JMIR Publications

ABSTRACT

Background:

Digital health interventions (DHI) that target physical inactivity and sedentary behaviour typically have modest beneficial effects on measures of activity. Effectiveness of DHIs is commonly assumed to be related to level of user DHI engagement, including measures of both use and users subjective experience. However, little is known about the relationships between measures of DHI engagement and physical activity or sedentary behaviour.

Objective:

To (i) describe the direction and strength of the association between engagement with DHIs and physical activity or sedentary behavior in adults, and (ii) explore whether the direction of association between engagement and physical activity or sedentary behaviour varies by the type of engagement with the DHI (i.e. subjective experience, activities completed, time and logins).

Methods:

Databases EMBASE, MEDLINE, PsycINFO and Scopus were searched from inception to December 2019. Grey literature and reference lists of key systematic reviews and journals were also searched. Studies were eligible if they examined a quantitative association between a measure of engagement with a DHI targeting physical activity and a measure of physical activity and/or sedentary behavior in adults (aged ≥18 years). We considered all quantitative study designs. Studies that purposely sampled or recruited individuals on the basis of pre-existing health-related conditions were excluded. In addition, studies were excluded if the individual engaging with the DHI was not the target of the physical activity intervention, the study had a non-DHI component, or where DHIs targeted multiple health behaviors. A random effects meta-analysis and direction of association vote counting (for studies not included in meta-analysis) were used to address objective 1. Objective 2 used vote-counting on the direction of the association.

Results:

Overall, 10,653 unique citations were identified and 375 full texts were reviewed. Of these, 19 studies (26 associations) were included in the review, with no studies reporting a measure of sedentary behaviour. Meta-analysis of 11 studies indicated a small statistically significant positive association between DHI engagement (based on all usage measures) and physical activity (0.08 [95% CI 0.01, 0.14]). Heterogeneity was high, with 77% of the variation in point estimates explained by between-study heterogeneity. Vote counting indicated the relationship between physical activity and DHI engagement was consistently positive for three measures: subjective experience measures (2 of 3 associations), activities completed (5 of 8 associations) and logins (6 of 10 associations). However, the direction of associations between physical activity and time based measures of usage were mixed (2 of 5 associations supported the hypothesis, 2 were inconclusive and 1 rejected the hypothesis).

Conclusions:

The findings indicate a weak, but consistent positive association between engagement with a physical activity DHI and physical activity outcomes. No studies targeted sedentary behaviour outcomes. Findings were consistent across most constructs of engagement, although associations were weak. Clinical Trial: CRD42018110657