Chronic diseases threaten the lives of a large portion of the world’s population. Among Americans, 60% have one or more chronic diseases such as heart disease, cancer, stroke, and diabetes, according to the CDC. The CDC works with communities to promote health and wellness by educating people about preventing chronic diseases through nutrition, physical activity, avoiding tobacco and drinking too much alcohol, and getting regular medical checkups to screen for illness. We have written previously about the role of wellness programs in discovering chronic disease.

Healthcare scientists and doctors from the University of Nottingham developed an artificial intelligence machine-learning algorithm that accurately predicts premature death in middle-aged people with chronic disease. In a study published in Machine Learning in Health and Biomedicine, the Nottingham team presented the results of its algorithm used with more than one half million people aged 40 to 69 with chronic diseases. The subjects were first surveyed from 2006 to 2010 and followed up through 2016. According to Dr. Stephen Weng, Assistant Professor of Epidemiology and Data Science, the initial surveys included demographic, biometric, clinical, and lifestyle factors for all participants. The data granularity for nutrition, for example, included daily consumption of meat, fruit, and vegetables.

The follow-up study showed that, based on national death records, the U.K. cancer registry, and hospital records, the Nottingham AI predictive model was significantly more accurate than traditional models that rely on age and gender and other multivariate models. The Nottingham team continues to develop new AI models. The current study followed earlier work with a neural network-based model that bested conventional guidelines at predicting cardiovascular disease.