Integrated stepped alcohol treatment for patients with HIV and at-risk alcohol use: a randomized trial

At-risk alcohol use, defined as alcohol consumption of ≥ 14 drinks per week or ≥ 4 drinks per occasion in men ≤ 65 years old or ≥ 7 drinks per week or ≥ 3 drinks per occasion in women or men > 65 years old [1], is an important problem that warrants intervention in routine medical settings [2]. This is particularly true for patients with HIV (PWH) given that at-risk levels of alcohol use may interfere with achievement of HIV viral suppression [3], increase risk of morbidity and mortality [4‐6], and lead to risk behaviors and ongoing HIV transmission [7]. Accordingly, guidelines recommend that patients with at-risk alcohol use, including those with HIV, should receive brief interventions with subsequent treatment to reduce their alcohol use [8, 9] and integrated with HIV care [10].

Despite its potential to improve individual and public health, brief intervention with subsequent indicated treatment is inconsistently delivered to PWH [11], and, to date, only a limited number of studies have been specifically designed to address unhealthy alcohol use (defined as the spectrum of alcohol use including at-risk drinking and alcohol use disorder) among PWH [12‐17]. These studies have generally focused on evaluation of a specific medication (i.e., naltrexone) [12] or behavioral intervention [13‐16] and some focused on a specific patient population (e.g., women, men who have sex with men). None of these studies offered a comprehensive package that allowed evaluation of initial patient response to a lower intensity intervention prior to adding additional services. Such “stepped care models” allow for tailoring of treatment based on patient response while employing multidisciplinary team members as needed to deliver specific components of care to maximize resource allocation. Stepped care models have been successfully applied to a variety of medical conditions (e.g., depression, hypertension, chronic pain) [18, 19] addressed in routine medical settings, but rarely applied to address alcohol use [20] and specifically among PWH [13]. We have recently reported the benefits of stepped alcohol treatment for PWH who have alcohol use disorder (as defined by the Diagnostic and Statistical Manual criteria to identify individuals with loss of control and adverse consequences from alcohol use) or lower levels of alcohol use in the presence of liver disease [21, 22]. Notably, this model, because of its approach involving integration of alcohol treatment into routine medical settings with stepped care and demonstrated impact on improving outcomes among PWH with alcohol use disorder, has been endorsed by the National Institute on Alcohol Abuse and Alcoholism [23, 24]. To our knowledge, no prior studies have examined the impact of stepped alcohol treatment integrated with HIV care to address alcohol and HIV outcomes specifically among PWH with at-risk alcohol use and that explicitly includes components designed to promote behavioral and medication-based treatments as indicated. Thus, the aims of this study were to examine the effectiveness of integrated stepped alcohol treatment (ISAT) versus treatment as usual (TAU) on alcohol use and HIV outcomes among PWH with at-risk alcohol use. We hypothesized that ISAT would be associated with improved drinking and health outcomes compared to TAU, with the primary hypothesis being that ISAT would lead to fewer drinks per week compared with TAU.

The proportion of participants with no heavy drinking days was not significantly different among those randomized to ISAT compared to TAU at week 24 [47 vs. 43%, adjusted odds ratio [AOR] (95% CI) = 0.8 (0.2, 2.6)] and at week 52 [52 vs. 58%, AOR (95% CI) = 0.4 (0.1, 1.5)] (Fig. 3b, Additional file 1: Table S1). The percentage of days abstinent did not differ among those randomized to ISAT compared to TAU at week 24 (LSmean = 60 vs. 66%), AMD [95% CI] = − 2 [− 10, 6] and at week 52 [LSmean = 54 vs. 67%, AMD = − 11 (− 25, 3)]. At week 24, PEth concentrations did not differ between participants in the ISAT and TAU groups.

Our findings extend the existing literature focused on developing interventions to address at-risk alcohol use among PWH [17]. It is notable that clinic-based interventions to address at-risk alcohol use among PWH are relatively scarce. Based on a systematic review of the literature and meta-analysis, Scott-Sheldon and colleagues reported in 2017 that there were only 21 studies that reported on individual-level interventions designed to address alcohol use among PWH, 71% of which were clinic-based [17]. Consistent with our approach, studies that targeted alcohol alone (vs. multiple HIV-related behaviors) and that were clinic-based (vs. recruiting from other/mixed settings), were found to be more successful at decreasing alcohol use[14‐17], though these effects may vary based on baseline level of motivation and intervention strategy[13]. Building on this literature and our experiences evaluating ISAT among PWH with higher and lower levels of alcohol use[21, 22], the current study provides additional support for approaches that include MI-based interventions that target alcohol use in clinic-based settings to promote alcohol reduction among PWH. Specifically, we found that some PWH – even when not specifically incentivized to do so—will attend visits to address their alcohol use. In addition, we found that graded interventions and those incorporating MI are appropriate; 50% of participants met criteria for MET after an initial brief intervention with booster and nearly 45% responded to these MET sessions and did not meet criteria for further “stepping up.” The stepped care model served to enhance receipt of delivery of evidence-based interventions to optimize resources given its sequential nature. This stepped care approach is relevant for primary care based settings given the potential to incorporate multidisciplinary team members to deliver specific components, such as the brief intervention [44]. Whether this model and treatment receipt translates into changes in alcohol use and HIV-related outcomes for PWH with at-risk alcohol use, however, is not yet clear. In addition, our findings should be interpreted in the context of recent VA-based analyses demonstrating that brief intervention is associated with decreased likelihood of receipt of specialty addiction services [45]; however, it is an empirical question how these findings generalize to PWH seen in HIV clinics, where mental health services are often embedded. In addition, the recognized harmful effects of concurrent alcohol and tobacco use [46, 47] and our findings that current smoking was negatively associated with alcohol reductions over time together reinforce the need for evaluation of targeted interventions to address these commonly co-occurring, mutually-reinforcing behaviors [48]. Stepped care interventions, including behavioral interventions (e.g., contingency management or cognitive behavioral therapy) in combination with medications, such as varenicline, merit further evaluation.

Our study should be interpreted in the context of its limitations. First, despite a multi-pronged approach to enroll participants, we were unable to meet recruitment targets and were underpowered to evaluate the impact of our intervention on participant-level outcomes. We believe this relates to low prioritization to address alcohol use among this patient population in the context of at-risk alcohol use [49‐51], underlying depressive symptoms, as well as challenges with logistics associated with study participation (e.g., travel, childcare responsibilities, employment) as has been observed in other alcohol intervention studies for PWH [52]. Formative evaluation to better understand why patients do and do not opt to participate in an alcohol intervention study and how these populations differs may be useful. Meanwhile, novel strategies, including use of incentives and remote methods for intervention delivery, to promote patient engagement in alcohol interventions may be useful and are actively being evaluated [53]. Second, we assumed missing data were missing at random; however, we acknowledge that there is no method to validate this assumption and that a missing not at random (MNAR) process may have biased results whereby participants with missing data may have been more likely to have higher levels of alcohol use. Given that more patients in the ISAT group were lost to follow-up, this may have biased us towards the null hypothesis. Third, our findings may not be generalizable to women or patients receiving care outside the VA. Fourth, we relied on research assistants to identify potentially eligible participants and to determine whether they meet criteria for being stepped up. Future study is needed to determine how this might be translated into routine clinical practice. Lastly, our control condition, TAU, may reflect a higher level of alcohol-related care than is routine in other HIV treatment settings, given routine implementation of AUDIT-C screening and electronic health record prompts to trigger brief intervention when indicated.

ISAT may hold promise as a model to promote alcohol-related care among PWH across the spectrum of alcohol use. Future studies are needed to enhance patient-level engagement to initiate and remain in alcohol interventions and bolster the impact of ISAT on patient-level outcomes among PWH with at-risk alcohol use.