Towards Data-Driven Hydration Monitoring: Insights from Wearable Sensors and Advanced Machine Learning Techniques

Apparaju Sreeharsha, Sarah McHale, Nonso Nnamoko, Ella Pereira

Advancements in wearable sensors and digital technologies/computational tools (e.g., machine learning (ML), general data analytics, mobile and desktop applications) have been explored in existing studies. However, challenges related to sensor efficacy and the application of digital technology/computational approaches for hydration assessment remain under-explored. Key knowledge gaps include applicable devices and sensors for measuring hydration and/or dehydration, the performance of approaches (e.g., ML algorithms) on sensor-based hydration monitoring; the potential of multi-sensor fusion to enhance measurement accuracy and the limitations posed by experimental datasets. This review aims to address the gaps by examining existing research to provide recommendations for future improvements. A systematic review was conducted following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Comprehensive searches across PubMed, Scopus, IEEE Xplore and MDPI databases for academic studies published between 2009 and 2024 were performed using predefined inclusion and exclusion criteria. Two reviewers independently screened and assessed studies, with disagreements resolved by a third reviewer. Data was synthesised narratively or through meta-analysis, where applicable. The database search yielded 1029 articles, with 999 unique studies remaining after duplicate removal. After title and abstract screening, 910 irrelevant studies were excluded. Full-text evaluation of 89 articles led to the inclusion of 20 studies for in-depth analysis. Findings highlight significant progress in hydration monitoring through multi-sensor fusion and advanced ML techniques, which improve accuracy and utility. However, challenges persist, including model complexity, sensor variability under different conditions, and a lack of diverse and representative datasets. This review underscores the need for further research to overcome these challenges and support the development of robust, data-driven hydration monitoring solutions.