AI-enabled Clinical Decision Support System (CDSS) for Common Non-communicable Diseases

In collaboration with AstraZeneca Vietnam, this project introduces a rule-based Clinical Decision Support System (CDSS) for hypertension management. Clinical guidelines contain essential knowledge for treatment decisions, but their complexity makes them difficult to apply consistently in routine practice.

The system encodes the VSH/VNHA 2022 hypertension guidelines as computable knowledge: 14 decision trees and 368 nodes covering initial visits, follow-up encounters, and comorbidity pathways including pregnancy, heart failure, chronic kidney disease, diabetes, and resistant hypertension. Patient data is evaluated through deterministic traversal and converted into actionable recommendations.

At the core is a transparent rule engine supported by an expert-reviewed knowledge base, where every recommendation traces back to its guideline source. Because the knowledge is stored as data rather than code, guideline updates require no engineering work. The system is a full-stack application: a frontend for entering, testing, and visualizing patient scenarios, and a backend API handling decision-tree traversal, recommendation generation, and medication regimen support, with FHIR R4 export for hospital integration.

By turning complex guidelines into explainable decision support, the project improves the consistency, traceability, and efficiency of hypertension management. Rather than replacing clinical judgment, the CDSS helps healthcare professionals understand what is recommended and why, supporting more transparent decisions.