A new study published in PLoS One suggests that prediction of premature mortality using artificial intelligence (AI) could be a significant step in preventative medicine in the future.
A research team at the University of Nottingham comprising data scientists and physicians have developed computer-based ‘machine learning’ algorithms to predict early mortality risk from chronic disease. The algorithms analysed data from more than half a million middle-aged individuals recruited in the UK Biobank. The predictions made by the AI system were mapped to the mortality data derived from the Office of National Statistics death records, the UK cancer registry and ‘hospital episodes’ statistics.
The findings showed that the AI system was able to make accurate predictions in a better way than the current standard predictive models developed by human experts. The machine learning models termed ‘random forest’ and ‘deep learning’ were pitted against the age and sex-based ‘Cox regression’ model, which had a low accuracy in mortality prediction, and also against a multivariate Cox model, which fared better but was likely to over-predict risk.
Dr, Stephen Weng, the lead author, said: "Preventative healthcare is a growing priority in the fight against serious diseases so we have been working for a number of years to improve the accuracy of computerised health risk assessment in the general population."