Artificial Intelligence can Predict Premature Death

Researchers have successfully tested a new system that can predict the risk of premature death in the middle-aged population.

According to the latest study at the University of Nottingham, ‘Machine Learning’ algorithms can prove extremely helpful in improving preventative healthcare in the future. This deduction was made on the basis of a computer-based system that can successfully predict the chances of premature death, due to chronic disease. The researching team used the health data of more than a million people (aged between 40 and 69) who were recruited to the UK Biobank between 2006 and 2010.

Researchers found that this newly-developed system is extremely accurate in its predictions of premature death as it performed much better than the current standard technique. Dr. Stephen Weng, an Assistant Professor of Epidemiology and Data Science at the University of Nottingham and the Lead Researcher of the study, highlighted the significance of this achievement and mentioned that it can prove very useful in the field of preventative healthcare. He 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 computerized health risk assessment in the general population. Most applications focus on a single disease area but predicting death due to several different disease outcomes is highly complex, especially given environmental and individual factors that may affect them.”  

Predicting Premature Death by Machine Learning

Weng described this system as a major step forward in the field of preventative medicine. He stated that it is a holistic and unique approach to use machine learning algorithms to determine the risks of premature death in an individual. He elaborated that the computer models they developed take a variety of factors into account including clinical, demographic, lifestyle, and biometric routines of the person. In addition to all these assessments, the dietary composition of that individual is also monitored to ensure accurate results. He compared this new technique with the standard prediction models in the following words:

“We mapped the resulting predictions to mortality data from the cohort, using Office of National Statistics death records, the UK cancer registry and ‘hospital episodes’ statistics. We found machine-learned algorithms were significantly more accurate in predicting death than the standard prediction models developed by a human expert.”

Machine Learning Models

Previous research of the team showed that 4 algorithms of Artificial Intelligence (Neural Networks, Logistic Regression, Random Forest, and Gradient Boosting) are much better at predicting cardiovascular disease in an individual than the standard models. This urged the researching team to develop an improved system that could accurately predict the risks of premature death in the middle-aged population. Consequently, they used a combination of Random Forest and Deep Learning to create a modified model which significantly enhanced the accuracy of predictions.

The newly-developed system was compared with both, ‘Cox Regression’ and ‘Multivariate Cox’ models. Cox Regression is the least accurate model at predicting mortality while Multivariate Cox tends to over-predict the risk of premature death. Joe Kai, a Professor at the University of Nottingham who was also a part of the study, praised the system by saying,

“There is currently intense interest in the potential to use ‘AI’ or ‘machine learning’ to better predict health outcomes. In some situations, we may find it helps, in others it may not. In this particular case, we have shown that with careful tuning, these algorithms can usefully improve prediction. These techniques can be new to many in health research, and difficult to follow. We believe that by clearly reporting these methods in a transparent way, this could help with scientific verification and future development of this exciting field for health care.”

Role of Artificial Intelligence in Future Tools

Researchers believe that AIis extraordinarily important for a lot of future tools capable of delivering personalized medicine. They mentioned that techniques like machine learning are ideal for tailoring risk management for different patients, each with individual healthcare data. They are hopeful that they will be able to use these computer-based systems to improve the overall efficiency of the medical world. As far as these algorithms are concerned, they will need to verify and validate them in other populations before implementing them into routine healthcare.