Effect of 2 Integrated Interventions on Alcohol Abstinence and Viral Suppression Among Vietnamese Adults With Hazardous Alcohol Use and HIV

Key Points Question What is the effect of 2 clinic-based behavioral interventions of varying intensity on alcohol use and viral suppression among individuals with HIV and hazardous alcohol use? Findings In this 3-group randomized clinical trial involving 440 participants, the combined intervention and brief intervention, which both incorporated motivational enhancement therapy and cognitive behavioral therapy, significantly increased the percentage of days abstinent from alcohol at 12 months compared with the standard of care. Viral suppression at 12 months was significantly higher after the brief intervention than standard of care but not significantly higher after the combined intervention than standard of care. Meaning These findings support the use of the brief intervention in antiretroviral therapy clinics to reduce alcohol use and increase viral suppression in this high-risk population.


Introduction
Hazardous alcohol use is highly prevalent among people living with HIV (PWH). 1 In a large multisite clinical cohort of PWH in the United States, 27% reported hazardous alcohol use and 34% reported binge drinking. 24][5][6][7][8][9] Alcohol use has been associated with decreased adherence to antiretroviral therapy (ART) [10][11][12] and decreased viral suppression. 135][16] Despite the known adverse health consequences of alcohol use, there is a lack of evidence-based interventions for PWH, [17][18][19][20][21] and hazardous and heavy alcohol consumption are frequently not addressed in HIV health care settings.
Hazardous and heavy alcohol use in specific subpopulations of PWH, including people who inject drugs (PWID), may place them at even higher risk for health consequences due to comorbidities as well as transmission of HIV and other infections.Hazardous alcohol use among PWID who are living with HIV is independently associated with decreased ART adherence and viral suppression 13 as well as sharing needles and/or syringes, having multiple sex partners, and engaging in sex without condoms. 10,15,22 conducted a 3-group randomized clinical trial to compare the effectiveness of the combined intervention and the brief intervention integrated into ART clinics with the standard of care (SOC) on percentage of days abstinent from alcohol (assessed with timeline follow-back interviews and confirmed using the alcohol biomarker phosphatidylethanol) 23 and viral suppression at 12 months after enrollment.The combined intervention and brief intervention have been highly effective in reducing alcohol use among people living with HIV with hazardous or heavy alcohol use in US settings [24][25][26] and were selected for their potential applicability to the Vietnamese context. 27

Study Setting
The study was conducted in Thai Nguyen, a semi-urban province in North Vietnam located 75 kilometers north of Hanoi with a population of approximately 1 million residents.Thai Nguyen has 12 government outpatient ART clinics that are the sole source of ART medications in the province.We recruited participants and conducted the study in the 7 largest outpatient ART clinics.

Ethical Approvals
The study was approved by the institutional review boards at the University of North Carolina Gillings School of Global Public Health, the Johns Hopkins University Bloomberg School of Public Health, and the Thai Nguyen Center for Preventive Medicine.All study participants provided written informed consent in Vietnamese.The study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.The trial protocol and statistical analysis plan are available in Supplement 1.

JAMA Network Open | Substance Use and Addiction
Effect of 2 Interventions on Alcohol Use and Viral Suppression Among People With HIV

Randomization and Masking
Participants were randomly assigned to the combined intervention, the brief intervention, or the SOC group at a ratio of 1:1:1.To achieve this, a randomization schedule was generated in SAS version 9.4 (SAS Institute) using permuted-block randomization with a block size of 3, such that each block contained a random assignment to each of the 3 groups.Participants randomly assigned within 1 triplet block all belonged to the same ART clinic.Masking in the field was not feasible due to the nature of the intervention; however, analysts and principal investigators were masked to study assignment until the analyses for this paper were complete.

Eligibility Criteria
Eligibility criteria included being enrolled at a study ART clinic and receiving ART; having hazardous alcohol use, defined as having an Alcohol Use Disorders Identification Test-Consumption score of at least 4 for men or at least 3 for women; being aged 18 years or older; and planning on residing in Thai Nguyen for the next 24 months.The sole exclusion criterion was having a Clinical Institute Withdrawal Assessment score of at least 10, out of concerns for risk of alcohol withdrawal.

Recruitment Procedures
Recruitment began in March 2016 and continued through May 2017.Participants were followed up for 12 months.Any ART client who accessed care in the study clinics during the recruitment period was approached for participation (eTable 1 in Supplement 2).Once saturation was reached (ie, the study team did not identify any new potential participants for at least 1 week), we moved to recruitment in the next clinic, using a random ordering of clinics.HIV clinicians introduced the study to patients and referred them to study staff if interested, at which time eligibility was assessed and written informed consent was obtained.Viral load was not assessed prior to enrollment.

Study Procedures
Participants in all 3 groups received a standard Ministry of Health recommendation from their HIV clinician to decrease alcohol use and a referral to harm reduction services and to treatment of hepatitis B and C viruses, sexually transmitted diseases, and tuberculosis.The processes for selecting and culturally adapting the 2 interventions are described in detail elsewhere. 27We conducted training and provided supervision through weekly meetings using a completion checklist and the Yale Adherence and Competence Scale (YACS; range 1-7, with 7 indicating excellent fidelity) 28 to ensure sessions were standardized across our team of paraprofessional counselors who delivered the interventions.A team of 2 counselors, 1 woman and 1 man, was assigned to the combined intervention group, and a different counseling team, 1 woman and 1 man, was assigned to the brief intervention group.
Because the concept of alcohol reduction is novel in this setting, we selected a combined intervention approach that includes cognitive behavioral therapy and motivational enhancement therapy. 27,29The brief intervention has similar components as the combined intervention but is briefer.
The combined group included 6 individual face-to-face sessions delivered 1 week apart and 3 optional group sessions delivered concurrently.The first session used motivational enhancement therapy, adopting a guiding but nonjudgmental style to enhance motivation.Subsequent sessions used cognitive behavioral therapy, building skills to effectively refuse alcohol, to manage cravings and high-risk situations for alcohol use, and to develop self-efficacy.Sessions also provided information on the harmful effects of drinking on HIV and overall health.Group sessions reinforced skills learned in the individual sessions and provided a forum to discuss experiences (eTable 2 in Supplement 2).
Participants in the brief intervention group participated in 2 individual face-to-face sessions and 2 individual booster telephone sessions.Based on Project TrEAT, 30 the content of the brief intervention sessions included elements used in the combined intervention, including information about alcohol and its effects as well as alcohol behavior change strategies (eTable 2 in Supplement 2).

JAMA Network Open | Substance Use and Addiction
Effect of 2 Interventions on Alcohol Use and Viral Suppression Among People With HIV

Statistical Analysis
The study was designed to have 80% power to detect the following: (1) greater than 50% effect size in percentage of abstinent days at 12-month follow-up minus baseline, comparing the combined intervention vs SOC (or the brief intervention vs SOC); (2) greater than 35% effect size in percentage abstinent days at 12-month follow-up minus baseline, comparing the combined intervention vs the brief intervention; and (3) greater than 6% change in viral suppression at 12-month follow-up minus baseline, comparing the combined intervention vs SOC (or the brief intervention vs SOC).For each outcome, we evaluated the extent to which the mean values differed across groups after baseline by intention-to-treat.To do this, first we estimated the proportion of patients in each of the 3 groups for whom the outcome was measured and compared among groups using the deviance χ 2 test.We also estimated the mean outcome level in each group and compared among groups using a Wald χ 2 test.We also assessed the possibility of selectively differential attrition across arms at each postbaseline visit.To do this, we estimated and compared the mean baseline outcome among groups for the patients who provided outcome data at that postbaseline visit.If an imbalance was found, we also produced postbaseline regression-adjusted outcome estimates by recalibrating the baseline distribution of that group to the baseline distribution across all groups combined. 33In short, the overall group distribution of the baseline outcome can be estimated at each value x as 1 / total sample size n.Therefore, if the regression of the mean postbaseline outcome y given baseline value x within arm a is estimated as y ¯a × x, then the regression-adjusted outcome average in that arm is

JAMA Network Open | Substance Use and Addiction
Statistical analyses were conducted in R version 3.4.3(R Project for statistical computing) and SAS version 9.4 (SAS Institute).Statistical significance was set at P < .05,and all tests were 2-tailed.

Results
During

Baseline Characteristics
Demographic and clinical characteristics at baseline were comparable across groups (Table 1

Percentage of Days Abstinent
At baseline, the mean (SD) percentage of days abstinent in the past 30 days was 40% (1.6%), with no significant differences across arms (

Viral Load
The percentage of participants achieving viral suppression (ie, <20 copies/mL) was similar at 3 and 6 The observed effect sizes for abstinent days were within the range of those planned in the protocol to have enough power.For viral load, the observed effect for BI vs SOC was similar to the planned power, although the observed combined intervention vs SOC effect was smaller than what was planned (eTable 5 in Supplement 2).

Secondary Outcomes
Relative

Discussion
In this randomized clinical trial, we observed that alcohol reduction can be achieved through a brief intervention, leading to substantial improvement in viral suppression.Both integrated alcohol As described elsewhere, 29 the combined intervention is more resource-intensive than the brief intervention ($30 additional cost per participant); however, the brief intervention was at least as, if not more, effective than the combined intervention on biologically validated alcohol measures and viral suppression.Part of the explanation for our findings may be that, given the lack of alcohol reduction programs in Vietnam, even 2 in-person sessions and 2 telephone sessions of alcohol reduction counseling can increase awareness of the harms of alcohol and teach coping skills to manage high risk moods and/or situations for alcohol use.Importantly, these results underscore that a brief intervention provided by paraprofessional counselors effectively reduced alcohol consumption and increased viral suppression among people with hazardous or heavy drinking being treated in ART clinics.Practical implications include integrating the brief intervention into ART clinics and screening for ART clients with hazardous drinking.Our findings highlight the need for a   rigorously evaluated implementation study in multiple settings to evaluate whether the same results could be achieved in real-world settings in a wide variety of low-to middle-income countries.
Heavy alcohol use among PWH is prevalent and is associated with lower ART adherence, lower rates of viral suppression, and increased sexual and injecting risk behaviors. 10,11,16People who drink alcohol excessively may be more likely to become HIV infected. 21But higher rates of drinking among PWH is likely due in part to coping with the stress of HIV-related stigma and negative life events, such as an HIV diagnosis, as well as lack of social support. 21,34Training in alternative means of coping may both reduce alcohol use and enhance well-being.But alcohol treatment has rarely been incorporated into HIV care settings.In the United States, a stepped alcohol treatment intervention for patients with HIV and alcohol use disorder incorporating a physician-managed medication-based intervention, motivational enhancement therapy, and referral to intensive outpatient or residential treatment led to improvements in the proportion of days abstinent and the proportion of patients with an undetectable HIV viral load at 52 weeks. 35We showed the effects of 2 counseling-only interventions on these same outcomes at 52 weeks, which, unlike physician-managed medicationbased interventions, can be realistically scaled in low-resource settings.Specifically, both the combined and brief interventions have a highly structured format that are manualized and can be delivered by trained paraprofessionals.Notably, alcohol reduction and associated improvement in viral load were achieved in patients who were drinking at hazardous or heavy levels but were not seeking alcohol treatment.
At 12 months, viral suppression was significantly higher in the brief intervention group compared with the SOC group; however, the percentage of virally suppressed participants was also higher in the brief intervention group than the SOC group at baseline.After standardizing both baseline levels of viral suppression and dropout rates over time, the brief intervention group still had a significantly higher level of viral suppression compared with the SOC group.This suggests that the brief intervention reduced alcohol use, which in turn improved ART adherence, leading to greater viral suppression compared with SOC.
Our study included only individuals receiving ART and did not exclude those who had already achieved viral suppression; therefore, a high proportion of our study population already had viral suppression at baseline.We anticipate that including people who are not receiving ART and excluding those who had achieved viral suppression would increase the effect of the interventions on viral suppression.In settings where viral suppression is already high, it is important to reach the final non-virally suppressed 10% to 15% of patients, who are often the most challenging to engage in HIV care.

Limitations and Strengths
This study has limitations.Self-reported alcohol abstinence is susceptible to social desirability bias.To assess the extent of this bias, we compared phosphatidylethanol level of less than 8 ng/mL, the established cut-off for abstinence, to self-reported alcohol abstinence.Only 1% to 6% reported no alcohol use in the past 21 days but had phosphatidylethanol levels of greater than 8 ng/mL, suggesting social desirability at play.However, we saw a similar percentage of those with phosphatidylethanol levels of less than 8 ng/mL who reported alcohol use in the past 21 days, suggesting possible imperfections in either or both measures that work in both directions.Therefore, while there was a strong association between phosphatidylethanol levels of less than 8 ng/mL and self-reported alcohol abstinence, it is still unclear how to best use phosphatidylethanol.
Most participants in this study were men.This was expected, given that 65% of PWH are men 36 and most individuals with alcohol use disorder in Vietnam are men. 36Therefore, we anticipate that our findings will generalize to our target population in Vietnam, although they may not be generalizable to other contexts in which women account for a greater proportion of PWH and people with alcohol use disorders This study also has several strengths.The interventions were designed to be scalable.Both interventions were conducted in the very clinic settings in which they could be implemented during scale-up.By integrating into an existing structure and enrolling current ART clients, we had the opportunity to evaluate the interventions in real-world clinic settings to directly inform future scale-up. 37The role of counselors can likely be fulfilled by peers, social workers, counselors, or clinicians.The counselor role did not require a high level of education, and some counselors did not have a bachelor's degree.Indeed, lay counselors have effectively delivered psychological treatment for men with alcohol dependence in primary care settings in India. 38Uptake of both interventions was excellent.Attendance was high compared with other trials of the combined and brief interventions 38 and was relatively higher for the brief intervention, which was expected given the lower burden on the participants (ie, 6 in-person sessions vs 2 in-person sessions plus 2 brief telephone sessions).Duration of sessions, including the brief intervention telephone sessions, was also longer than expected, reflecting high engagement during the sessions.Counselor fidelity to the intervention manual was also acceptable.

Conclusions
In summary, the brief intervention, which combined in-person and telephone sessions, and the combined intervention, which included individual and group sessions, increased the percentage of days abstinent from alcohol (confirmed using phosphatidylethanol), decreased the number drinks per drinking day, and decreased the number of heavy drinking days.These effects were sustained through 52 weeks.Furthermore, the brief intervention increased viral suppression at 52 weeks.
Based on the strength of the brief intervention compared with the combined intervention as well as the lower resources required to sustain the brief intervention compared with the combined intervention, implementation of the brief intervention should be a priority in other populations of ART clients.
Excluding participants who died, 405 of 435 participants (93.1%) completed the 3-month follow-up, 410 of 432 (94.9%) completed the 6-month follow-up, and 390 of 425 (91.8%) completed the 12-month follow-up.The most common reasons for missed visits at 12 months were incarceration and relocation.By the 12-month visit, 390 of 440 participants (88.6%) provided abstinence and viral load outcomes.

Table 1 .
Baseline Characteristics by Group

Table 1 .
Baseline Characteristics by Group (continued)

Table 2 .
Percentage of Days Abstinent From Alcohol in the Last 30 Days, by Trial Group and Visit

Table 3 .
Percentage of Participants With Viral Suppression by Trial Group and Visit a

Table 4 .
Percentage of Participants With Viral Suppression With Covariate Standardized to Have the Same Baseline Percentage of Viral Suppression a b Each arm's covariate has been standardized to have 84% of participants virally suppressed at baseline.cTwoparticipantswho did not complete their 3-month visit had 3-month viral load data.JAMA Network

Open | Substance Use and Addiction Effect
of 2 Interventions on Alcohol Use and Viral Suppression Among People With HIV Self-reported alcohol abstinence was a primary outcome, but other measures of alcohol reduction may also achieve viral suppression.Comparing the proportion of patients who reduced alcohol consumption between the SOC and combined intervention groups and the SOC and brief intervention groups at 12 months is important to understand how the interventions may operate to affect viral suppression.We found that both number of drinks per drinking day and number of heavy drinking days at 12 months were reduced in the combined and brief intervention groups compared with the SOC group, providing additional support for the effectiveness of the combined and brief interventions.

Open | Substance Use and Addiction SUPPLEMENT 2. eTable 1.
Number of Participants Enrolled in the Study From Each ART Clinic, Thai Nguyen Province, Vietnam eTable 2. Descriptions of Each Alcohol Reduction Intervention Evaluated eTable 3. Cohen d for Difference Among Groups for Mean Percentage of Days Abstinent From Alcohol in the Last 30 Days eTable 4. Difference Among Groups for Percentage Viral Suppression in the Last 30 Days eTable 5. Observed Effect Size vs Planned Power for Primary Outcomes eTable 6. Mean Number of Drinks per Drinking Day in the Last 30 Days, by Groups and Visits eTable 7. Cohen d for Difference Among Groups for Mean Number of Drinks per Drinking Day in the last 30 Days eTable 8. Mean Number of Heavy Drinking Days in the Last 30 Days, by Groups and Visits eTable 9. Cohen d for Difference Among Groups for Mean Number of Heavy Drinking Days in the last 30 Days eTable 10.Self-reported Alcohol Abstinence in the Last 3 weeks and Phosphatidylethanol Levels Less Than 8 ng/mL