Complex solutions for a complex problem: A meta-analysis of the efficacy of multiple-behavior interventions on change in outcomes related to HIV.

Abstract

The purpose of this meta-analysis was to examine the success of multiple-behavior interventions and to identify whether the efficacy of such programs depends on the number of recommendations prescribed and the type of outcomes measured. We conducted a synthesis of 136 research reports (N = 59,330) using a robust variance estimate model (Tanner-Smith et al., 2016) to study change between baseline and the first follow-up across multiple-behavior interventions, single-behavior interventions, and passive controls. Multiple-behavior interventions were more efficacious than their single-behavior counterparts (multiple-behaviors: d = .44 [95% confidence interval, CI [.27, .60]); single-behavior: d = .21 [95% CI [.00, .43]), with efficacy varying based on the type of outcomes measured. Publication bias analysis revealed a small asymmetry but controlling for it did not eliminate these effects. There was a strong linear relation between the number of recommendations prescribed by an intervention and intervention efficacy (B = .07, SE = .01, p < .001), with strongest improvements observed for interventions making five or more recommendations. These patterns remained when controlling for other intervention and population characteristics. Multiple-behavior interventions are successful in the HIV domain and increasing the number of recommendations made in the intervention generally maximizes improvements. These findings provide insights that may guide the design and implementation of integrated interventions. (PsycInfo Database Record (c) 2021 APA, all rights reserved).