JMIR Publications

ABSTRACT

Background:

The number of people suffering from non-communicable diseases (NCDs) is rapidly increasing. NCDs are the major cause of disability and premature mortality worldwide, and are associated with negative workplace outcomes, such as sickness absence and reduced work productivity. There is a need to identify scalable interventions and their active components to relieve disease and treatment burden and facilitate work participation. eHealth interventions have shown potential in clinical and general populations when it comes to increasing well-being and physical activity and could be well suited for workplace settings.

Objective:

The aim of this review was to provide an overview of the effectiveness of eHealth interventions at the workplace targeting health behaviors of employees, and to map behavior change techniques used in these interventions.

Methods:

A systematic search of literature was performed in PubMed, EMBASE, PsychINFO, Cochrane CENTRAL, and CINAHL databases in September 2020 and updated in September 2021. Extracted data included participant characteristics, setting, type of eHealth intervention, mode of delivery, reported outcomes, effect sizes, attrition rates, etc. Quality assessment and risk of bias of the included studies was assessed with the Risk of Bias 2 tool by Cochrane Collaboration. Behavior change techniques (BCTs) were mapped in accordance with the Behavior change techniques taxonomy v1. The review was reported according to PRISMA checklist.

Results:

Seventeen randomized controlled trials fulfilled the inclusion criteria. There was a high level of heterogeneity across the measured outcomes, treatment and follow up periods, content of eHealth interventions and workplace contexts. Out of 17 studies 24% reported unequivocally significant findings for all primary outcomes, with effect sizes ranging from small to large. Furthermore, 53% of studies reported mixed results and 24% reported non-significant results. The most frequently targeted behavior was physical activity (88% of the studies), the least frequently targeted behavior was smoking (12% of the studies). Attrition varied greatly across the studies (0-37%). Risk of bias was high in 65% of the studies and with some concerns for the rest 35%. Interventions employed various BCTs, and the most often used were “feedback and monitoring” (82%), “goals and planning” (59%), “antecedents” (59%) and “social support” (41%).

Conclusions:

Evidence from this review suggests that although eHealth interventions may have potential, there are still many unanswered questions regarding their effectiveness and what drives the mechanism behind the effects. Low methodological quality, high heterogeneity and complexity, characteristics of included samples and often high attrition rates challenge the investigation of the effectiveness and making sound inferences about the effect sizes and significance of the results. To address this, new studies and methods are needed. A megastudy design where different interventions are evaluated in the same population, over the same period, on the same outcomes may solve some of the challenges. Clinical Trial: PROSPERO (CRD42020202777)