Digital health and machine learning in population health and wellness management: A scoping and bibliometric review of effectiveness, equity, and implementation challenges

Machine learning (ML) in digital health applications is becoming more popular for the general management of population wellness and the promotion of large-scale prevention, risk stratification, and the use of data to formulate data-driven decision-making in the field of population health. Although there has been a rapid increase in this area, the available reviews are highly fragmented and frequently tend to concentrate on particular technologies or predictive performance, yet provide minimal synthesis of the effectiveness, equity, implementation, and governance at the population-level. This review followed the Preferred Reporting Items for Systematic Reviews and Scoping Reviews (PRISMA-ScR) guidelines and employed a hybrid scoping and bibliometric approach. of the peer-reviewed literature published between 2018 and 2025 and indexed in PubMed/MEDLINE, Scopus, Web of Science, and IEEE Xplore. The review of the literature was in the form of scoping and concerned the digital form of healthcare, roles of ML analytic, the domains of outcome, and bibliometric analysis was performed to map the trends in publications, thematic groups, and evolving research areas. The results show significant increases in interdisciplinary studies in the field of public health, medical informatics, and data science in health. Most studies reported predictive-performance and short-term behavioral outcomes, whereas evidence on long-term population-level health effects and healthcare utilization remained limited. Nevertheless, there is little evidence regarding long-term population-level health effects and changes in healthcare use. Equity assessments have rarely conducted, and there have been repeated debates pertaining to data representativeness and algorithmic bias. Governance and implementation issues (such as model interpretability, privacy, data sharing, and regulatory uncertainty) have been continuously found to be obstacles to wide-scale and responsible deployment. All in all, this review offers a combined evaluation of efficacy, equity, and governance in ML-enabled, digital health to manage population wellness. The conclusions also show that there is a need to shift away from technology-focused assessments in favor of outcome-driven, equity-based, and governance-informed ways of encouraging sustainable and responsible population-level action.