SGH and NUS develop algorithm-driven tests for mortality risk indication

The PCAge tool adopts an analytical matric factorisation approach to simplify complex data.

The Singapore General Hospital (SGH) and the National University of Singapore's Yong Loo Lin School of Medicine (NUS Medicine) have jointly developed algorithm-powered tests to indicate mortality risks for ageing intervention strategies.

PCAge, a machine learning algorithm, uses an analytical matrix factorisation technique to reduce complex high-dimensional data into lower dimensions.

The team utilised publicly available data from the National Health and Nutrition Examination Survey (NHANES), comprising over 3,000 participants aged between 40 and 89 years.

This method provides data to identify the underlying causes of co-morbidities based on an informative representation of a patient’s status.

Furthermore, the team has generated a streamlined clinical ageing clock, LinAge, based directly on the developed tool, maintaining equivalent predictive power but requiring fewer parameters. 

To date, LinAge has been applied to 40 Singaporeans, aged 65 to 95, in SGH’s geriatric clinic and can be determined from routine clinical blood examinations, a urine test, and a health questionnaire.