When am I going to die? AI may be able to tell you, research suggests

According to a new study conducted by the University of Nottingham, Artificial Intelligence (AI) can accurately predict when someone will die. Led by Dr Stephen Weng, an assistant professor of epidemiology and data science, scientists have been using AI in the medical field with patients suffering from chronic diseases. This AI tool uses data collected from more than 5 lakh people and predicts who is at the risk of early death.

The data collected on patients consists of their family history as well as their salt consumption. It also collects data on their medication usage. Researchers claim that the AI tool was 'very accurate' in its tests. In fact, it is said to be 10 per cent more reliable than existing machine-learning (ML) predictions.

“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,” said Dr Weng.

READ | Apart from calling and texting, smartphones now diagnose medical conditions

Researchers explain that most of the existing applications concentrate on one area, which is not enough for predicting someone’s death. Plus, taking into account many different disease outcomes is extremely complicated. As a result, scientists went to develop a new approach to predict an individual’s early death using Artificial Intelligence (AI).

The tool created by researchers is based on data collected from 502,628 people between the age of 40 and 69. These people participated in UK Biobank study between 2006 and 2010 and were followed until 2016. The tool also uses 60 health predictors, such as BMI, blood pressure and vitamin or supplement use. Some other data collected on participants include their consumption of vegetable, meat, cheese, cereal, fish, fruit and alcohol.

“We found machine-learned algorithms were significantly more accurate in predicting death than the standard prediction models developed by a human expert,” Dr Weng added.

READ | Scientists develop graphene-based biosensor that could diagnose lung cancer early

According to scientists and researchers, Artificial Intelligence will play a significant role when it comes to developing personalised medicine. However, more research is needed into machine learning's efficacy in other population groups and how to bring AI into more common healthcare.