Artificial intelligence is being used to find people with heart conditions before they even have symptoms.

In an ongoing trial, a ground-breaking tool scours GP records for “red flags” which could indicate whether a patient was at risk of developing atrial fibrillation (AF).

John Pengelly, a former Army captain, said he was “really grateful” that the algorithm had detected his AF risk. He now takes a “couple of pills a day” to reduce his heightened chance of a potentially deadly stroke.

AF causes an irregular and often abnormally fast heart rate, and people with it have a significantly higher risk of stroke.

Some sufferers experience heart palpitations, dizziness, shortness of breath and tiredness – but others have no symptoms and are unaware their heart rate is irregular.

Around 1.6 million people in the UK have been diagnosed with AF but the British Heart Foundation (BHF) said there were probably many thousands of undiagnosed cases in the UK.

When AF is identified and treated early it can be managed and the stroke risk reduced.

The new AI tool is being assessed in a trial, called Find-AF, funded by BHF and the Leeds Hospitals Charity.

Its algorithm was developed by scientists and clinicians at the University of Leeds and Leeds Teaching Hospitals NHS Trust, with funding from the BHF.

They trained the AI tool using anonymized electronic health records of over 2.1 million people, training the algorithm to find warning signs that could indicate a person is at risk of developing AF.

The tool was validated with medical records from a further 10 million people and is now examining GP records at several surgeries in West Yorkshire.

Experts are assessing how effective it is at finding people who are at risk of developing AF in the next six months, with those identified then offered further testing.

The software works out someone’s risk based on a number of factors including age, sex, ethnicity and whether or not they have other medical conditions including heart failure, high blood pressure, diabetes, ischaemic heart disease and chronic obstructive pulmonary disease.