Treatment seeking as a mechanism of change in a randomized controlled trial of a mobile health intervention to support recovery from alcohol use disorders

Abstract

Background: We estimated the efficacy of the Addiction-Comprehensive Health Enhancement Support System (A-CHESS) in increasing the use of services for addiction and examined the extent to which this use of services mediated the effects of A-CHESS on risky drinking days and abstinence from drinking.

Methods: We conducted secondary data analyses of the A-CHESS randomized controlled trial. Recruitment occurred in five residential treatment programs operated by two addiction treatment organizations. Participants were 349 adults with alcohol use disorders recruited two weeks before discharge from residential treatment. We provided intervention arm participants with a smartphone, the A-CHESS application, and an 8-month service plan. Control arm participants received treatment as usual. Telephone interviews at 4, 8, and 12-month follow-ups assessed past-month risky drinking days, past-month abstinence, and post-discharge service utilization (past-month outpatient addiction treatment and past-week mutual help including Alcoholics Anonymous and Narcotics Anonymous). Using mixed effects latent variable models, we estimated the indirect effects of A-CHESS on drinking outcomes, as mediated by post-discharge service utilization.

Results: Approximately 50.5% of participants reported outpatient addiction treatment and 75.5% reported mutual help at any follow-up interview in the year following randomization. Assignment to the A-CHESS intervention was associated with an increased odds of outpatient addiction treatment across follow-ups, but not mutual help. This use of outpatient addiction treatment mediated the effect of A-CHESS on risky drinking days, but not abstinence. The effect of A-CHESS through outpatient addiction treatment appeared to reduce the expected number of risky drinking days across follow-ups by 11%.

Conclusions: The mobile health (mHealth) intervention promoted the use of outpatient addiction treatment, which appeared to contribute to its efficacy in reducing risky drinking. Future research should investigate how mHealth interventions could link patients to needed treatment services and promote the sustained use of these services.

Keywords: Aftercare; Continuing care; Mediation; Mobile health; Randomized controlled trial; Smartphone; Treatment utilization.

Figure B.1

The system of equations underlying the mediation models. Preacher, Zyphur, and Zhang (2009) have named this approach the “unconflated multilevel model” implemented within a multilevel structural equation model framework. Note: RDD= risky drinking days, OP=post-discharge outpatient addiction treatment. An analogous approach was used for the outcome of abstinence and for the mediator of mutual help attendance. γ₁ in this figure corresponds to path c’ in Figure 1. γ₂ in this figure corresponds to path b₁ in Figure 1. γ₁₁ in this figure corresponds to path a₁ in Figure 1.

Figure 1

Path diagram for the analyses examining the mediation of the intervention effect on alcohol outcomes through post-discharge outpatient addiction treatment and mutual help. Mediation was examined with the product of coefficients approach (e.g., a₁*b₁). Tables 4 and 5 contain corresponding estimates from this figure computed with Mplus. Appendix B contains the system of equations underlying this path diagram.