Key points are not available for this paper at this time.
Background Physical inactivity remains a persistent global health challenge despite long-standing evidence that regular physical activity (PA) reduces chronic disease risk, cognitive decline and premature mortality. In parallel, digital health technologies have expanded rapidly, yet it remains unclear how distinct platform types have emerged, diffused and been differentially adopted in clinical vs. non-clinical populations. Methods We conducted a large-scale bibliometric trend analysis of technology-supported PA interventions indexed in Scopus and SportDiscus, covering records published from 1953 to 2025. Using an active-learning screening workflow with title/abstract screening, sensitivity checks and consensus adjudication, we identified 2,981 eligible studies published between 1988 and 2025 across 757 journals and 63 countries. Studies were coded by population (clinical vs. non-clinical), platform cluster, and author-reported PA outcome direction abstracted from the publication record. Results Publications increased markedly after 2008, with smartphone/mHealth and wearable sensors becoming the dominant platform clusters in the 2010s and early 2020s. In the smoothed overall trajectories, publication activity reached its highest level around 2022, followed by a contraction concentrated mainly in mature clinical platform clusters. Non-clinical studies generally adopted newer platforms earlier, whereas clinical studies showed a recurrent lag before converging for more established, accessible technologies. A distinct population-level reversal was visible in mature platforms: non-clinical studies led early smartphone, wearable and web-based uptake, clinical studies became more prominent around the COVID-19 period, and non-clinical studies again predominated by 2025 for smartphone/mHealth, wearable sensors and web applications. Multi-component designs were common, with the strongest network backbone centred on smartphone, wearable and web-based combinations. Reported PA improvement signals were common across both population strata, but these findings reflect patterns in published study reporting rather than comparative effectiveness. Conclusions By providing a platform-centred, cross-population and temporally resolved map of technology-enabled PA interventions, this study identifies mature technology backbones, emerging areas of experimentation, and recurrent translational gaps between clinical and non-clinical contexts. The findings support more theory-informed and implementation-aware intervention design while underscoring that bibliometric prominence should not be equated with real-world efficacy.
Baumann et al. (Fri,) studied this question.