Alcohol Use, HIV Infection, and Antiretroviral Adherence

by Broyles, Lauren Matukaitis

Abstract (Summary)
Alcohol use appears to negatively impact antiretroviral (ART) adherence, though conclusions about its effects are inconsistent, and the mechanisms of these effects are unclear. Accurate assessment of alcohol use is important for adherence counseling in HIV/AIDS. This secondary data analysis aimed to 1) determine if positive alcohol screening tests can predict ART adherence; 2) compare the effects of two ART adherence interventions with usual care across alcohol screening status; 3) explore mediation by self-efficacy in the relationship between adherence and several psychosocial variables; and 4) evaluate the psychometric properties and factor structure of the Alcohol Use Disorders Identification TestConsumption (AUDIT-C). The sample included 310 HIV+ adults on ART. Over 25% of the sample was AUDIT-C positive. Through sequential multiple linear regression analyses, AUDIT-C status (but not AUDIT-3 status) significantly added to the prediction of dose adherence (p=.005) and days under-dosing (p=.021) after controlling for confounders and covariates. In repeated measures analysis to determine if alcohol use impacts the effect of the interventions on dose adherence over time, only main effects for time and alcohol screening status were significant. Adherence was significantly lower at Time 2 than at baseline, F (1, 236.287) = 25.595, p = .000, and significantly lower for AUDIT-C positive individuals than for AUDIT-C negative individuals, F (1, 340.338) = 12.304, p = .001. In path analysis, near-significant results suggest partial mediation of the relationship between adherence and conscientiousness by self-efficacy. The internal consistency and test-retest reliability of the AUDIT-C were high. Multi-sample confirmatory factor analysis revealed factor invariance for sex, but the best-fitting model for race allowed partial invariance where AUDIT-C item 3 (episodic heavy drinking) was free to vary across whites/nonwhites, X2 (3, 310) = 1.818, p = .6111. Inconsistent AUDIT-C data and missing Time 2 adherence data were significantly related to baseline opioid use. In conclusion, positive AUDIT-C status may serve as an indirect indicator for ART nonadherence. The AUDIT-C appears to reliably assess alcohol use in PWHIV, but common modifications may risk compromising validity, particularly in drug users. Further attention to the cultural equivalence of the AUDIT-C across racial groups may be warranted.
Bibliographical Information:

Advisor:Judith A. Erlen, PhD, RN, FAAN; Susan Sereika, PhD; Adam Gordon, MD, MPH; Christopher M. Ryan, PhD

School:University of Pittsburgh

School Location:USA - Pennsylvania

Source Type:Master's Thesis



Date of Publication:08/09/2008

© 2009 All Rights Reserved.