Overcoming Big Data Interoperability Challenges Generated by Ubiquitous Devices Outside Traditional Health Systems: A...

Ubiquitous devices such as smartphones, wearables, and personal gadgets generate large volumes of personalized health data outside of traditional health systems. Despite its abundance, differences in data formats and semantics across hardware, platforms and sectors keep information siloed. As a result, valuable precision health and social insights are lost when training or validating predictive health models. Interoperable data pipelines that operate across non-clinical settings are needed so this data can be responsibly translated into actionable information for personal and public health.

This systematic review protocol aims to evaluate existing methodologies that address the interoperability challenges when integrating multimodal health data from ubiquitous devices into non-clinical health infrastructures. The systematic review seeks to identify primary barriers (technical, semantic, organizational, and regulatory), describe current solutions, and pinpoint gaps that hinder seamless exchange and use of such data.

A comprehensive literature search will be conducted across multiple databases (PubMed, IEEE Xplore, ACM Digital Library, and Web of Science) for peer-reviewed primary studies from 2014 to 2024. Gray literature and conference proceedings are excluded. Eligible studies empirically evaluate a framework, architecture or tool that enables technical, semantic, organizational, and regulatory interoperability for person-generated health data outside clinical settings. Two reviewers will conduct title and abstract screening and full-text screening in Covidence using predefined criteria, with blinding of authors, journal, and year. Disagreements will be resolved by consensus or a third reviewer. Data will be extracted in Microsoft Excel, and methodological quality will be assessed with the Mixed Methods Appraisal Tool (2018). Interoperability methodology synthesis will be primarily narrative with evidence tables. Where comparable quantitative outcomes exist, we will compute effect sizes and consider robust statistically significant findings (p < 0.05).

This systematic review protocol will sufficiently assemble a comprehensive and reproducible evidence base to conduct the review. The completed review will synthesize and critically appraise peer-reviewed approaches to interoperability for ubiquitous device-generated personal health data outside of traditional health systems; identify recurrent barriers and current solutions; and outline good practices to enable cross-sector data usage, thereby increasing the volume of interoperable data available for predictive modeling of personalized health outcomes.