BACKGROUND/RATIONALE:
Adherence to antiretroviral therapy is important for HIV-infected persons, because nonadherence can lead to progression of disease and development of resistant strains of HIV. Few studies have focused on adherence to the newer combination antiretroviral therapies. There is no consensus on how to measure adherence to antiretrovirals, and little is understood about the deteminants of adherence to antiretrovirals.
OBJECTIVE(S):
To identify the important factors associated with adherence to combination antiretroviral therapy, and the impact of adherence on outcomes.
We will use data from patient survey and from VA's national HIV Registry to conduct cross-sectional and longitudinal analyses of patient self-reported adherence, adherence measured by prescrption refills, and outcomes. This project will interface with two Service-Directed Research (SDR) Projects proposed by the San Diego campus of VA's Sepulveda Center for Health Care Provider Behavior (hereafter the San Diego group), the coordinating center for the HIV QUERI. The San Diego Group has agreed to collaborate on our study.
METHODS:
An Expert Panel will develop six to ten measures of patient self-reported adherence, and four to six measures of adherence based on prescription refill data. HIV-infected veterans were recruited from the Houston and San Diego VAMCs to participate in focus group interviews. Analysis of the transcripts from the focus groups will be used to develop survey questions to elicit patients' perceptions about adherence. Survey items will also be constructed to elicit patients' self-reported adherence, using the measures developed by the Expert Panel. We will use the HIV Registry database to conduct longitudinal multiple regression analyses of prescription refill adherence and outcomes using data from 1996 through 1999. The pharmacy package of the HIV Registry contains information on all prescriptions filled, including drug names, dates, and dosages. The adherence measures used will be constructed by the Expert Panel. Outcome measures will include achievement of undetectable viral load, change in CD4 count, progression to AIDS, survival, and hospitalization. The explanatory variables will include patient-level factors (socioeconomic characteristics; illness severity), characteristics of the treatment regimen (dosing schedule, side effects), and characteristics of the medical facility (dedicated AIDS units, physician experience with HIV patients).
The Generalized Estimating Equation (GEE) approach will be used for estimation. GEE will account for repeated observations on the same patient and for time-dependent covariates relevant to the longitudinal analyses of prescription refills. GEE will also account for clustering of patients within treatment regimen and facility. A logistic link function will be used for dichotomous dependent variables, and a normal link function for continuous dependent variables.
FINDINGS/RESULTS:
None at this time.
IMPACT:
Not yet measured.
PUBLICATIONS:
Journal Articles
- Menke TJ, Rabeneck L, Hartigan PM, Simberkoff MS, Wray NP. Clinical and socioeconomic determinants of health care use among HIV-infected patients in the Department of Veterans Affairs. Inquiry : A Journal of Medical Care Organization, Provision and Financing. 2000; 37(1): 61-74.
