John A. Fleishman

Effect of Case Management on Unmet Needs and Utilization of Medical Care and Medications among HIV-Infected Persons

Case management has been advocated as a strategy to decrease unmet need for supportive services, improve utilization of ambulatory services, and avoid costly institutional care for HIV-infected persons (17). Previous studies have demonstrated that HIV-infected persons with case managers are more likely to receive benefits advocacy, psychological services, and emotional and practical support (2, 813). However, because many studies used process indicators (for example, receipt of benefits counseling) rather than outcome indicators (for example, receipt of Medicaid) and relied on nonprobability-based samples, the generalizability of their results is limited. Moreover, most studies had cross-sectional designs, which limits the ability to draw causal inferences. Few studies have evaluated the relationship between case management and utilization of medical services by HIV-infected persons, and results from these studies have been somewhat contradictory (4). The impact of case management on medication use is particularly important given that receipt of antiretroviral and prophylactic treatment has been shown to decrease morbidity and mortality among HIV-infected persons (1417). We assessed the effect of case management on unmet need for supportive services and on utilization of medical care and medications in a longitudinal probability sample of HIV-infected adults. Methods Design The HIV Cost and Services Utilization Study is a national multistage probability sample of HIV-infected adults drawn from medical providers in the contiguous United States (18, 19). In the first stage, we randomly selected 28 metropolitan statistical areas and 24 clusters of rural counties. In the second stage, we randomly selected 58 providers known to care for patients with HIV in urban areas and 28 providers in rural areas. We also selected 87 other providers in urban areas and 23 in rural areas who had confirmed in a screening survey of approximately 4000 physicians that they cared for eligible patients with HIV. In the third stage, we randomly selected anonymous patients (using unique codes) from lists of all eligible patients seen at the participating providers during January and February 1996. In the first two stages, we set sampling rates proportional to caseload. In the third stage, we set sampling rates to equalize the probability of selection within subgroups while increasing the overall sampling rate for women and members of private staff-model health maintenance organizations. After replacement of a single urban provider with an equivalent one in the same community, we obtained agreement to participate from 100% of known providers in urban areas and 79% in rural areas, 70% of other providers in urban areas and 83% in rural areas, and 84% of the selected patients. We completed baseline interviews for 2864 of the 4042 (76%) selected patients between January 1996 and April 1997 (20). Follow-up interviews were conducted between December 1996 and July 1997; 2466 persons responded (69% of surviving sampled persons). The median interval between the two interviews was 243 days (range, 36 to 517 days). The study was approved by institutional review boards at RAND and at local institutions. For each respondent, we constructed an analytical weight to adjust the sample to represent the entire reference population. Each weight is the product of the sampling weight, which adjusts for differential sampling probabilities; a multiplicity weight, which adjusts for patients who could have entered the sample through their visits to multiple providers; a nonresponse weight; and an attrition weight (21). Measurements Case Management At baseline, participants were asked whether they had a case manager and, if so, the number of times that they had seen or spoken with the case manager in the previous 6 months. Participants were told that case manager meant a social worker, nurse, AIDS service organization staff member, staff in other service organizations, or anyone else who is assigned to help you get and coordinate care. Participants were considered to have contact with a case manager if they reported seeing or talking to a case manager at least once in the 6 months before the baseline interview. Unmet Need for Supportive Services Unmet need for supportive services was assessed only among persons who had an identified need for a specific service at the baseline interview. We defined need for income assistance as being unemployed and not receiving public income assistance (Social Security Income, Social Security Disability Insurance, or Aid for Families with Dependent Children) other than general welfare at the time of the baseline interview. Participants were considered to have an unmet need for income assistance if they were not receiving any form of public income assistance other than general welfare at the time of the follow-up interview. We defined need for health insurance as having no insurance (Medicaid, Medicare, private insurance, Civilian Health and Medical Program of the Uniformed Services [CHAMPUS], or Veterans Administration benefits) at the time of baseline interview. Unmet need for health insurance was defined as not having a form of insurance at the time of the follow-up interview. Need for housing was defined as being unstably housed (living in shelter; single-room occupancy hotel; doubled up with a friend or relative; hospital; nursing home; residential care facility; or situation other than ones own house, own apartment, or AIDS housing) during the 6 months before the baseline interview or self-reported as homeless or with no usual place to live at the time of the baseline interview. Unmet need for housing was defined as being unstably housed during the interval between the baseline and follow-up interviews. Need for home health care (for example, someone to help with medical or personal care at home) was based on self-reported need at baseline. Unmet need for home health care was based on having no paid or volunteer home health care workers (excluding family members) in the interval between the baseline and follow-up interview. Need for emotional counseling was based on self-reported need in the 6 months before the baseline interview. Unmet need for emotional counseling was defined as not having seen a mental health provider, attended a support group, or seen a spiritual provider (such as a minister) during follow-up. Need for substance abuse services was based on reporting drug dependence during the year before the baseline interview. Drug dependence was defined as use of illicit drugs or prescription drugs known to cause dependence (such as tranquilizers), use of more drugs than intended, or the presence of an emotional or psychological problem associated with drug use during the year before the baseline interview. Unmet need for treatment of substance abuse was defined as not having received professional substance abuse services or participated in a self-help group during follow-up. We defined any unmet need as needing at least one of the six services at the baseline interview and having at least one unmet need at follow-up. Utilization of Medical Care and HIV Medications We constructed three measures of utilization of medical care during the interval between baseline and follow-up visit: an average of at least one ambulatory visit per 3-month period, at least one hospitalization, and at least one emergency department visit not leading to a hospitalization. We selected one ambulatory visit per 3 months because it is the upper limit of an acceptable interval for monitoring progression of disease and response to treatment. Participants who had a follow-up interview within 90 days of the baseline interview (n = 68) were excluded from this analysis because insufficient time had passed to judge the appropriateness of outpatient care. Although some hospitalizations and emergency department visits are inevitable, higher rates of hospitalization and emergency department visits (22) are more likely to occur among persons receiving inadequate outpatient care and those with unstable living situations. Four measures of antiretroviral utilization in the period between the baseline and follow-up interviews were evaluated among persons with a lowest lifetime CD4 cell count less than 0.50 109 cells/L: 1) use of any antiretroviral medication, 2) simultaneous use of at least two antiretroviral medications, 3) simultaneous use of at least three antiretroviral medications, and 4) use of a protease inhibitor or non-nucleoside reverse transcriptase inhibitor. Prophylaxis against Pneumocystis carinii pneumonia or toxoplasmosis in the period between the baseline and follow-up interviews was assessed for participants whose lowest CD4 cell count was less than 0.20 109 cells/L, the accepted threshold for this treatment. Multivariable Models To assess the impact of case management on unmet need for supportive services and utilization of medical care and medications, we adjusted the analysis for covariates that may confound these relationships. The Behavioral Model of Health Services Use was chosen as the conceptual framework for our study because we believed that unmet need and utilization of medical care would be influenced by predisposing factors, enabling resources, and need (23). We included in our analysis predisposing factors (age, sex, ethnicity, HIV risk group, education, geographic region, and drug dependence), enabling resources (insurance status, income, housing status, and living with others), and need (sum of supportive services needed and CD4 lymphocyte count). We categorized HIV risk group as injection drug user if participants had used any injection drug since 1979. Noninjection drug users were categorized as either men who have sex with men or other risk group (includes heterosexual transmission, persons who received infected blood products, and persons with other known or unknown exposures). Drug dependence (yes or no) was categorized

Coping and depressive symptoms among young people with AIDS.

This study examined coping behaviors of people with AIDS, using a large sample (N = 736) that was both geographically and sociodemographically diverse. In-person interviews were conducted with people receiving AIDS-related medical or social services; follow-up interviews were conducted approximately 11 months later. Factor analyses of 16 coping behaviors revealed three factors: Positive Coping, Seeking Social Support, and Avoidance Coping. Respondents with a history of injected drug use, as compared with gay or bisexual men, had higher scores for Avoidance Coping and lower scores for Positive Coping. Each coping scale was significantly related to depressive symptoms in cross-sectional analyses. In longitudinal analyses that controlled for prior depressive symptoms, Positive Coping was significantly related to decreases in symptoms.

Variation in health service use among hiv-infected patients

The effects of sociodemographic factors on health service use among people with human immunodeficiency virus (HIV) infection are assessed. Data are from a survey of 939 clients of the Robert Wood Johnson Foundations AIDS Health Services Program in nine communities across the country. Dependent variables are the number of outpatient visits, use of the emergency room, and whether the respondent had been admitted as an inpatient. In the 3 months before the interview, the sample averaged 7.46 outpatient physician/clinic visits: 35.9% reported an emergency room visit, and 29.9% had been hospitalized. The data suggested differential patterns of health service use, such that those who are white, male, and non-intravenous drug users have higher rates of outpatient clinic/physician use, whereas those who are nonwhite, female, and intravenous drug users have higher rates of emergency room use. Whether these observed differences are attributable to the systems response to different socioeconomic groups, or to differences in individual orientations toward use of medical care is discussed.

Differences in access to zidovudine (AZT) among symptomatic HIV-infected persons

Object:To evaluate socioeconomic factors that determine whether symptomatic HIV-infected persons are offered zidovudine (AZT).Design:Cross-sectional survey conducted as part of the Robert Wood Johnson Foundation’s AIDS Health Services Program.Setting:Public hospital clinics and community-based AIDS organizations in nine American cities.Patients:880 HIV-seropositive outpatients interviewed between October 1988 and May 1989.Main results:Males were more likely to have been offered AZT than were females (adjusted odds ratio 2.99; 95% confidence interval 1.67 to 5.36), those with insurance were more likely to have been offered AZT than were those without (adjusted odds ratio 2.00; 95% confidence interval 1.25 to 3.21), and whites more likely to have been offered AZT than were non-whites (adjusted odds ratio 1.73; 95% confidence interval 1.11 to 2.69). Intravenous drug users were less likely to have been offered AZT than were non-drug users (adjusted odds ratio 0.44; 95% confidence interval 0.28 to 0.69). Persons who had had an episode of Pneumocystis cariniipneumonia were more likely to have been offered AZT than were persons who had AIDS and had not had Pneumocystis cariniipneumonia (adjusted odds ratio 2.95; 95% confidence interval 1.71 to 5.11).Conclusion:The authors conclude that traditionally dis-advantaged groups have less access to AZT, the only antiretroviral agent demonstrated to increase survival of patients who have symptomatic HIV infection.

Rural-urban differences in usual source of care and ambulatory service use: analyses of national data using Urban Influence Codes.

Background. Rural‐urban disparities in access to and utilization of medical care have been a long‐standing focus of concern. Objective. Using the nine‐category Urban Influence Codes, this study examines the relationship between place of residence and having access and utilization of ambulatory health services. Research Design. Data come from the Medical Expenditure Panel Survey, conducted in 1996. Linear and logistic regression analyses assess the relationship between county type and having a usual source of care and ambulatory visits, controlling for demographic and health status measures. Results. Residents of counties that were totally rural were more likely to report having a usual source of care (adjusted OR: 1.98; CI: 1.01, 3.89) than residents of large metropolitan counties. Residents of places without a city of 10,000 or more, but adjacent to a metropolitan area, were also more likely to report having a usual source of care (adjusted OR: 1.92; CI: 1.16, 3.22). In a regression analysis, residents of the most rural places reported fewer visits during the year (B = ‐2.42, CI: ‐3.68, ‐1.32). Conclusions. Results suggest that using rural and urban definitions that go beyond the traditional dichotomy of metropolitan and non‐metropolitan may assist policymakers and researchers in identifying types of places where there is a disparity in access and subsequent utilization of health care. Rural residents, defined as totally rural in the urban influence coding scheme, may report having a health care provider but report fewer visits to health care providers during a year.

Self-Rated Mental Health and Racial/Ethnic Disparities in Mental Health Service Use

Background:Studies of health service use for emotional problems show that the majority of those with disorders do not seek professional help. In addition, mental health service use is lower among members of minority communities, compared with non-Hispanic whites. Objective:To examine the role of self-reported mental health as an indicator of awareness of mental conditions and as an influence in the process of seeking mental health care. Research Design:We conducted cross-sectional analyses of nationally representative data from the Medical Expenditure Panel Survey (MEPS) for 2000–2004. Measures:In-person interviews obtained data on self-rated mental health (SRMH), ambulatory mental health visits, and purchase of prescription medications to treat mental conditions. Respondents completed the SF-12 health status survey; analyses included the SF-12 mental component summary (MCS) as a measure of emotional symptoms. Analyses included only those who provided self-reports of MCS and SRMH. Results:SRMH was related to any ambulatory visit and any medication purchase for mental health treatment, controlling for MCS, and other sociodemographic and clinical variables. The association between SRMH and service use was weaker for black and Hispanic respondents than for whites. In addition, the magnitude of the association between SRMH and MCS was weaker for black and Hispanic respondents than for whites. Conclusions:Racial/ethnic differences in service use may arise in part from different propensities to interpret emotional symptoms as reflecting ones mental health and then to seek professional intervention for emotional problems. SRMH may be useful as an indicator of the extent to which people acknowledge the existence of emotional problems.

Physical and Role Functioning Among Persons With Hiv: Results From a Nationally Representative Survey

Background.Functional limitations of persons living with HIV affect their care needs and ability to perform social roles such as employment. Earlier estimates are outdated, and nationally representative estimates of prevalence and distribution of these limitations have not previously been available. Objectives.The objectives of this study were to characterize physical and role limitations experienced by adults in care for HIV disease in the United States and to analyze variations by demographic and health characteristics in a multivariate framework. Methods.Among 2,836 respondents in the HIV Cost and Services Utilization Survey, we assessed physical functioning with a 9-item scale designed to span a broad continuum of tasks and administered the 2-item ACTG SF-21 role functioning scale. Linearization methods were used to account for the multistage survey design. Results.Limitation in complex roles—working at a job, working around the house, or going to school—was more prevalent than limitation in most specific physical tasks. Among physical tasks, limitation was more prevalent in energy-demanding activities such as climbing stairs (43%) or walking >1 block (26%) than in self-care tasks such as bathing and dressing (14%). Greater limitation was associated with older age, lower educational attainment, more advanced disease, and higher symptom burden. Protease inhibitor treatment was associated with somewhat less physical limitation but no difference in role limitation. Conclusions.Functional status varied widely, suggesting the need for flexible, individualized care system responses. Results identified subgroups whose needs warrant special attention. Symptom intensity, pain, and fatigue were strongly associated with limitation; improved management of these disease manifestations might improve physical and social functioning.

Global self-rated mental health : Associations with other mental health measures and with role functioning

Background:A large body of research shows that global self-rated health is related to important outcome variables. Increasingly, studies also obtain a single global self-rating of mental health, but understanding of what this item measures is limited. Objective:To clarify interpretation of self-reported mental health, we examine its associations with other validated measures of mental health and role functioning. Research Design:We conducted cross-sectional analyses of nationally representative data from the Medical Expenditure Panel Survey. Measures:In-person household interviews obtained data on global self-reported mental health and any limitations in work, school, or housekeeping activities. Adult respondents (N = 11,109) completed the SF-12 health status survey, the K6 scale of nonspecific psychologic distress, and the Patient Health Questionnaire (PHQ-2) depression screener in a self-administered questionnaire. We used the SF-12 Mental Component Summary and the mental health subscale. Analyses examined associations among mental health measures and regressed activity limitations, and the SF-12 physical and emotional role functioning scales on mental health measures, controlling for demographics and selected chronic conditions. Results:The 4 multi-item mental health measures were strongly correlated with each other (r > 0.69), but correlated less strongly with the self-reported mental health item (r ≈ 0.4). In an exploratory factor analysis, self-reported mental health loaded on both mental and physical health factors. In multivariate analyses, each mental health variable was significantly associated with activity limitations and with role functioning, but the association of self-reported mental health with emotional role functioning was relatively weak. Conclusions:Although global self-rated mental health is related to symptoms of psychologic distress, it cannot be considered to be a substitute for them.

Using the SF-12 health status measure to improve predictions of medical expenditures

Background:Relatively few studies have used self-reported health status in models to predict medical expenditures, and many of these have used the SF-36. Objectives:We sought to examine the ability of the briefer SF-12 measure of health status to predict medical expenditures in a nationally representative sample. Methods:We used data from the 2000–2001 panel of the Medical Expenditure Panel Study. Respondents (n = 5542) completed the SF-12 in a questionnaire. Interviews obtained data on demographics and selected chronic conditions. Data on expenditures incurred subsequent to the interview were obtained in part from provider records. We examined different regression model specifications and compared different statistical estimation techniques. Results:Adding the SF-12 to a regression model improved the prediction of subsequent medical expenditures. In a model with only age and gender, adding the SF-12 increased R2 from 0.06 to 0.13. The coefficients for the Physical Component Summary (PCS) and the Mental Component Summary (MCS) of the SF-12 for this model were −0.045 (P < 0.01) and −0.012 (P < 0.01), respectively. In a model including demographic characteristics, chronic conditions, and previous expenditures, adding the SF-12 increased the R2 from 0.26 to 0.29. The coefficients for the PCS and the MCS for this model were −0.025 (P < 0.001) and −0.005 (P = 0.15), respectively. A single general health status question performed almost as well as the full SF-12. Models estimated using ordinary least squares had undesirable properties. In terms of R2, a generalized linear model (GLM) with a Poisson variance function was consistently superior to a GLM with a gamma variance function. Conclusions:Information on self-reported health status is useful in predicting medical expenditures. The extent to which the SF-12 adds predictive power over a comprehensive array of diagnostic data remains to be examined.

Demographic variation in SF-12 scores: True differences or differential item functioning?

Background. Demographic differences have been reported in summary measures of physical and mental health based on the SF‐12 instrument. Objectives. This study examines the extent to which differential item functioning (DIF) contributes to observed subgroup differences in health status. DIF refers to situations in which the psychometric properties of items are not invariant across different groups. The presence of DIF confounds interpretation of subgroup differences. Subjects. A national sample of 11,626 adult respondents in the 2000 Medical Expenditure Panel Survey who completed a self‐administered questionnaire. Measures. In addition to the SF‐12, we collected data on demographic characteristics (age, gender, education, and race/ethnicity) and whether the person had ever been diagnosed with six chronic medical conditions. Results. Multiple‐indicator multiple‐cause latent variable models showed significant differences in physical health by gender, age, and education. Adjusting for DIF reduced but did not eliminate age and education differences. However, for mental health, adjusting for DIF resulted in Black‐White differences becoming nonsignificant, and the effect for the oldest age group switched from positive to negative. Race/ethnicity was not associated with physical health status. Conclusions. Age group comparisons of mental health may be particularly affected by DIF. Differences in education, as well as age and gender, need to be controlled when making group comparisons. Additional work is needed to understand factors that give rise to demographic differences in reported health status.