A new artificial intelligence tool could transform the way heart conditions are detected, identifying patients at risk before they even exhibit symptoms.
The groundbreaking system uses millions of anonymised health records to identify red flags associated with atrial fibrillation (AF).
AF is a heart condition characterised by an irregular and often rapid heartbeat, which can increase the risk of stroke. People who have it may experience palpitations, dizziness, shortness of breath and fatigue, but in many cases, the condition does not present any symptoms at all.
Currently, about 1.6 million people in the UK have been diagnosed with AF, but the British Heart Foundation (BHF) believes many more remain undiagnosed and unaware of their elevated stroke risk.
The tool, FIND-AF, was developed by scientists and clinicians at the University of Leeds and Leeds Teaching Hospitals NHS Trust, with funding from the BHF.
It uses machine learning to examine a patient’s medical records and assess the likelihood they will develop AF based on factors including age, sex, ethnicity, and whether they have other medical conditions including heart failure, high blood pressure, diabetes, ischaemic heart disease and chronic obstructive pulmonary disease.
The algorithm is trained on 2.1 million anonymised health records and validated with a further 10 million health records.
Patients identified as high risk are given a handheld electrocardiogram (ECG) device to use at home twice a day for four weeks, or anytime they feel symptoms like palpitations. This data is sent directly to the trial team for analysis.
If the ECG readings suggest irregular heart rhythms indicative of AF, the tool triggers an alert to the patient’s GP, who can then discuss treatment options. These may include medication to reduce the risk of stroke, or lifestyle changes to improve heart health.
John Pengelly, a retired captain in the Army Catering Corps who participated in the trial had no noticeable symptoms of AF, but the AI tool flagged him as being at risk.
The 74-year-old from Apperley Bridge, Bradford, West Yorkshire, said he got a letter inviting him to take part in the study and thought: “If it benefits somebody then great, I want to help.”
“They sent me a little digital monitor and a few times a day I had to put my thumbs on it so it could take a reading, which took about two minutes. Then I pressed send and the reading went to the trial team. I did that for a few weeks, and I sent the kit back – it was really straightforward.
“I was diagnosed with AF a few weeks after that. I’d heard of it, but you never think that these things will happen to you. I didn’t have any symptoms. I’d occasionally get a bit breathless when I’m out and about, but that’s because there are so many hills around us and some of them are really steep.
“I’m really grateful it has been picked up. I now take a couple of pills every day to reduce my risk of having a stroke. It’s just a few pills every day that will hopefully keep me going for a good few more years yet.”
What causes atrial fibrillation?
When the heart beats normally, its muscular walls tighten and squeeze (contract) to force blood out and around the body.
They then relax so the heart can fill with blood again. This process is repeated every time the heart beats.
In atrial fibrillation (AF), the heart’s upper chambers (atria) contract randomly and sometimes so fast that the heart muscle cannot relax properly between contractions. This reduces the heart’s efficiency and performance.
AF happens when abnormal electrical impulses suddenly start firing in the atria. These impulses override the heart’s natural pacemaker, which can no longer control the rhythm of the heart. This causes you to have a highly irregular pulse rate.
The cause is not fully understood, but it tends to affect certain groups of people, such as older people and people living with long-term (chronic) conditions such as heart disease, high blood pressure or obesity. It may be triggered by certain situations, such as drinking too much alcohol or smoking.
AF can be defined in various ways, depending on the degree to which it affects you. For example:
- Paroxysmal atrial fibrillation – episodes come and go, and usually stop within 48 hours without any treatment;
- Persistent atrial fibrillation – each episode lasts for longer than seven days (or less when it is treated);
- Permanent atrial fibrillation – when it is present all the time;
- Longstanding atrial fibrillation – when atrial fibrillation has been experienced – usually – for more than a year.
Estimates suggest that AF is a contributing factor in around 20,000 strokes every year in the UK.
Chris Gale, professor of cardiovascular medicine at the University of Leeds and honorary consultant cardiologist at Leeds Teaching Hospitals NHS Trust, said: “All too often the first sign that someone is living with undiagnosed atrial fibrillation is a stroke.
“This can be devastating for patients and their families, changing their lives in an instant.
“It also has major cost implications for health and social care services – costs which could have been avoided if the condition were spotted and treated earlier.”
Dr Sonya Babu-Narayan, associate medical director at the British Heart Foundation and consultant cardiologist at Royal Brompton Hospital, said: “We have effective treatments for people with atrial fibrillation who are at high risk of having a stroke.
“But right now some people are missing out because they don’t know that they may be living with this hidden threat to their health.
“By harnessing the power of routinely collected health care data and prediction algorithms, this research offers a real opportunity to identify more people who are at risk of atrial fibrillation and who may benefit from treatment to reduce their risk of a devastating stroke.”
Dr Ramesh Nadarajah, from Leeds Teaching Hospitals NHS Trust, said: “Data are collected about patients in every interaction they have with the NHS.
“These data have huge potential to make early identification of and testing for conditions like AF easier and more efficient.
“If it’s successful, this study will be the launchpad for a larger nationwide trial to determine whether our algorithm could become part of everyday clinical practice.
“Ultimately, we hope that this approach will lead to an increase in the number of people diagnosed with AF at an early stage who get the treatment they need to reduce their risk of stroke.”
Earlier in December, Health Secretary Wes Streeting highlighted how using AI and big data would be “game-changing” for healthcare.
Talking to MPs on the Health and Social Care Committee, he suggested that “we can use AI, machine learning, genomics, big data, to not only intervene early with earlier diagnosis and earlier treatment, but to actually predict and prevent illness, which is the game-changing paradigm shift in healthcare in this century.”
If you are affected by heart disease, you can contact the British Heart Foundation for emotional support and for advice from a healthcare professional.
Additional reporting from the wires.
