Artificial Intelligence (AI) enabled clinical decision support in resource constrained settings

Advancements in machine learning (ML) have enabled development of a vast array of models to support clinical decision-making. However, few systems have been deployed in real-world clinical settings. Resource-limited settings provide a fertile ground to harness the benefits of ML-based decision support systems. These settings are characterized by a lower physician/specialist to population ratio, as well as lacking in medical expertise leading to high referral rates.

The goals of this project are to investigate the feasibility of using ML based clinical decision support in resource-limited settings. We are focussing on risk prediction in cardiovascular disease which is the leading cause of morbidity and mortality in developing countries. Many ML models are available to predict the risk of the top three diseases in cardiovascular medicine, namely heart failure, acute coronary syndromes, and atrial fibrillation. However, no previous study has investigated their deployment and use in resource-limited settings.