|
 |
|
|||||||
|
|||||||||
© 2002 American Public Health Association
RESEARCH AND PRACTICE |
Jeanne C. Keruly and Richard D. Moore are with the Johns Hopkins University School of Medicine, Baltimore, Md. Richard Conviser is with the Health Resources and Services Administration, Rockville, Md.
Correspondence: Requests for reprints should be sent to Richard D. Moore, MD, Johns Hopkins University School of Medicine, 1830 E Monument Street, Room 8059, Baltimore, MD 21205 (e-mail: rdmoore{at}jhmi.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Objectives. This study was designed to assess sociodemographic and medical insurer factors associated with receipt of highly active antiretroviral therapy (HAART).
Methods. Patients included (n = 959) were enrolled in the Johns Hopkins HIV Clinic after April 1, 1996, received 
90 days of care, and had a CD4 count 500 cells/mm3 or HIV-1 RNA > 20 000 copies/mL. We assessed the associations of sociodemographic factors and medical insurance with receipt of HAART, stratified by 2 time periods (April 1996 through March 1997 versus April 1997 through March 1999).
Results. HAART was more likely to be used in patients who were > 39 years, White, had CD4 counts < 350 cells/mm3, had fewer missed clinic visits, and did not have intravenous drug use as their risk factor for HIV transmission. In period 1 (April 1996 through March 1997), HAART was more likely to be used in patients who were commercially insured than in other payer groups; differences between payers narrowed in period 2 (April 1997 through March 1999), however, as did differences by race.
Conclusions. Differences in use of HAART on the basis of payer have narrowed since 1996. This encouraging finding may demonstrate the importance of programs that lower economic barriers to medical care.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Demographic and socioeconomic differences have been reported in access to care for persons who have HIV infection and AIDS.1–7 A recent national survey of HIV-infected persons in care, the HIV Care and Services Utilization Study (HCSUS),8 found that medical insurance was a particularly important correlate of receipt of highly active antiretroviral therapy (HAART). Patients with commercial or private insurance were most likely to receive HAART; patients who were insured by a public insurer (e.g., Medicaid or Medicare) were less likely to receive HAART; and patients with no insurance were least likely to receive HAART. The HCSUS found that demographic differences (e.g., racial/ethnic, sex, age) were less important than insurance as predictors of HAART utilization. Notably, the HCSUS analysis was based on data principally from the period during which HAART was being introduced in the United States. Compared with 1996, by late 1997 differences between privately insured and government or uninsured patients had narrowed but were still relatively large.8 We sought to determine whether there has been a continued narrowing of these differences by assessing the relation of utilization of HAART among our patients to their insurance status and other demographic characteristics over time from 1996 through 1999.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
In 1990, a clinical database was established within the Johns Hopkins HIV Clinic that was designed to capture comprehensive longitudinal data on patients attending this clinical practice. A structured, protocol-driven process is used to abstract clinical data at enrollment into the practice and then longitudinally thereafter. This database includes demographic, psychosocial, and clinical information at entry to the practice and any subsequent changes in these variables during follow-up. This process is described elsewhere.9
Patients were included in this analysis if they were enrolled in the Johns Hopkins HIV Clinic from April 1, 1996, through March 31, 1999; had encounters in the system of at least 2 visits and for at least 90 days, with follow-up extending to December 31, 1999; and had a CD4 count of 500 cells/mm3 or fewer or an HIV-1 RNA load of > 20 000 copies/mL or greater at their first encounter in our system. Our goal was to include only patients who were potentially eligible for antiretroviral therapy under prevailing guidelines.10,11
Demographic variables included in this analysis were age, sex, race/ethnicity, HIV risk behavior, level of education, and residence. We established patient residence by using the zip code from the first study period encounter; we categorized residence as city, Maryland county, or outside Maryland. We measured level of education at enrollment into the HIV Clinic. Clinical variables were psychiatric diagnoses, CD4 cell count, HIV-1 RNA viral load, and prescribed antiretroviral therapies. We defined HAART as receipt of 3 or more drugs, with at least 1 of the drugs being a protease inhibitor or a non-nucleoside reverse transcriptase inhibitor.
We assessed payer status through the institution's visit registration database, which captures the method of payment for each completed visit. For each completed encounter with the patient, an insurance field indicates the verified insurance status for the completed visit. For example, eligibility status for receipt of Medical Assistance is verified before registration is complete. Uninsured patients were defined as those who present for care and have no insurance or have lost their Medical Assistance eligibility. A review of our Ryan White–supported activity accounting database, which was initiated in 1997, indicates that more than 90% of the visits of our uninsured patients qualified for a reduction in fee schedules and were supported exclusively by Ryan White CARE Act funds. Data in the payer field were reformatted into an insurance field that included the following values: commercial/private insurance, Medicaid-funded (including Medical Assistance only, Medicaid, and Medicare dually insured), and uninsured. We did not analyze patients with Medicare only. We then constructed a summary payer variable for each patient by examining the payer status for all completed visits and assigning the following values: 80% commercial/private, 80% Medicaid funded, 80% uninsured, 20% to 80% partially insured. Approximately 12% of our patients were enrolled in clinical trials in which they received HAART therapy. For the analysis presented in this article, we excluded these patients to assess other factors associated with the use of HAART without confounding by clinical trial enrollment.
Scheduled visits that are missed or canceled also are contained within the database. We calculated a missed visit ratio by dividing the number of missed (not canceled) appointments by the total number of scheduled appointments.
For our analyses, we assessed the association of each of the sociodemographic and clinical patient characteristics with receipt of HAART. We used the 
2 test for categorical variables and the t test for continuous variables and the Wilcoxon rank sum test for variables that were not normally distributed. We used Cox proportional hazards regression to assess multivariate associations between the sociodemographic and clinical variables and the use of HAART. For the Cox regression analyses, we stratified patients by enrollment date and follow-up time. The first regression analyzed patients who were enrolled into our clinic from April 1, 1996, through March 31, 1997, and followed until December 31, 1997. The second regression analyzed patients who were enrolled into our clinic from April 1, 1997, through March 31, 1999, and followed through December 31, 1999. The third analysis included all patients. In this analysis, we created a time-dependent variable in the model specifying early enrollment (April 1, 1996, through March 31, 1997) versus late enrollment (April 1, 1997, through March 31, 1999). The third analysis included interactions between calendar period and insurance status, which we also modeled as timedependent variables.
Finally, in patients who received HAART, we assessed achievement of an undetectable HIV-1 RNA load (defined as < 400 copies/mL on HIV-PCR) within 6 months of starting therapy. We used logistic regression to assess the association of patient sociodemographic and clinical characteristics with achievement of an undetectable HIV-1 RNA load.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
We analyzed data from 959 HIV-infected patients. Major demographic and clinical characteristics of the sample are shown in Table 1
. The sample population is predominantly male (70%), African American (78%), and has intravenous drug use as the major risk behavior for HIV transmission (47%). HAART was received by 66% of the patients. The median CD4 count was 243 cells/mm3; the median HIV-1 RNA load was 36 800 copies/mL. There were 163 (17%) patients who were 80% commercially/privately insured, 278 (28%) who were 80% government insured, 247 (26%) who were 80% uninsured, and 271 (29%) who were 20%–80% partially insured. Of the 271 partially insured patients, 16 had 50%–79% commercial/private insurance, 150 had 50%–79% Medicaid, and the remaining 95 were 50%–79% uninsured. The median (range) duration of time in days in each insurance category was 540 (109 to 1138) for 80% commercial/private; 611 (91 to 1149) for 80% Medicaid; 392 (91 to 1151) for 80% uninsured; and 581 (94 to 1148) for 20%–80% partially insured.
|
Univariate associations of the sociodemographic and clinical patient characteristics with HAART use are shown in Table 2
. HAART was significantly less likely (P < .05) to be received by women, non-White patients, injection drug users, patients with less than a high school education, patients from Baltimore City, patients who do not have a psychiatric diagnosis, patients with a higher CD4 count, and patients with a lower HIV-1 RNA load. Commercially/privately insured patients were more likely than Medicaid partially insured patients or uninsured patients to receive HAART. Medicaid and partially insured patients were more likely than uninsured patients to receive HAART. Missed visits were more common among patients who did not receive HAART.
|
Results of the multivariate Cox regression analyses of receiving HAART are shown in Table 3
. The first of these analyses represents 462 patients enrolled from April 1, 1996, through March 31, 1997, and followed until December 31, 1997. The second analysis represents 497 patients enrolled from April 1, 1997, through March 31, 1999, and followed until December 31, 1999. In both analyses, patient characteristics that were independently associated with a greater likelihood of receiving HAART included having a CD4 count of 
200 cells/mm3, not being an injection drug user, being older than 39 years, and having had 25% or fewer missed visits. In the earlier-enrolled patients, HAART receipt also was associated significantly with commercial/private insurance and White race. In the later-enrolled patients, however, insurance and race were not significant except for a significant difference for 20%–80% partially uninsured (relative hazard [RH] = 0.76; 95% confidence interval [CI] 0.59, 0.97). In the combined analysis of all 959 patients, patients with any category of noncommercial insurance were significantly less likely to receive HAART than were patients who were insured commercially/privately. The interactions between insurance category and time period demonstrate, however, that in the late time period after January 1, 1998, there was a significant increase in the use of HAART for Medicaid (RH = 1.83; 95% CI 1.08, 3.12) and uninsured patients (RH = 1.72; 95% CI 1.00, 2.97) compared with commercially/privately insured patients. Those who were partially insured had a nonsignificant increase in HAART use (RH = 1.61; 95% CI 0.94, 2.76).
|
The percentages of patients who received HAART from 1996 through 1999, stratified by payer, are plotted in Figure 1
. Consistent with the results of the foregoing multivariate analyses, there was a narrowing of the difference in HAART use by payer category over time.
|
Finally, we analyzed achievement of an undetectable HIV-1 RNA load for patients who received HAART (n = 632; undetectable = 328). This analysis is shown in Table 4
. Patients who had a baseline CD4 count of less than 50 cells/mm3, a HIV-1 RNA load of greater than 10 000 copies/mL, or 25% or fewer missed visits were less likely to have undetectable HIV-1 RNA. Insurance, age, race, sex, psychiatric diagnosis, and HIV risk factor were not significant variables. Time of enrollment (April 1, 1996, through March 31, 1997, versus after April 1, 1997) also was not a significant variable.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
HIV-infected patients who were uninsured or were insured by Medicaid were less likely than commercially/privately insured persons to receive HAART before December 1997, but these differences in receipt of HAART diminished after that date. Although patients who were partially insured (almost exclusively a combination of uninsured and government insured) remained somewhat less likely to receive HAART in the later time period, differences also decreased between patients in this payer category and the commercially/privately insured patients. These data from our practice suggest that early differences in HAART use have narrowed over time. The disparities in receipt of HAART by insurer in 1996 and early 1997 in HCSUS and among our patients, which still were apparent in HCSUS even by the end of 1997, were no longer present among our patients in 1998–1999.
In our multivariate analysis, the association of commercial insurance with early HAART use was independent of other sociodemographic and clinical factors. Commercial insurance, therefore, may be an important socioeconomic and behavioral marker that is not otherwise captured by any of our other variables. When HAART first became available, HIV care providers may have preferentially used this therapy in commercially insured patients because of their belief that employment (the source of insurance in almost all of these patients) connoted a stability and self-discipline that was important for the most effective use of HAART. As providers' experience with HAART has increased, its use also has increased in all patients. The fact that use of HAART could increase in other payer groups emphasizes, however, the importance of access to these expensive drugs. It is notable that in Maryland, the Medicaid program has never restricted access to any antiretroviral drugs. They were all available to Medicaid-insured patients from 1996 through 1999. Similarly, in our uninsured patients, antiretroviral drugs were available shortly after licensing through the Ryan White CARE Act–funded Maryland AIDS Drug Assistance Program. More than 90% of visits by our uninsured patients qualified for reduction in fee schedules and were supported exclusively by Ryan White CARE Act funds from Titles I, II, and III. In addition to medication, the funding also provides access to primary and subspecialty care, laboratory monitoring of therapy, and other associated care that is critical in the use of these drugs. The Ryan White CARE Act, which has been in existence since 1990, provided $3 billion in medical and ancillary care assistance throughout the United States through its 4 titled programs from 1996 through 1999.12 Although neither the Maryland Medicaid program nor the Ryan White CARE Act funding can explain the narrowing over time in HAART use differences by payer group, they provided the means to support patients who otherwise would have been excluded from this care.
Our analysis also demonstrated some demographic differences in the use of HAART in 1996 and early 1997. Among early enrollees, White patients were more likely to receive HAART than non-White patients, but this difference narrowed later. Injection drug use as an HIV transmission risk factor was associated with decreased use of HAART in the early cohort, and although it was less strongly associated with use of HAART in the later cohort, it continued to be a factor associated with less HAART use. This association may reflect a perceived difficulty for substance users to maintain the rigorous schedule required to use HAART and follow up with visits to monitor HAART use. Nevertheless, there may be inappropriate underutilization of HAART in substance users.
Although mental health diagnoses have been associated with decreased access to care for other illnesses,13 a psychiatric diagnosis did not serve as a barrier to HAART use in our patient population, and neither did patient age. Missing scheduled visits, however, was strongly associated with decreased HAART use. Lack of HAART receipt may have been appropriate management for patients with a high proportion of missed visits. This patient population, however, also may be a target for intervention to improve visit adherence. Our analysis of achievement of an undetectable viral load also showed that missing visits was most strongly associated with not achieving an undetectable viral load, even after control for injection drug use as an HIV risk factor. This finding may be supportive of our providers' decisions not to prescribe HAART as frequently to these patients. These patients may be least likely to respond to HAART, possibly as a result of poorer adherence, because there was no difference in prior use of antiretroviral drugs that could have led to a higher prevalence of drug-resistant HIV. Notably, no other patient characteristics (except low baseline HIV-1 RNA) were associated with achievement of an undetectable viral load.
In summary, we found that medical insurance was strongly associated with receipt of HAART soon after its introduction into clinical practice. These early differences had narrowed significantly by 1999, however, as had other demographic differences in HAART receipt. This encouraging finding demonstrates that effective, yet expensive, medical care can become available to all patients if socioeconomic barriers are removed.
|
Acknowledgments |
|---|
This work was supported by the Health Resources and Services Administration (98-HAB-A260935), the National Institute on Drug Abuse (RO1-DA-11602), the Agency for Health Care Policy and Research (RO1-HS-07809), and the Food and Drug Administration (FD-U-000977).
|
Footnotes |
|---|
J . C. Keruly and R. D. Moore planned the study. J. C. Keruly conducted the analysis and wrote the paper. R. D. Moore and R. Conviser contributed to the analysis and the writing of the paper.
Accepted for publication February 6, 2001.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
1. Stein MD, Piette J, Mor V, et al. Differences in access to zidovudine (AZT) among symptomatic HIV-infected persons. J Gen Intern Med. 1991;6:35–40.[Medline]
2. Bastian L, Bennett CL, Adams J, Waskin H, Divine G, Edlin BR. Differences between men and women with HIV-related Pneumocystis carinii pneumonia: experience from 3,070 cases in New York City in 1987. J Acquir Immune Defic Syndr. 1993;6:617–623.
3.
Moore RD, Stanton D, Gopalan R, Chaisson RE. Racial differences in drug therapy for HIV disease in an urban community. N Engl J Med. 1994;330:763–768.
4. Davidson AJ, Bertram SL, Lezotte DC, et al. Comparison of health status, socioeconomic characteristics, and knowledge and use of HIV-related resources between HIV-infected women and men. Med Care. 1998;36:1676–1684.[Medline]
5. Kupek E, Dooley M, Whitaker L, Petrou S, Renton A. Demographic and socioeconomic determinants of community and hospital services costs for people with HIV/AIDS in London. Soc Sci Med. 1999;48:1433–1440.
6. Palacio H, Shiboski CH, Yellin EH, Hessol NA, Greenblatt RM. Access to and utilization of primary care services among HIV-infected women. J Acquir Immune Defic Syndr. 1999;21:293–300.
7. Sorvillo F, Kerndt P, Odem S, Castillon M, Carruth A, Contreras R. Use of protease inhibitors among persons with AIDS in Los Angeles County. AIDS Care. 1999;11:147–155.[Medline]
8.
Shapiro MF, Morton SC, McCaffrey DF, et al. Variations in the care of HIV-infected adults in the United States. JAMA. 1999;281:2305–2315.
9. Moore RD. Understanding the clinical and economic outcomes of HIV therapy: the Johns Hopkins HIV Clinical Practice Cohort. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;17(suppl 1):S38–S41.
10.
Carpenter CCJ, Fischl MA, Hammer SM, et al. Antiretroviral therapy for HIV infection in 1998. JAMA. 1998;277:1962–1969.
11. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. US Department of Health and Human Services and Henry J. Kaiser Family Foundation. MMWR Morb Mortal Wkly Rep. 1998;47(RR-5):43–82.[Medline]
12. HRSA Fact Sheet. Rockville, Md: HIV/AIDS Bureau, Health Resources and Services Administration; March 2000.
13.
Druss BG, Rosenheck RA. Mental disorders and access to medical care in the United States. Am J Psychiatry. 1998;155:1775–1777.
This article has been cited by other articles:
|
|
 |
|
  K. Clements-Nolle, R. Marx, M. Pendo, E. Loughran, M. Estes, and M. Katz Highly Active Antiretroviral Therapy Use and HIV Transmission Risk Behaviors Among Individuals Who Are HIV Infected and Were Recently Released From Jail Am J Public Health, April 1, 2008; 98(4): 661 - 666. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.T. Fang, Y.Y. Chang, H.M. Hsu, S.J. Twu, K.T. Chen, C.C. Lin, L.Y.L. Huang, M.Y. Chen, J.S. Hwang, J.D. Wang, et al. Life expectancy of patients with newly-diagnosed HIV infection in the era of highly active antiretroviral therapy QJM, February 1, 2007; 100(2): 97 - 105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Fitzpatrick, N. R. Powe, L. S. Cooper, D. G. Ives, and J. A. Robbins Barriers to Health Care Access Among the Elderly and Who Perceives Them Am J Public Health, October 1, 2004; 94(10): 1788 - 1794. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |