Improving predictability and effectiveness in preventive digital health interventions: A scoping review. (Preprint)

Abstract

This research studies preventive digital health interventions (P-DHIs) using mobile apps to improve diet and physical activity (PA) behaviors. RCT studies indicate that P-DHIs can prevent health problems for some citizens in some situations, but the results from these studies are mixed. Adoption studies have identified multiple factors related to individual citizens and the context they live in that complicate the transfer of positive results from RCT studies to practical use because different citizens may respond differently to the same P-DHI. Implementation studies have revealed barriers to large-scale implementation of mHealth solutions in general. Consequently, it is not clear how to deliver predictable outcomes from P-DHIs, and it is not clear how to achieve effectiveness when scaling up interventions to reduce health problems in society in general.

This research attempts to expand our understanding of how to increase the predictability of outcomes from P-DHIs focusing on PA and diet behaviors and amplify our understanding of how to improve effectiveness from large-scale implementations.

The objective of this research is pursued through a multidisciplinary scoping review. The literature review was conducted using Web of Science and Google Scholar, reflecting the multidisciplinary nature of P-DHIs. Theoretical models from Information Systems and Public Health and quantitative and qualitative mHealth studies were included in the literature review. The first step was to use Information Systems theory to identify key constructs influencing outcomes from IT in general. The second step was to identify factors from the Public Health and mHealth literature influencing the outcome specifically from P-DHIs and influencing the possibilities for successful large-scale mHealth implementation. In the third step, the factors from the Public Health and mHealth literature were categorized using the constructs from Information Systems theory. Finally, the P-DHI investment model was developed based on the Information Systems constructs and factors from the Public Health and mHealth literature.

196 articles met the inclusion criteria. The included articles use a variety of methodologies, including literature reviews, interviews, surveys, and RCT studies. The P-DHI investment model suggests which constructs and related factors to focus on to increase the predictability of P-DHI outcomes and to improve effectiveness from large-scale implementations. The key constructs are: The P-DHI itself, which includes IT and non-IT investments, the context in which P-DHIs are implemented and used, the processes that are changed by the P-DHI, and the lag effects which influence when outcomes from P-DHIs are realized. The model describes the interdependence between the constructs and the difficulties in understanding and controlling the many factors related to them.

The research suggests that outcome predictability could be improved by including descriptions of the constructs and factors in the P-DHI investment model when reporting from empirical studies. Doing so would increase our understanding of when and why they succeed or fail. Effectiveness from large-scale implementations might be improved by using the P-DHI investment model to evaluate the difficulties and possibilities in implementing P-DHIs to create better environments for the use of P-DHIs before investing in them and when designing the P-DHIs. The cost-effectiveness of large-scale implementations is unknown, implementations are far more complicated than just downloading and using apps, and there is uncertainty accompanying implementations given the lack of coordinated control over the many constructs and factors that influence the outcome.