The tragic death of a nine-year-old boy which lead to an AI tool for asthma patients

In 2015 a nine-year-old boy called Michael Uriely from North West London died from a chronic asthma attack.  

 It was a tragic, and avoidable, loss of life and one which as a GP made me wonder what we could do to help avoid such cases happening again. 

 It was one of the cases that led my colleague Sukin Natarajan and Jay Verma to set up SmartLife Health to address the problem of poor data analysis in primary care.  

According to analysis from Asthma + Lung UK, the UK has the worst death rates for lung conditions such as asthma and COPD. Over a seven-year period, half a million people have died from such conditions.[i] 

Asthma is a chronic condition and frequent reviews are essential to provide patients with rigorous monitoring. But to allow that to happen, a patient’s record needs to have the right information recorded and that means finding the ‘missing asthmatics’ from our patient lists to ensure they get the right care.  

In Primary Care, we have to navigate a balancing act of patient access, support, choice and costs against funding, ensuring we end up with the best possible patient outcomes.  

Thanks to funding from The Health Foundation, SmartLife Health was able to investigate the best ways to identify patients who were potential asthmatics and allow them to be diagnosed by doctors.  

We estimated it would take a GP 15.6 hours to manually review patient records to identify 100 asthma patients. 

But by analysing some 9000 patient records from two GP practices using various machine learning techniques SmartLife Health identified a way to cut this time by 4.9 hours – time which could be spent on the myriad of other demands on a GPs time.