Application of machine learning in recognising and treating asthma Heathrow Medical Centre

Asthma is one of the most prevalent diseases in the UK, with approximately 5.4 million people affected being on prescribed medication. Poor asthma management is responsible for a large proportion of preventable deaths in the UK.

A lack of consistency in the coding of GP records is largely responsible for many patients missing out on regular asthma reviews. GP practices within the Long Lane and First Care Group Primary Care Network (PCN) have found that medical record inconsistencies can be responsible for under-reporting of practice prevalence by as much as 1–2% of the whole patient population.

In a search of the records of Heathrow Medical Centre and Shakespeare Health Centre, 171 patients with possible asthma were identified. Of these, 61% were found to be asthmatic based on national guidelines. These patients had not previously been correctly coded or formally diagnosed.

This project will involve applying advanced analytical techniques based on machine learning to automate this time-consuming checking process and quickly find inconsistencies in records at the point of care. It will run analysis on clinical codes, free text and documents. This has the potential to outperform the manual process.

This artificial intelligence (AI) algorithm support tool will help identify patients who have suspected asthma, and illustrate the best asthma treatment option(s) available for patients with asthma. The aim is to use this tool across the 50,000 patients within the PCN.

A similar methodology could also be used to deal with other long-term conditions, such as diabetes.

Contact

For more information about this project, please contact Sukin Natarajan, Programme Director, Heathrow Medical Centre.