted several studies to evaluate its effect on the access to and quality of Medicaid services. Data collected in these studies provide the opportunity to investigate access and quality of care disparities among Medicaid beneficiaries of different racial/ethnic backgrounds. The specific aims of this study were to answer the following questions: Are there differences in eligibility status and patterns of change in eligibility status among different racial/ethnic groups? What are the patterns for Supplemental Security Income (SSI) and Temporary Aid for Needed Families (TANF) beneficiaries? Are there disparities between racial/ethnic groups in physical and mental health service utilization rates and service mix among Medicaid SSI beneficiaries? If so, what is the nature of these differences? Are there differences in perceptions of unmet physical and/or mental health service needs among the different racial/ethnic groups? Are there differences in physical health and/or mental health status among different racial/ethnic groups? The first component of this study examined changes in Medicaid beneficiaries eligibility status and eligibility duration for both Medicaid SSI and TANF across different racial/ ethnic groups, to address Research Questions 1 and 2. Due to cultural and language barriers existing in current health care system, we hypothesized that racial/ethnic minorities would be more likely to lose their Medicaid eligibility and to have a longer temporal gap before regaining eligibility. In theory, once individuals become eligible for Medicaid, they are able to access the physical and mental health services covered by the Medicaid plan. This is not always the case. Previous AHCA physical and mental health services evaluations (Shern, Giard, Robinson, & et al., 2002) have revealed that 15 percent to 20 percent of individuals diagnosed with serious mental illness (SMI) did not use any Medicaid-billed mental health services. However, due to broadly defined race/ethnicity classifications, these studies had a limited ability to identify any racial/ethnic disparities that might exist in the Florida Medicaid population. Component II of this study examined differences in service penetration rates among Medicaid SSI beneficiaries of different racial/ethnic groups, using Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries

Medicaid service encounter data from July 1, 2002 to June 30, 2003, to answer Research Question 3. We hypothesized that the minority Medicaid SSI beneficiaries would have lower service penetration rates compared to White Medicaid beneficiaries. Since 1998, researchers in Louis de la Parte Florida Mental Health Institute have collected self-reported data from a stratified sample of state Medicaid beneficiaries. These data, including information related to access to services, physical and mental health status, and satisfaction with services, allowed us to examine whether unmet health needs exist overall among Medicaid beneficiaries, to explore whether various subgroups were more likely to experience unmet health care needs, and to identify possible barriers that keep them from meeting those needs. Component III results provided answers to Research Questions 4 and 5. Researchers have identified important factors related to health disparity such as language, cultural competence, patient-provider communication, and patient-provider concordance that influence the differences in access and outcomes between minority groups and Whites (Perez-Stable, et al., 1997; Minnesota Department of Human Services, 2003; Weech-Maldonado, et al., 2004). Therefore, we hypothesized that minority Medicaid beneficiaries would be more likely to experience unmet health care needs, have lower health functioning levels, and have a lower level of satisfaction with services received compared to White Medicaid beneficiaries. Methods The study employed analyses of secondary data, specifically, Medicaid eligibility data files, Medicaid claims data, and self-reported data collected by the FMHI researchers for the evaluation of managed mental health waiver programs in between 1998 and 2004. The design of this study was based on the framework presented in Figure 1, which defines the populations addressed by each of the four Research Questions. Three components of the data were analyzed to address the specific questions: (1) eligibility and racial/ethnic distribution, (2) service utilization among different racial/ethnic groups, and (3) unmet health/mental health needs. Figure 1 Conceptual Framework for Examining Medicaid Health Services Used by Different Racial/Ethnic Groups Medicaid Enrollees Maintained Medicaid Eligibility Yes Sought Yes Used Yes Services Services No Yes Regained Medicaid Eligibility No No No No Question 1 Are there differences in eligibility status? Question 2 Are there differences in patterns of change in eligibility status between groups? Question 3 Are there differences in service used? Question 4 Are there differences in perceptions of unmet health and/or mental health service needs? Question 5 Are there differences in health and mental health status? Louis de la Parte Florida Mental Health Institute June 2005

Definitions The following definitions are used for the purpose of this study: Maintaining Medicaid Eligibility The begin and end dates in the Medicaid eligibility file was used to identify beneficiaries Medicaid eligibility status during the time period between July 1, 1999 and June 30, 2004. Individuals having a continuous eligibility status without any interruption, or who had an eligibility gap of no more than 30 days between an eligibility end date and the following eligibility begin date, are considered as maintaining Medicaid eligibility. Medicaid Eligibility Gap A duration of greater than 30 days between an end-date of eligibility and the following begin date of Medicaid eligibility is defined as a Medicaid eligibility gap. Unmet Physical/Mental Health Service Need Unmet needs were determined from the self-reported data, and defined as when the beneficiary: Reported needed services but did not seek or receive services, or Reported a problem obtaining services, or Reported being unable to get the type of services needed, or Had a physical/mental health status that was at or above the threshold indicating service need, but did not use services. Race/Ethnicity The race/ethnicity variable was derived from the Medicaid eligibility data file and from self-reported data. Race/ethnicity categories in Medicaid eligibility data included White, Black, Native American, Oriental, Hispanic, and Other. Self-reported race groups included White, Black, Native American, Asian, and Other. Self-reported ethnicities include Hispanic origin or non-hispanic origin. For Study Component I and II, the race/ethnicity data were taken from the Medicaid eligibility data, due to the unavailability of any self-report information. The Component III analysis employed self-reported race/ ethnicity information. Individuals self-reported to be of Hispanic origin were grouped as being of Hispanic ethnicity. The self-reported race group was used to identify individuals as non-hispanic White, non-hispanic Black, Native American, Asian American, and Other. Race/ethnicity data from the Medicaid eligibility files were used when individuals did not themselves provide any race/ethnicity identity. Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries

Sample, Data Sources and Analytical Approaches Study Component I: Eligibility Status Medicaid beneficiaries were included for analysis based on Medicaid eligibility during the period from July 1, 1999 through June 30, 2004. State Medicaid data was obtained and examined for the Medicaid beneficiaries distribution in terms of gender, race/ethnicity, age, program, type of health plan enrolled in, duration of maintaining Medicaid eligibility prior to the first Medicaid eligibility gap, and duration of the first Medicaid eligibility gap before reinstatement of Medicaid eligibility. In the Component 1 section of the study, the identification of sex and race/ethnicity was based solely on Medicaid eligibility data, which includes the classifications White, Black, Native American, Asian, Hispanics, and Others. Age was determined at the time of earliest Medicaid eligibility during the 5-year study period. Beneficiaries who continuously held Medicaid eligibility without a gap of more than 30 days until within 30 days of the end of the study period (i.e., 5/31/2004) were classified as having maintained Medicaid eligibility (regardless of date of eligibility commencement). Beneficiaries who had a gap in eligibility longer than one month, or who lost eligibility before 5/31/2004, and did not regain it before the end of the study period, were classified as having lost Medicaid eligibility. A general descriptive analysis was performed to describe the trends in racial/ethnic distributions over the five-year study period. Kaplan-Meier survival analysis was used to describe the first loss of eligibility status within the study period. A Cox survival model was also used to examine the effect of race/ethnicity on losing eligibility, while controlling for age and gender effects for two subgroups, i.e., individuals who received Supplemental Security Income (SSI) and individuals enrolled in the Temporary Aid for Needed Families (TANF) program. Study Component II: Mental/Physical Health Service Utilization Rates Component II included Medicaid beneficiaries who were eligible for Medicaid benefits without dual eligibility for Medicare, and who received SSI between July 1, 2002 and June 30, 2003. Due to the limited availability of service encounter data from managed care organizations, only Areas 1 and 6 included Medicaid SSI beneficiaries who were in either fee-for-services or managed care plans. For the other AHCA areas, only Medicaid SSI beneficiaries who enrolled in MediPass were included in the Component II analyses. Medicaid service claims and eligibility data files for the fiscal year 2002 were used to examine differences among different racial/ethnic groups in utilization rates for both physical and mental health services in the categories of inpatient care, emergency services, outpatient care, and day treatment services. Data sources included the service encounter data of both statewide Medicaid FFS claims data and Medicaid managed care service encounter data of Area 1 and 6 from July 1, 2002 through June 30, 2003. The service encounter data provided information about the service date, type of service received, and provider. A total of eight different health service categories were identified: mental health outpatient services, mental health day treatment services, mental health emergency services, mental health inpatient services, physical health outpatient services, physical emergency services, physical inpatient services, and substance abuse inpatient services. After recoding of the data, quarterly service utilization rates in 10 Louis de la Parte Florida Mental Health Institute June 2005

each service category were calculated and the general linear model was used to examine differences in service utilization by different racial/ethnic groups. For purposes of the Component II analyses, all service events were recoded into four categories: inpatient/residential services, emergency services, day treatment services, and outpatient services according to Catcaid codes (Appendix B) developed by the Policy and Services Research Data Center (PSRDC) at FMHI. Study Component III: Unmet Physical/Mental Health Service Needs Subjects were included in this study component if they (a) responded to one of the mail surveys of adults and caregivers of Medicaid children from Area 1, 4, 5, 6 and 8 that were conducted between 1998 and 2004 and (b) were eligible for Medicaid at the time they responded to the mail survey. Since we have conducted mail surveys multiple times in Areas 1, 4 and 6, some subjects could have responded more than once and we included all available data in this analysis. Self-reported information taken from the surveys included demographic data (e.g., gender, race/ethnicity, age), health/mental health functioning levels, specific problems or barriers associated with access to services, and level of satisfaction with services. These data were supplemented with Medicaid eligibility data, which was used to determine the subjects eligibility status and plan. Adult beneficiaries physical health status was determined using the SF-12 questionnaire (Brazier, Jones, & Kind, 1993; McHorney, Kosiniski & Ware, 1993; Ware, Bayliss, Rogers, Kosinski, & Tarlov, 1996; Ware, Kosinski, & Keller, 1995, 1996). The mental health status of adult beneficiaries was determined using the Colorado Symptom Index, a 14-item self-reported measure of psychiatric symptoms with a high internal consistency (Cronbach α =.87) (Shern, Lee, & Coen, 1997). Children s physical health status was assessed using a 25-item subset of the Child Health Questionnaire (Landgraf, Abetz, & Ware, 1999; Walters, Salmon, Wake, Wright, & Hesketh, 2001), and the Pediatric Symptom Checklist (Jellinek & Murphy, 1990; Jellinek, Murphy, & Burns, 1986; Jellinek, et al., 1988) was used to screen children for potential psychosocial problems. We conducted simple descriptive analyses to explore the frequency of unmet health care needs and barriers or problems in accessing services, and applied a general linear model with univariate analysis of variance to compare the unmet health needs and health status among different racial/ethnic groups enrolled in different health plan conditions. Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries 11

Results Study Component 1 Study Component I was designed to answer Research Question 1: Are there differences in eligibility status and patterns of change in eligibility status among different racial/ ethnic groups? And Research Question 2: What are the patterns for SSI and for TANF beneficiaries? A total of 4,533,784 beneficiaries of all age groups received Medicaid benefits between July, 1999 and June, 2004. Over this time period, the racial/ethnic composition of the beneficiary population changed (See Appendix A for details of race/ethnic distribution changes in each AHCA area). The relative proportion of both White and Black beneficiaries gradually decreased (from 38.68 percent to 36.94 percent and from 32.10 percent to 27.28 percent, respectively) while Hispanic, Asian and other unclassified beneficiaries gradually increased (from 18.37 percent to 21.43 percent, from 0.45 percent to 0.6 percent, and from 10.35 percent to 13.7 percent, respectively) (Figure 2). Figure 2 Race/Ethnicity Distribution of the Medicaid Population 100% 90% White Black Native American Asian Hispanic Other 10.4% 11.3% 12.4% 13.6% 13.7% 80% 70% 18.4% 19.2% 19.9% 20.6% 21.4% 60% 50% 40% 30% 32.1% 30.7% 29.1% 28.2% 27.3% 20% 38.7% 38.3% 38.0% 36.9% 36.9% 10% 0% 1999 2000 2001 2002 2003 Significant differences in eligibility loss rates (p <.01) and new enrollment rates (p <.01) contributed to the changes in the Medicaid beneficiary s racial/ethnic distribution. New enrollment rates of minority groups other than Blacks were significantly higher than the White s (Figure 3). Though Blacks had a significantly lower rate of new enrollment than Whites, their eligibility loss rate was also significantly lower than the Whites rate. Native American, Asian, and Hispanics had significantly higher eligibility loss rates than Whites (Figure 4). 12 Louis de la Parte Florida Mental Health Institute June 2005

Figure 3: Medicaid New Enrollment Rates White Black Native American Asian Hispanic Other 60% 50% 40% 30% 20% 10% 0% 2000 2001 2002 2003 Figure 4 Medicaid Eligibility Loss Rates White Black Native American Asian Hispanic Other 40% 35% 30% 25% 20% 15% 10% 5% 0% 1999 2000 2001 2002 2003 We further examined the pattern of eligibility status change for TANF and SSI beneficiaries, the two major Medicaid programs that provide health coverage for the most vulnerable population in the state. Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries 13

Eligibility Changes Among TANF Beneficiaries A total of 2,744,776 individuals comprising 34.8 percent White, 30.2 percent Black, 0.06 percent Native American, 0.7 percent Asian American, 27.9 percent Hispanic, and 6.4 percent Other racial/ethnic group - received TANF benefits during the five-year study period. Among all TANF beneficiaries, 79.83 percent were adults, 20.14 percent were children and 0.03 percent were seniors. Significant differences were found in the overall average duration of Medicaid eligibility prior to the first loss of benefits among different racial/ethnic groups in the TANF program. Black TANF beneficiaries had the longest eligibility durations prior to losing Medicaid eligibility (mean = 706 days, SD = 594) of the racial/ethnic groups, followed by the Other group. White, Native American, Asian American and Hispanic Medicaid beneficiaries all had similar eligibility durations (mean durations of 531 days, 526 days, 531 days, and 529 days, respectively) (Figure 5). Figure 5 Duration of TANF Eligibility 800 700 708 Length of Eligibility (days) 600 500 400 300 200 531 526 531 529 466 100 0 White Black Native American Asian Hispanic Others Survival analyses indicated that Black children and adults both had lower attrition rates of eligibility and thus longer durations of eligibility. Black TANF children were 29.6 percent less likely than to lose TANF eligibility compared White TANF beneficiary children, while Black TANF adults were 11.2 percent less likely to lose benefits than Whites. The risks of eligibility loss for TANF as derived from the survival analysis are shown in Table 1 (for TANF children) and Table 2 (for TANF adults). The relative chance of eligibility loss compared to Whites is given by the survival curve exponent, Exp(B). 14 Louis de la Parte Florida Mental Health Institute June 2005

Table 1 Likelihood of TANF Children Losing Medicaid Eligibility, Relative to White Children B S.E. Exp(B) 95% CI p Black -0.350 0.002 0.704 0.701 0.707 <.0001 Native American -0.003 0.034 0.997 0.933 1.065 0.9238 Asian -0.105 0.010 0.900 0.882 0.918 <.0001 Hispanic -0.099 0.002 0.906 0.902 0.910 <.0001 Other 0.130 0.003 1.139 1.131-1.147 <.0001 Table 2 Likelihood of TANF Adults Losing Medicaid Eligibility, Relative to White Adults B S.E. Exp(B) 95% CI p Black -0.112 0.004 0.888 0.881 0.896 <.0001 Native American -0.021 0.060 0.980 0.872 1.101 0.7301 Asian -0.104 0.019 0.902 0.868 0.936 <.0001 Hispanic 0.102 0.005 1.107 1.098 1.117 <.0001 Other 0.077 0.011 1.080 1.057 1.103 <.0001 More than 25 percent of Black TANF children maintained their eligibilities throughout the five-year study period, while fewer than 25 percent of TANF children in the Other group maintained their Medicaid eligibility. Hispanic child TANF beneficiaries had a lower attrition rate for Medicaid eligibility than White TANF children. Adult Hispanic TANF beneficiaries had a higher eligibility attrition rate compared to adult White TANF beneficiaries (Figures 6 & 7). Both adult and child Native American TANF beneficiaries had similar eligibility attrition rates compared to White TANF beneficiaries. Asian American children enrolled in the TANF program had a slightly lower eligibility attrition rate compared to White TANF children, while Asian American adult TANF beneficiaries have similar eligibility attrition Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries 15

rate as adult White TANF beneficiaries (Figures 6 & 7). Figure 6 Eligibility Attrition for TANF Children Figure 7 Eligibility Attrition for TANF Adults 16 Louis de la Parte Florida Mental Health Institute June 2005

Continuation of health care is an important issue for individuals in need of physical and/or mental health care. After loss of Medicaid eligibility, individuals can reapply and reinstate their Medicaid eligibility. Examination of the duration of the gap between the first loss of Medicaid eligibility and its reinstatement found that Blacks, both children and adults, were more likely to regain their Medicaid eligibility (36.6 percent and 41.2 percent respectively) compared to respective White TANF beneficiaries. Hispanic TANF children also had higher rates (approximately 19.1 percent) of regaining Medicaid eligibility compared to White TANF children. However, Hispanic adults, Native American (children and adults), and Asian Americans were less likely than Whites to regain Medicaid eligibilities (Tables 3 & 4). Table 3 Likelihood of Regaining Medicaid Eligibility by TANF Children B S.E. Exp(B) 95% CI p Black 0.312 0.003 1.366 1.358 1.374 <.0001 Native American -0.200 0.056 0.819 0.734 0.913 0.0003 Asian -0.178 0.018 0.837 0.809 0.866 <.0001 Hispanic 0.175 0.004 1.191 1.182 1.200 <.0001 Other -0.052 0.006 0.950 0.939 0.960 <.0001 Table 4 Likelihood of Regaining Medicaid Eligibility by TANF Adults B S.E. Exp(B) 95% CI p Black 0.345 0.006 1.412 1.394 1.430 <.0001 Native American -0.292 0.114 0.747 0.597 0.934 0.0105 Asian -0.332 0.038 0.717 0.666 0.773 <.0001 Hispanic -0.067 0.008 0.936 0.921 0.950 <.0001 Other -0.184 0.021 0.832 0.798 0.866 <.0001 Medicaid Eligibility Changes Among SSI Beneficiaries A total of 411,245 adults and children received SSI benefits during the five year study period comprising 36.0 percent White, 27.7 percent Black, 0.02 percent Native American, 0.25 percent Asian American, 4.98 percent Hispanic, and 30.1 percent Other race/ethnic group between July 1, 1999 and June 30, 2004. Among all SSI beneficiaries, 39.3 percent (161,721) lost their eligibility at least once between 1999 and 2004. Of those who lost eligibility, 37.3 percent (n = 70,101) were White, 17.1 percent (n = 39,904) were Black, 0.1 percent (n = 17) were Native Americans, 2.0 percent (n = 514) were Asian, 20.4 percent (n = 8,100) were Hispanic, and 23.1 percent (n = 43,086) were Other. As was the case for TANF beneficiaries, Black SSI beneficiaries had the longest duration of eligibility (average 1,197 days) before the first eligibility loss during the study period. White SSI beneficiaries had the second longest duration of eligibility (average 924 days). The Native American, Asian American and Other groups had substantially lower eligibility durations (average 739 days, 609 days, and 705 days, respectively) (Figure 8). Racial and Ethnic Disparities in Health Service Use and Perceived Unmet Health Needs Among Florida Medicaid Beneficiaries 17

Figure 8 SSI Eligibility Duration 1,400 1,200 1197 Eligibility Duration (days) 1,000 800 600 400 924 739 609 705 200 0 White Black Native American Asian Hispanic Unlike child TANF beneficiaries, child SSI beneficiaries were more likely to retain Medicaid eligibility. More than 50 percent of SSI beneficiaries maintained their Medicaid eligibility through the five-year study period, as shown by the survival curve in Figure 9. Figure 9 Attrition of Eligibility for SSI Children 18 Louis de la Parte Florida Mental Health Institute June 2005