Feasibility and Behavioral Impact of an AI-Enabled Workplace Health Screening Machine in a Low-Resource Urban Setting: A Pilot Implementation Study in the Philippines (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Background: Artificial intelligence (AI)-enabled health screening offers new opportunities to detect noncommunicable disease (NCD) risks in resource-limited settings. However, evidence on real-world feasibility, user acceptance, and behavioral outcomes of such tools in low- and middle-income countries (LMICs) remains limited. </sec> <sec> <title>OBJECTIVE</title> Objective: This study evaluated the feasibility, acceptability, and short-term behavioral impact of DigiHealth, an AI-enabled health screening machine deployed among public school teachers in the Southern Philippines. </sec> <sec> <title>METHODS</title> Methods: A cross-sectional implementation study was conducted among 384 teachers who underwent biometric and biochemical screening (BMI, blood pressure, fasting blood sugar, HbA1c, and lipid profile) using DigiHealth. Post-screening surveys measured perceived ease of use, reliability, privacy, and follow-up health behaviors. Quantitative data were analyzed using Welch’s t-test, χ², and logistic regression. The study was guided by the Technology Acceptance Model (TAM) and Health Belief Model (HBM). </sec> <sec> <title>RESULTS</title> Results: Most participants were female (81.2%; median age 44 years). Males showed higher systolic blood pressure (130.7 vs 125.2 mmHg, P=.04) and triglycerides (171 vs 144 mg/dL). Overall, 85% rated DigiHealth as “good” or “excellent,” and 93% found it easy to use. Seventy percent consulted a health professional, and 67% reported lifestyle modification after screening. Age was inversely associated with clustering of ≥3 metabolic risk factors (P=.01). </sec> <sec> <title>CONCLUSIONS</title> Conclusions: AI-assisted workplace screening is feasible, acceptable, and behaviorally activating in low-resource contexts. DigiHealth demonstrates how pragmatic, fit-for-purpose AI innovations can complement national NCD programs and promote early detection in institutional settings within LMICs. </sec> <sec> <title>CLINICALTRIAL</title> NA </sec>

Ähnliche Arbeiten