| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Division of General Internal Medicine, the Center for Research in Medical Education and Health Care, the Biostatistics Section, Division of Clinical Pharmacology, Jefferson Medical College of Thomas Jefferson University, Philadelphia, Pennsylvania; and New York State Department of Health, Albany, New York.
Address reprint requests to: Barbara J. Turner, MD College Building Room 132 1025 Walnut Street Philadelphia, PA 19107-5083 E-mail: barbara.turner{at}mail.tju.edu
 |
Abstract |
|---|
Top
 Abstract MethodsResultsDiscussionReferences |
|---|
Methods: We retrospectively studied a cohort of 2714 HIV-infected women in New York State who delivered live infants from 1993–96. Among 682 women prescribed antiretroviral therapy in the last two trimesters, we studied 549 who started therapy more than 2 months before delivery. Adherence was defined as adequate if the supplied drug covered at least 80% of the days from the first prescription in the last two trimesters until delivery. Multivariable analyses were used to examine associations between maternal and health care factors and adherence.
Results: Only 34.2% of 549 subjects had at least 80% adherence based on pharmacy data, a rate that remained stable over time. The adjusted odds ratios (ORs) of adherence for black (OR 0.47, 95% confidence interval [CI] 0.30, 0.75) and Hispanic (OR 0.49, 95% CI 0.29, 0.82) women were nearly 50% lower than for white women. The OR of adherence was 0.32 (95% CI 0.12, 0.90) for teenagers compared with women aged 25–29 years and 0.56 (95% CI 0.34, 0.92) for women in New York City versus those residing elsewhere. Women on antiretroviral therapy before pregnancy were more likely to adhere (OR 1.55, 95% CI 1.02, 2.35).
Conclusion: Teenagers, women of minority groups, and women living in New York City had greater risks of poor antiretroviral adherence, whereas women already prescribed antiretrovirals before pregnancy had better adherence. Our conservative pharmacy data–based measure showed that most HIV-infected women adhered poorly and adherence did not improve over the 4-year study.
Since the Pediatric AIDS Clinical Trial Group Protocol 076 reported a significant decrease in vertical transmission of human immunodeficiency virus (HIV) with zidovudine given during pregnancy, intrapartum, and to infants,1 antiretroviral therapy for infected pregnant women has become standard care.2 Prescriptions of antiretroviral therapy for HIV-infected pregnant women have increased,3–6 but prescriptions do not ensure adherence. Poor antiretroviral adherence puts mothers and children at risk, but little is known about antiretroviral adherence in pregnant women. We conducted a population-based examination of adherence in pregnant women.
|
Methods |
|---|
|
Top
AbstractMethodsResultsDiscussionReferences |
|---|
We linked deliveries with vital statistics data on education, residence, racial-ethnic group, parity, self-reported substance abuse during pregnancy, birth weight, and gestational age at delivery from doctors’ estimates or, rarely, mothers’ self-reports of last menstrual cycles. Each woman was assigned an identification number, and other identifying information was removed from the file.
We identified 3310 deliveries of live infants during the study and retained only the most recent deliveries of each woman (n = 2921). We matched vital statistics records for 2796 (95.7%) deliveries. We excluded the few Asian or Native American women (n = 36) or those who lacked ethnicity data (n = 16) or maternal country of birth (n = 32), leaving 2714 women. Institutional review boards at Jefferson Medical College, the New York State Department of Health, and the Medicaid Confidential Data Review Committee approved the study. This same base population of 3310 deliveries was used to select women eligible for separate analyses by our group that addressed postpartum adherence (n = 2648) (unpublished data) and changes in antiretroviral prescription after publication of Pediatric AIDS Clinical Trial Group Protocol 076 (n = 2607).3 Each of the three analyses had different eligibility requirements, given the different periods of focus. However, the total eligible group for this analysis (n = 2714) included all women in the other two studies plus approximately 50–110 women who were excluded from other analyses because of such factors as limited duration of postpartum eligibility or multiple deliveries.
Outcome Measures
Claims files provided diagnoses, payment rates, and filled prescriptions. From pharmacy claims, we identified all Food and Drug Administration–approved nucleoside reverse transcriptase, protease, and non–nucleoside reverse transcriptase inhibitors. At least one paid claim for an approved antiretroviral drug during the study interval signified receipt of therapy. We restricted our analysis to the last two trimesters. We initially categorized women as nonusers, users, and late starters (ie, treatment only within 2 months of delivery). We excluded nonusers and late starters from analyses of adherence because at least 2 months of pharmacy claims were necessary to identify gaps in adherence.9 Our approach follows that of other studies that used pharmacy data to estimate medication adherence.10
Similar to other studies,11 adequate adherence was defined as sufficient antiretroviral drug, based on each woman’s filled pharmacy claims, to cover at least 80% of days from the first prescription until delivery. Our analytic decisions led to a generous estimate of adherence. We determined the quantity of a specific type and strength of medication prescribed and determined the number of days that amount covered, assuming the lowest recommended daily dose. We then examined the days supplied on the claim. When those numbers differed, we used the estimate that suggested better adherence. When a woman received antiretroviral therapy within 2 weeks before a hospitalization, the hospital days were counted as adherent. When a woman switched drugs, she received credit for carry-over days covered by the earlier drug. However, multiple claims for the same amount of one drug on the same day were only counted as one filled prescription.
Demographic and Clinical Data
Data on subject age, ethnicity, education, parity, and delivery year came from claims and vital statistics files. Data on viral load or CD4 counts were unavailable. We measured severity of HIV disease from diagnoses on claims.12 Chronic medical (eg, diabetes) or psychiatric (eg, depression) conditions also were identified from claims.
Illicit drug use was identified by a tested approach8 applied to claims or patient reports from vital statistics and categorized as methadone treatment in pregnancy; illicit drug use in pregnancy, as indicated by claims diagnoses or vital statistics; illicit drug use outside pregnancy based on diagnoses on claims or methadone treatment claims; and no evidence of illicit drug use. Women who qualified for more than one of the listed categories were classified into the first applicable one. Smoking or drinking during pregnancy was determined from vital statistics.
We created indicators for receipt of HIV-focused services when women visited providers paid at enhanced rates by New York State to deliver HIV-specific ambulatory care.13 We also identified any care from providers in an HIV-related specialty (ie, infectious disease, allergy-immunology, hematology-oncology, and HIV specialty designation). Nearly all (341 of 346) subjects who received HIV-focused services also received care from an HIV specialty provider. Thus, we created a three-part variable for analysis: HIV-focused services with or without HIV-specialty care, HIV specialty care only, and neither HIV-focused services nor specialty care. We also determined whether women were treated at sites participating in trials of antiretroviral drugs in pregnancy.
To assess prenatal care, we applied the Kotelchuck Adequacy of Prenatal Care Utilization14 to claims from obstetrician-gynecologists, general internists, family practitioners, or HIV specialty providers as reported (Berlin M, Cocroft J, Newschaffer CJ, Cosler L, Turner BJ. Adequacy of prenatal care: Vital statistics vs. medical claims data [abstract]. American Public Health Association Meeting, Washington, DC, November 1998) and created a three-part measure: inadequate, intermediate-adequate, and adequate plus. From pharmacy claims, we created an indicator for women who started antiretroviral therapy before pregnancy. We also created a categoric variable for the duration of antiretroviral therapy from first-filled prescription during pregnancy until delivery, grouped by quartile.
Analysis
We estimated proportions of women with any antiretroviral exposure during the last two trimesters stratified by calendar month of delivery and plotted the proportion of women with any exposure by 3-month groups, smoothing the plot line using the cubic spline method.15 We examined the association of maternal and health care characteristics with adherence among women with prescribed antiretroviral therapy in the last two trimesters, but before the last 2 months before delivery. Bivariate associations with adherence were examined using the 
2 test. For age, 5-year categories were analyzed with the exception of two larger age groups for women younger than 20 and at least 35 years old. We also combined 1993 and 1994 for analysis because of a small sample for 1993. Multivariable logistic regression models were estimated for adherence. We estimated a series of models to reduce the number of variables considered. We entered all demographic and clinical variables listed in Table 1
in a backward selection model using P < .20 for retention in the model, then added all health care delivery variables in Table 2 to the variables retained in the first model.
|
|
|
Results |
|---|
|
Top
AbstractMethodsResultsDiscussionReferences |
|---|
shows that the proportion of women adherent to antiretroviral therapy did not change over the study period. Tables 1
and 2 summarize the unadjusted associations between adherence to antiretroviral therapy and maternal and health care factors. Adherence was lowest among women younger than 20 years of age (n = 30) compared with older age groups. Ethnicity was related significantly to adherence, with the proportions of black and Hispanic women who adequately were adherent approximately 20% lower than the propor tions of white women who were adequately adherent. Other maternal characteristics associated significantly with lack of adherence were other chronic diseases such as asthma, illicit drug use, or methadone treatment during pregnancy, and living in New York City. Multiparity and smoking or drinking during pregnancy had borderline associations with inadequate adherence. No health care factors were statistically significantly related to adherence, but higher proportions of women were adherent if they received antiretroviral therapy before pregnancy, HIV-focused services, or HIV specialty care.
|
shows the results of multivariable analysis. Significant associations persisted between poor adherence and minority ethnicity, teen age, and New York City residence. Adherence also was lower for women with other chronic diseases (odds ratio [OR] 0.67, 95% confidence interval [CI] 0.44, 1.0). Women treated with antiretroviral therapy before pregnancy had greater than 50% higher ORs of adherence than women first prescribed antiretrovirals during pregnancy. Other associations did not quite achieve statistical significance. Somewhat lower ORs of adherence were seen in multiparous women (OR 0.63, CI 0.37, 1.08 for at least three versus no prior births) and for pregnant women who used illicit drugs or methadone compared with women without evidence of illicit drug use. Treatment at clinical trial sites was not associated significantly with adherence, and prenatal care did not show a consistent association. Women who received HIV-focused services or had care from a provider in an HIV-related specialty had 40–60% higher ORs of adherence, but lower confidence limits crossed one.
|
|
Discussion |
|---|
|
Top
AbstractMethodsResultsDiscussionReferences |
|---|
Identifying factors likely to put women at risk for poor adherence to antiretroviral therapy could help target adherence interventions. Our adjusted analyses showed that relatively few factors were associated strongly with inadequate adherence. Women who did not have those characteristics still had low adherence rates, suggesting global problems with adherence. Previous research also showed the difficulty of predicting compliance with HIV therapy.17–20
Demographic factors associated with poor adherence were minority ethnicity, teen age, and New York City residence. Others described relationships between minority ethnicity and less use of antiretroviral therapy.21–23 Blacks and Hispanics had rates of adherence that were nearly 50% lower than those of whites in our current study. Hispanics had more than twofold greater ORs of antiretroviral treatment in pregnancy than whites in previous analyses.3 Perhaps cultural or language barriers caused Hispanics to be reluctant to disagree with providers when offered therapy, but they showed their aversion by poor adherence. Residence in New York City was associated with worse adherence. Interviews of 71 HIV-infected women in New York City revealed frequent negative attitudes toward zidovudine,24 which might be more prevalent in New York City than other areas of the state. Older women might have better knowledge about HIV disease and its management or might be more organized and tolerant of strict therapeutic regimens. Young persons, especially teenagers, probably had never taken a medication regularly and could have been overwhelmed by the task.
Women who had been prescribed antiretroviral therapy before pregnancy showed better adherence than women who first took it during pregnancy. Possibly those women had greater experience with taking medications, had developed reminder systems to help with adherence, had more advanced disease, or were most convinced of the benefits of therapy. The presence of a chronic disease other than HIV infection showed lower ORs of adherence. Those women might have been overwhelmed by or intolerant of taking medications for HIV and their other chronic conditions.
Research suggested that factors such as illicit drug use might be related to poor adherence to HIV-related therapies,25–27 but illicit drug users adhere to therapy when aided by social supports.28,29 Odds ratios of adherence were lower for women with active drug use or methadone treatment in our study, but they were not statistically significant.
Characteristics of health care settings have been shown to influence receipt of therapies for HIV.3 Specialists in HIV care appear to be quicker to adopt advances in antiretroviral therapy than other physicians.30 We found that ORs of adherence were approximately 40–60% greater for persons who received care from providers in HIV-related specialties such as infectious diseases or who were paid to deliver enhanced HIV services. However, that association did not achieve statistical significance, nor did the association between adequacy of prenatal care and adherence.
Our pharmacy-based measure of adherence allows population-based analyses and avoids the Hawthorne effect that might influence other measures of adherence. Pharmacy data–based methods have been described offering overly optimistic estimates of adherence.31 Our measure could be viewed as underestimating adherence because it did not consider treatment of women who received drugs through clinical trials or free drug programs such as the AIDS Drug Assistance Program. However, such women would not have been in our study because inclusion required at least one Medicaid claim for antiretroviral therapy, and women in trials or free drug programs would not need Medicaid to pay for drugs. We could not distinguish drug holidays from a woman failing to take doses of her medication, but both types of poor adherence have potentially detrimental effects on suppression of HIV replication.16,32 Although this study did not examine the relationship of adherence and clinical outcomes, Woodward et al used a similar pharmacy data–based approach to monitor antiretroviral adherence and showed an association with suppression of HIV viral load (Woodward J, Wareham PS, Grohskopf L, Madigan D, Hooton TM. Protease inhibitors and HIV-1 RNA response [abstract]. 12th World AIDS Conference, Geneva, Switzerland, June 1998).
Our results are worrisome in that relatively few HIV-infected pregnant women had adequate antiretroviral adherence. Minority ethnicity, young age, and large urban residence might indicate women at particular risk for poor adherence, but adherence was problematic in all groups examined. Research should explore further which factors predict adherence.
|
Footnotes |
|---|
The authors’ opinions do not necessarily reflect those of the New York State Department of Health.
Received April 28, 1999. Received in revised form July 30, 1999. Accepted August 12, 1999.
|
References |
|---|
|
Top
AbstractMethodsResultsDiscussionReferences |
|---|
2. Anonymous. Centers for Disease Control and Prevention. Public Health Service Task Force recommendations for the use of antiretroviral drugs in pregnant women infected with HIV-1 for maternal health and for reducing perinatal HIV-1 transmission in the United States. MMWR Morb Mortal Wkly Rep 1998;47(RR-2):1–30.[Medline]
3. Turner BJ, Newschaffer CJ, Zhang D, Fanning T, Hauck WW. Translating clinical trial results into practice: Predictors of antiretroviral use by HIV-infected pregnant women. Ann Intern Med 1999;130:979–86.
4. Healton C, Taylor S, Burr C, Dumois A, Lowenstein N, Kaye J. The impact of patient education about the effect of zidovudine on HIV perinatal transmission: Knowledge gain, attitudes, and behavioral intent among women with and at risk of HIV. Am J Prev Med 1996;12:47–52.[Medline]
5. Cooper ER, Nugent RP, Diaz C, Pitt J, Hanson C, Kalish LA, et al. After AIDS clinical trial 076: The changing pattern of zidovudine use during pregnancy, and the subsequent reduction in the vertical transmission of human immunodeficiency virus in a cohort of infected women and their infants. J Infect Dis 1996;174:1207–11.[Medline]
6. Wiznia AA, Crane M, Lambert G, Sansary J, Harris A, Solomon L. Zidovudine use to reduce perinatal HIV type 1 transmission in an urban medical center. JAMA 1996;275:1504–6.[Abstract]
7. Turner BJ, McKee L, Silverman NS, Hauck WW, Fanning T, Markson LE. Prenatal care and birth outcomes of a cohort of HIV infected women. J Acquir Immune Defic Syndr 1996;13:227–34.
8. Fanning TR, Turner BJ, Cosler LE, Oetjen-Gerdes L, Markson LE, McKee L, et al. Quality of Medicaid data for HIV/AIDS research: Examination of a statewide database. AIDS Pub Pol J 1995;10:39–47.
9. Christiansen DB, Williams B, Goldberg HI, Martin DP, Engelberg R, LoGerfo JP. Assessing compliance to antihypertensive medications using computer-based pharmacy records. Med Care 1997;35: 1164–70.[Medline]
10. Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy records: Methods, validity, and applications. J Clin Epidemiol 1997;50:105–16.[Medline]
11. Eldred LJ, Wu AW, Chaisson RE, Moore RD. Adherence to antiretroviral and pneumocystis prophylaxis in HIV disease. J Acquir Immune Defic Syndr Hum Retrovirol 1998;18:117–25.[Medline]
12. Hauck WW, McKee LJ, Turner BJ. Two-part survival models applied to administrative data for determining rate of and predictors for maternal-child transmission of HIV. Stat Med 1997;16: 1683–94.[Medline]
13. Anonymous. Designated care programs for patients with AIDS and HIV-related illnesses in designated care centers. NY State J Med 1989;89:542–3.[Medline]
14. Kotelchuck M. An evaluation of the Kessner adequacy of prenatal care index and a proposed adequacy of prenatal care utilization index. Am J Public Health 1994;84:1414–20.
15. Pizer SM. Numerical computing and mathematical analysis. Chicago: Science Research Associated Inc., 1975.
16. Gallant JE, Block DS. Adherence to antiretroviral regimens in HIV-infected patients: Results of a survey among physicians and patients. J Int Assoc Physicians AIDS Care 1998;4:32–35.
17. Eraker SA, Kirscht JP, Becker MH. Understanding and improving patient compliance. Ann Intern Med 1984;100:258–68.
18. Stephenson BJ, Rowe BH, Haynes RB, Macharia WM, Leon G. Is this patient taking the treatment as prescribed? JAMA 1993;269: 2779–81.
19. Urquhart J. Role of patient compliance in clinical pharmacokinetics. Clin Pharmacol 1994;27:202–15.
20. Rudd P, Byyny RL, Zachary V, LoVerde ME, Titus C, Mitchell WD, et al. The natural history of medication compliance in a drug trial: Limitations of pill counts. Clin Pharmacol Ther 1989;46:169–76.[Medline]
21. Singh N, Squier C, Sevek C, Wagener M, Nguyen MH, Yu VL. Determinants of compliance with antiretroviral therapy in patients with human immunodeficiency virus: Prospective assessment with implications for enhancing compliance. AIDS Care 1996;8:261–9.[Medline]
22. Moore RD, Stanton D, Gopolan R, Chaisson RE. Racial differences in the use of drug therapy for HIV disease in an urban community. N Engl J Med 1994;330:763–8.
23. Hendrickson G, Nevins JM, Chesnut TJ, Cross LT, Agins BD. Barriers to participation in AIDS Drug Assistance Programs in New York City. AIDS Pub Pol J 1993;8:126–35.
24. Siegel K, Gorey E. HIV-infected women: Barriers to AZT use. Soc Sci Med 1997;45:15–22.
25. Mehta S, Moore RD, Graham NM. Potential factors affecting adherence with HIV therapy. AIDS 1997;11:1665–70.[Medline]
26. Samet JH, Libman H, Steger KA, Dhawan RK, Chen J, Shevitz AH, et al. Compliance with zidovudine therapy in patients infected with human immunodeficiency virus, type 1: A cross-sectional study in a municipal hospital clinic. Am J Med 1992;92:495–502.[Medline]
27. Freeman RC, Rodriguez GM, French JF. Compliance with AZT treatment regimen of HIV-seropositive injection drug users: A neglected issue. AIDS Educ Prev 1996;8:58–71.[Medline]
28. Samuels JE, Hendrix J, Hilton M, Marantz PR, Sloan V, Small CB. Zidovudine therapy in an inner city population. J Acquir Immune Defic Syndr 1990;3:877–83.
29. Briers B, Morabia A, Hirschel B. A cohort study of drug users’ compliance with zidovudine treatment. Arch Intern Med 1994;154: 1121–7.[Abstract]
30. Markson LE, Cosler LE, Turner BJ. Implications of generalists’ slow adoption of zidovudine in clinical practice. Arch Intern Med 1994;154:1497–504.[Abstract]
31. Paes AH, Bakker A, Soe-Agnie CJ. Measurement of patient compliance. Pharm World Sci 1998;20:73–7.[Medline]
32. Mayer KH. Practical application and limitations of "eradication" therapy. AIDS Reader 1997;7:130–3.
This article has been cited by other articles:
|
|
 |
|
  D. J. Merenstein, M. F. Schneider, C. Cox, R. Schwartz, K. Weber, E. Robison, M. Gandhi, J. Richardson, and M. W. Plankey Association Between Living With Children and Adherence to Highly Active Antiretroviral Therapy in the Women's Interagency HIV Study Pediatrics, April 1, 2008; 121(4): e787 - e793. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Schectman, M. M. Nadkarni, and J. D. Voss The Association Between Diabetes Metabolic Control and Drug Adherence in an Indigent Population Diabetes Care, June 1, 2002; 25(6): 1015 - 1021. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
80% of days from first prescription to delivery.