<![CDATA[Multimorbidity represents a critical challenge for primary healthcare systems in low- and middle-income countries (LMICs). In Indonesia, fragmented electronic health record (EHR) infrastructure limits effective chronic disease management. This research protocol presents an end-to-end information systems approach to: (1) design a validated ETL framework for heterogeneous health data; (2) develop and compare machine learning models (Random Forest, XGBoost, LSTM) for predicting multimorbidity risk; (3) identify critical determinants in an Indonesian population; and (4) evaluate a Clinical Decision Support System (CDSS) prototype. A mixed-methods, three-phase design will analyze 150,000 chronic disease patients from SIMPUS EHR data (2021-2025). Phase 2 focuses on CDSS development using explainable AI (XAI), while Phase 3 evaluates user acceptance using the Technology Acceptance Model (TAM). The study expects to produce a predictive model with AUC-ROC $\ge0.75$ and an operational CDSS prototype integrated with the Satu Sehat platform. This protocol addresses gaps in Southeast Asian LMIC data, implementation, and interpretability.]]>
Copyrights © 2026
