JMIR Publications

ABSTRACT

Background:

mHealth for substance misuse (i.e., alcohol/drugs) has the potential to bypass common barriers to treatment-seeking. Ten years following the release of the first mobile app targeting substance misuse, their effectiveness, usage and acceptability remains unclear.

Objective:

To conduct a systematic literature review of trials evaluating mobile app interventions for substance misuse.

Methods:

The review was conducted according to PRISMA guidelines and risk of bias was assessed for each study. Relevant databases were searched and articles in any language were included in the review if the intervention was delivered via a mobile app, a controlled trial and reported consumption outcomes (e.g., frequency of daily alcohol use).

Results:

Twenty studies met eligibility criteria across a range of substances: alcohol (11), tobacco (6), alcohol and tobacco (1), illicit drugs (1), illicit drugs and alcohol (1). Most samples were from the general community, with five studies using clinical samples. Analysed intervention sample sizes ranged from 22 to 14,228, and content was considerably diverse, from simple stand-alone apps delivering monitoring or psychoeducation, to multi-component apps with interactive features, audio-content, access to counsellors, or used as adjuncts alongside face-to-face or pharmacotherapy treatment. Intervention duration ranged from 1 to 35 weeks, with notification frequency ranging from none (use as needed) to multiple times per day. Six of the 20 app interventions reported significant reductions in substance use at post or follow-up compared to a control condition, with small to moderate effect sizes. A further two reported significant reductions during the intervention but not at post-treatment, and a third reported a significant interaction of two app intervention components in an exploratory analysis.

Conclusions:

While most app interventions in this review were associated with reductions in substance misuse, less than one-third were significantly better than comparison conditions. Five out of six of the apps that reported intervention effects targeted alcohol (of those, one targeted alcohol and illicit drugs, and another alcohol and tobacco) and one targeted tobacco. Three out of six included feedback (personalised, peer, or normative) and three had high risk of bias and three low risk. All six included interventions of 6-weeks or longer. Common study limitations were small sample sizes, risk of bias, lack of relevant details and poorly balanced control conditions. Large rigorous trials are required to understand: which app components are most effective, length of engagement required, and sub-groups most likely to benefit. In sum, evidence to date for the effectiveness of substance misuse apps is not compelling, although comparison conditions often preclude conclusions about apps per se. We discuss future approaches that can help to ascertain whether the promise of mHealth interventions for substance misuse can be fulfilled.