Health ministry unveils homegrown AI model built for patients

It aims to address healthcare AI issues, which is mostly trained on the Western population.

Singapore's Ministry of Health has launched Singapore Medical Foundation AI Model (SIMFONI), a national research and development initiative adapting artificial intelligence (AI) foundation models to the country's multi-ethnic population, local disease patterns, and clinical guidelines.

The programme aims to address a gap in existing healthcare AI, which is largely trained on Western population data and can be less accurate in Singapore's clinical context.

SIMFONI is backed by the Ministry of Health through the National Medical Research Council Office and MOH Holdings Pte Ltd, and was jointly developed by Robert Morris, Ngiam Kee Yuan, and Daniel Ting.

Early-phase work targets two clinical areas—decision support for cardiometabolic conditions in primary care, and multimodal AI for diagnosing and managing eye diseases.

Cardiometabolic conditions — diabetes, hypertension, and hyperlipidaemia — represent one of Singapore's largest chronic disease burdens.

Asian populations show distinct disease patterns; diabetes, for instance, manifests earlier and at lower BMIs than in Western populations.

SIMFONI's decision-support tools use a three-layer approach—validated risk prediction models, national clinical guidelines, and foundation models that reason across patient history and clinical context.

The long-term goal is to embed SIMFONI's tools directly into Singapore's Electronic Medical Records system, functioning as a clinical decision-support companion across care settings.