Using artificial intelligence (AI) in primary care settings could help to prevent harm caused by polypharmacy, by identifying risks before they occur, a new large-scale study has found.

AI was used to analyse over 300,000 patient records to calculate a weighted drug interaction risk score for each individual.

And it was able to identify high anticholinergic drug burdens – considered to be a particular risk for hospital admission due to adverse drug interactions – in identified groups before any symptoms were reported or harm occurred.

Yvette Agyako, lead pharmacist at Operose Health, which conducted the study, said that using AI in the management of complex polypharmacy medication reviews had been ‘amazing’.

‘Without AI, the process of identifying the risk for our patients would have been difficult and very time consuming,’ she said.

‘However, now we can easily identify all patients at risk and systematically review them, prioritising those with the greatest risk.’

Dr Tarek Radwan, GP director at Operose Health, said the study ‘clearly demonstrates the benefits of integrating AI with large scale data for patients, clinicians and the NHS’, and ‘should not be under-estimated’.

He added: ‘It is vitally important that we recognise when harm could be caused by the interaction of a person’s different medicines.

‘The findings of our study show we can use data and AI to quickly identify potential risks across large groups of people and take action before their health is impacted.’

In a paper on the study published in the International Journal of Environmental Research and Public Health in June, researchers said that the AI system was able to automatically identify a small number of high-risk patients for clinical manual appraisal, ‘typically extracting tens from potentially hundreds of thousands of patient records’.

Operose Health manages 66 GP practices across England, providing NHS primary care services for more than 640,000 patients.