Publication: PRECISION HEALTH THROUGH WEARABLE TECHNOLOGY: K-MEANS CLUSTERING FOR CULTURALLY ADAPTED NCD PREVENTION
0
0
Issued Date
2025-01-01
Resource Type
ISSN
24080071
eISSN
24079529
Scopus ID
2-s2.0-105027282075
Journal Title
Scientific Culture
Volume
11
Issue
4
Start Page
100
End Page
111
Rights Holder(s)
SCOPUS
Bibliographic Citation
Scientific Culture Vol.11 No.4 (2025) , 100-111
Suggested Citation
Pochai N., Klabchom K., Siriussawakul A., Pariyavuth P., Panurushthanon P., Al-Umaree P., Chottidao M., Punthipayanon S. PRECISION HEALTH THROUGH WEARABLE TECHNOLOGY: K-MEANS CLUSTERING FOR CULTURALLY ADAPTED NCD PREVENTION. Scientific Culture Vol.11 No.4 (2025) , 100-111. 111. doi:10.5281/zenodo.11042509 Retrieved from: https://hdl.handle.net/20.500.14740/55096
Corresponding Author(s)
Other Contributor(s)
Abstract
Introduction: Non-communicable diseases (NCDs) represent a critical global health challenge, accounting for 71% of deaths worldwide, with disproportionate burden in underserved populations including religious communities. Thai Buddhist monks face exceptionally high NCD prevalence attributed to sedentary lifestyles, dietary constraints, and limited physical activity opportunities inherent to monastic practices. This study introduces a novel machine learning framework for culturally-adapted health monitoring, addressing the urgent need for scalable, technology-driven solutions in traditional religious communities globally. Methods: This cross-sectional study employed an innovative K-means clustering approach to analyze wearable device data from 28 Thai Buddhist monks over one month (November-December 2024). Polar Pacer Pro devices captured daily step counts, average heart rate (beats per minute), and energy expenditure (kilocalories). Following Min-Max normalization, unsupervised K-means clustering identified distinct physical activity phenotypes. Optimal cluster determination utilized the Elbow Method through Within-Cluster Sum of Squares (WCSS) analysis. This represents the first application of unsupervised machine learning for health pattern recognition in monastic populations, demonstrating methodological innovation in small-sample clustering validation for culturally-specific healthcare contexts. Results: K-means clustering successfully identified distinct activity profiles within the monastic population, revealing significant heterogeneity in physical activity patterns. The analysis differentiated monks into meaningful clusters based on step count (range: 700-18,650 daily steps), heart rate (50-87 bpm), and energy consumption (1,623-3,758 kcal) profiles. Cluster centroids demonstrated clear stratification: low-activity groups (1,079-3,763 steps daily) representing 39% of participants with sedentary behavior patterns, moderate-activity clusters (4,495-5,164 steps), and high-activity groups (9,381-12,664 steps) approaching recommended cardiovascular health guidelines. These quantitative classifications provide empirical foundations for precision health interventions tailored to individual risk profiles. Discussion: The research evaluates scalable digital health system that holds great potential for implementation in many religious and cultural groups globally. The clustering approach was able to identify actionable health phenotypes that inform targeted NCD prevention strategies that do not violate cultural restrictions. Low-activity clusters could be seen as the straightforward target of intervention, and the higher-activity cluster indicates that the promotion of physical activity practices is successful in the traditional approach. The methodology is a potentially effective method that is feasible in resource-restricted environments, about consumer-level wearable technology and open-source machine learning algorithms. Strategies of cultural adaptation, such as the application of Buddhist walking meditation and mindful movements in health promotion, serve as sustainable avenues of health promotion at the community level. This framework offers a replicable template related to health disparities in religious community members around the world (estimated 500+ million) and can lead to Sustainable Development Goal 3 (Good Health and Well-being) via innovative and culturally relevant digital health technologies. These results can guide policymaking based on evidence related to community health initiatives and illustrate how the precision health vision can empower traditional in the development of precision health to address modern NCDs in contemporary society.