Drinking motives as prospective predictors of outcome in an intervention trial with heavily drinking HIV patients

Abstract

Background

Heavy alcohol consumption in HIV patients is an increasing health concern. Applying the drinking motivational model to HIV primary care patients, drinking motives (drinking to cope with negative affect, for social facilitation, and in response to social pressure) were associated with alcohol consumption at a baseline interview. However, whether these motives predict continued heavy drinking or alcohol dependence in this population is unknown.

Methods

Participants were 254 heavy-drinking urban HIV primary care patients (78.0% male; 94.5% African American or Hispanic) participating in a randomized trial of brief drinking-reduction interventions. Drinking motive scales, as well as measures of alcohol consumption and alcohol dependence, were administered at baseline. Consumption and dependence measures were re-administered at the end of treatment two months later. Regression analyses tested whether baseline drinking motive scale scores predicted continued heavy drinking and alcohol dependence status at the end of treatment, and whether motives interacted with treatment condition.

Results

Baseline drinking to cope with negative affect predicted continued heavy drinking (p < 0.05) and alcohol dependence, the latter in both in the full sample (adjusted odds ratio [AOR] = 2.14) and among those with baseline dependence (AOR = 2.52). Motives did not interact with treatment condition in predicting alcohol outcomes.

Conclusions

Drinking to cope with negative affect may identify HIV patients needing targeted intervention to reduce drinking, and may inform development of more effective interventions addressing ways other than heavy drinking to cope with negative affect.

Introduction

Over 1.1 million individuals in the US are infected with HIV (Centers for Disease Control and Prevention, 2012). Maintaining low levels of alcohol consumption is highly relevant to maintaining the health of these HIV-infected individuals, for several reasons. HIV-infected individuals with heavy drinking or alcohol dependence may experience lowered immune functioning (Shuper et al., 2010) and higher viral loads (Hahn and Samet, 2010). Liver disease is also a significant contributor to mortality among HIV-infected individuals (Antiretroviral Therapy Cohort Collaboration, 2010, Palella et al, 2006), due to high rates of Hepatitis B and C, antiretroviral (ART) medication hepatotoxicity, and alcohol consumption (Barve et al., 2010). However, high rates of alcohol use disorders among HIV-infected individuals dying of liver disease suggest considerable contribution of alcohol use to liver disease in this population (DeLorenze et al., 2011). Additionally, providers are sometimes reluctant to provide ART medication to patients with alcohol use disorders (Loughlin et al., 2004), and when they are prescribed, these patients have lower adherence (Azar et al., 2010). Due to these and other consequences of heavy alcohol involvement among those with HIV, better understanding of who is most at risk for prolonged heavy drinking and alcohol dependence in this population is important.

Relatively little is known about predictors of heavy alcohol involvement among individuals with HIV. Clinical trials with drinking-reduction outcomes in HIV populations show varied results (Meade et al., 2010, Papas et al., 2011, Parsons et al., 2007, Rotheram-Borus et al., 2009, Samet et al., 2005, Velasquez et al., 2009, Wong et al., 2008), with no consistent relationship between study design and drinking outcome, demonstrating that we do not yet understand what precedes changes in drinking. In cross-sectional observational studies, correlates of drinking include stressful experiences, depressive symptoms, poor coping strategies, and low self-efficacy (Longmire-Avital et al., 2012, Pence et al., 2008). A longitudinal study of women with HIV replicated depression as a prospective predictor of drinking (Cook et al., 2009, Cook et al., 2012). However, given the limited knowledge base on predictors of heavy drinking and alcohol dependence status over time in HIV samples, more information is needed, particularly from samples with both genders.

The motivational model of alcohol use may help identify predictors of persistent heavy alcohol involvement among individuals with HIV. This theory posits that alcohol consumption is driven by the anticipated benefits of drinking, which may include decreasing negative states or achieving positive effects (Cox and Klinger, 1988). Commonly studied motives address coping with unpleasant emotions (e.g., depressed affect, anxiety), enhancement of positive feelings, and responding to social situations (Kuntsche et al., 2005). Drinking motives (or reasons for drinking) predict alcohol consumption in general population samples, as shown by cross-sectional (Abbey et al., 1993, Carpenter and Hasin, 1998b, Mezquita et al., 2011, Tragesser et al., 2007, Trocki and Drabble, 2008) and prospective (Beseler et al., 2011, Crutzen et al., 2012, Tragesser et al., 2007) studies. Drinking motives also predict alcohol dependence status/symptoms, in both cross-sectional (Tragesser et al., 2007, Trocki and Drabble, 2008) and prospective (Beseler et al., 2008, Carpenter and Hasin, 1998a, Tragesser et al., 2007) community studies.

Recently, we evaluated cross-sectional associations between drinking motives and alcohol consumption in a heavily drinking HIV primary care sample recruited to participate in a drinking-reduction randomized trial (Elliott et al., 2013). Three of the four original subscales from the Reasons for Drinking Scale (Carpenter and Hasin, 1998b) had clear factor structure and good internal consistency in this sample: (a) drinking to cope with negative affect (e.g., sadness, boredom, irritability), (b) drinking for social facilitation, and (c) drinking in response to social pressure. These motives were associated with many aspects of past-year drinking measured at baseline: coping with negative affect was associated with higher levels of all drinking and heavy drinking measures, drinking for social facilitation was associated with more frequent drinking to intoxication, and drinking in response to social pressure was associated with lower drinking quantity and binge frequency. Although identifying cross-sectional correlates of heavy drinking in HIV patients is informative, identifying predictors of continued, persistent heavy drinking or alcohol dependence may be particularly important to improve clinical practice. Determining predictors of drinking despite intervention may help identify the most robust drinking predictors. Given the cross-sectional associations we found, we hypothesized that drinking motives would also predict continued heavy drinking and alcohol dependence symptoms in this heavily drinking HIV-infected sample, even despite participation in a drinking reduction trial. Drinking to cope with negative affect was a drinking motive of particular interest, given elevated rates of depression among individuals with HIV (Bing et al., 2001, Ciesla and Roberts, 2001, Zanjani et al., 2007), and studies showing that depression itself predicts later drinking in HIV-infected women (Cook et al., 2009, Cook et al., 2012). An additional question of interest was whether drinking motives would be differentially predictive among patients receiving different types of drinking-reduction interventions.

The present study therefore aimed to determine if drinking motives predict continued heavy alcohol consumption and alcohol dependence despite intervention among heavy-drinking HIV patients, and to determine if the effects of motives differed by treatment type. To do this, we conducted additional analysis of data from a randomized trial of brief drinking-reduction interventions with urban minority HIV primary care patients (Hasin et al., 2013), also used for the cross-sectional validation of the scale (Elliott et al., 2013). In this study, patients were randomized to one of three conditions, all involving brief (20–25 min) sessions at baseline and briefer (5–10 min) sessions at 30 and 60 days. One condition was a Motivational Interview (MI) session. Another was the MI session plus HealthCall (MI + HealthCall), which involved daily self-monitoring (2–3 min) via automated telephone interactive voice response technology with personalized feedback from the self-monitoring data provided at 30 and 60 days. The third, an attentional control condition, consisted of advice to reduce drinking and a video on HIV self-care without alcohol content. Although all groups reduced drinking, patients receiving MI + HealthCall had significantly greater drinking reduction than others (Hasin et al., 2013). In the current study, we examined three questions. First, do baseline motives predict whether these heavily drinking patients continued to drink heavily at end-of-treatment (i.e., do drinking motives predict continued heavy drinking despite intervention)? Second, do baseline motives predict alcohol dependence status at end-of-treatment? Third, do the effects of motives differ by treatment condition?