This site is not optimized for Internet Explorer 8 (or older).
Please upgrade to a newer version of Internet Explorer or use an alternate browser such as Chrome or Firefox.
Prediction of Premature All-Cause Mortality: A Prospective General Population Cohort Study Comparing Machine-Learning and Standard Epidemiological Approaches
Researchers from the University of Nottingham studied 502,628 adults aged 40 to 69 years whose health information was logged in the UK Biobank between 2006 and 2010. Using demographic data and taking into account biometric, clinical, and lifestyle factors, the authors developed predictive mortality models using deep learning, random forest, and Cox regression.
Death occurred in 14,418 adults (2.9%) over a total follow-up time of 3,508,454 person-years, and mortality data was corroborated with national records. The age- and gender-based Cox model was the least predictive, with an area under the curve (AUC) of 0.689, followed by the multivariate Cox regression model, which improved discrimination by 6.2% for an AUC of 0.751. The application of machine-learning algorithms further improved discrimination by 3.2% using the random forest model (AUC = 0.783; 95% confidence interval (CI), 0.776 - 0.791) and 3.9% using deep learning model (AUC = 0.790; 95% CI, 0.783 - 0.797). The two machine-learning algorithms improved discrimination by 9.4% and 10.1% respectively from a simple age and gender Cox regression model.
This work suggest machine learning significantly improved accuracy of prediction of premature all-cause mortality in this middle-aged population, compared to standard methods. This study illustrates the value of machine learning for risk prediction within a traditional epidemiological study design.