Diabetes’ Predictive Scoring Model in Adults at Jakarta Based on Waist-to-Hip Ratio and Handgrip Strength

Abstract

Background: Diabetes mellitus is an escalating public health concern, particularly among productive-age populations in urban settings. This study aimed to evaluate the potential of WHR and HG as low-cost, non-invasive tools for early diabetes risk detection.

Methods: A cross-sectional study was conducted among 539 productive-age adults (>18 years) across multiple districts in Jakarta during 2024–2025 using consecutive sampling. Fasting blood glucose, WHR, and HG were assessed. Multiple linear regression was applied to examine associations between WHR, HG, and fasting glucose. A binary logistic regression model with backward likelihood ratio selection was used to identify predictors of diabetes.

Results: WHR showed a significant positive association with fasting glucose (B = 72.647; p = 0.001), while HG demonstrated a significant inverse association (B = –0.468; p = 0.047). Logistic regression analysis indicated that lower HG (B = –0.867; Exp(B) = 0.420; p = 0.001) and higher WHR (B = –2.413; Exp(B) = 0.090; p = 0.018) were independently associated with diabetes risk. The model constant was 4.535 (p < 0.001).

Conclusion: A model integrating WHR and HG shows promise for early diabetes risk stratification among productive-age adults in Jakarta, supporting their utility in community-based screening.