Skip to content
View commentaries on this research

This is a plain English summary of an original research article. The views expressed are those of the author(s) and reviewer(s) at the time of publication.

An easy-to-use ‘smart’ stethoscope could screen for heart failure in primary care. Research found that the technology, which uses artificial intelligence, could accurately identify people with the disease.

Heart failure means that the heart is too weak to pump sufficient blood around the body. Currently, the condition is usually diagnosed in hospital.

In this study, researchers used a familiar stethoscope (to listen to heart sounds) with additional sensors to record an ECG (an electrocardiogram, a measure of the heart’s electrical activity). They incorporated artificial intelligence (a computer programme) to analyse the ECG reading.

The smart stethoscope identified people with heart failure correctly 9 out of 10 times. Few were missed by the stethoscope, and few were incorrectly identified as having heart failure when they did not.

The researchers say the tool could be used to diagnose heart failure earlier, in primary care. It could help GPs prioritise referrals to secondary care for the people who need specialist treatment.

For more information about heart failure, visit the NHS website.

The issue: can AI help diagnose heart failure?

Heart failure means the heart is too weak to pump sufficient blood around the body to meet its needs. About 1 in 100 adults in Europe have this condition, which is long-term and tends to gradually worsen over time. Having heart failure can severely limit quality of life.

Heart failure is usually diagnosed through an emergency hospital admission. However, many people first have symptoms such as breathlessness, ankle swelling or fatigue. GPs can offer checks and a blood test to see how well the heart is working. But the symptoms could be caused by other conditions, which means referrals to specialists can be delayed.

The sooner people receive treatment, the better their outcomes. Early diagnosis and treatment of heart failure is a goal of the NHS Long-Term Plan.  

ECG machines use electrodes to monitor the heart's rhythm and electrical activity, but these machines are often not available in primary care. An alternative, cheap and easy-to-use tool for taking ECGs could help GPs diagnose heart disease earlier.

Researchers used a stethoscope with additional sensors (electrodes) to take ECGs. Readings were analysed using artificial intelligence (a problem-solving computer programme) to detect heart failure. This would normally require an expensive, time-consuming ultrasound test (echocardiogram).

In this study, researchers compared the accuracy of the smart stethoscope with the standard echocardiogram.

What’s new?

The study was carried out in 7 NHS hospitals and community health centres in London. More than 1,000 people took part. They had been referred for an echocardiogram to investigate heart failure symptoms (such as breathlessness, chest pain) or to check for high blood pressure, stroke or other conditions related to the heart and blood vessels.

Using a regular echocardiogram, the researchers confirmed that 1 in 10 had heart failure (less than 40% of blood pumped out of the left side of the heart). Participants were then examined with the smart stethoscope. The researchers took readings at 4 different positions on the chest. Recording and analysis with artificial intelligence took 2 minutes per person.

Of the positions tested, the reading taken over the upper left side of the heart was best at detecting heart failure. When this reading was combined with another position, the results were even better.

With the combined readings, the smart stethoscope:

  • correctly identified the signs of heart failure 9 out of 10 times
  • missed few people with findings of heart failure on echocardiogram (sensitivity of 92%, compared with 81% in other studies using standard ECG)
  • rarely identified heart failure when it was not there (specificity of 80%, compared with 67% for standard ECG).

A normal heart pumps out more than 50% of blood with each beat. Most people incorrectly identified as having heart failure had reduced heart function and were close to the cut-off (pumping 41–50% of blood).  

Neither age, sex, nor race affected the smart stethoscope’s ability to detect heart failure.

Why is this important?

The smart stethoscope accurately screened for heart failure 9 out of 10 times. Its sensitivity and specificity were similar to previous studies using a standard ECG machine. The researchers say this could allow primary care clinicians to detect heart failure earlier. Referrals to secondary care would be more accurate, they say, because GPs would not have to rely only on the symptoms of heart failure (such as breathlessness and being overly tired) that can be caused by many conditions.

The device requires little training, so non-experts can use it following simple instructions. However, doctors will need to trust the tool before they use it. GPs may be more willing to use the smart stethoscope than other ECG devices with a single lead because they are familiar with stethoscopes.

The tool is relatively cheap and could be used in poorer countries to diagnose heart failure where there is limited access to specialist heart services.  

What’s next?

The researchers are continuing to follow-up the group of participants who were falsely identified as having heart failure. Many in this group had reduced heart function on the border of heart failure. Other research shows these people could be up to 4 times more likely to go on to develop heart failure (compared to the general population).

The smart stethoscope could encourage GPs to refer patients for specialist care more quickly. This could mean they start treatment without waiting for an echocardiogram, the current gold standard test for heart failure. 

In future, the smart stethoscope could be used to detect other heart conditions related to electrical activity (such as atrial fibrillation). Similarly, it could be used to diagnose conditions based on heart sounds, such as conditions related to heart valves. This would not be possible with a regular ECG.

Further research could explore whether including blood tests or people’s age, sex, blood pressure or other characteristics could improve the accuracy of diagnoses. Future studies could also examine the value for money offered by the device, and whether it improves survival and other outcomes.

The participants in the study were all likely to have trouble with their heart. It is not clear if the smart stethoscope would work as well on the general public. The researchers are now testing the smart stethoscope’s accuracy in a large-scale trial across 200 GP practices in London and Wales.

You may be interested to read

This NIHR Alert is based on: Bachtiger P, and others. Point-of-care screening for heart failure with reduced ejection fraction using artificial intelligence during ECG-enabled stethoscope examination in London, UK: a prospective, observational, multicentre study. Lancet Digital Health 2022;4:e117–25.

Easy-to-read information about heart failure is available on the Pumping Marvellous website.

Simple information on heart failure from the British Heart Foundation.

Research regarding a 12-lead ECG that used artificial intelligence, which inspired this study: Attia ZI, and others. Screening for cardiac contractile dysfunction using an artificial intelligence–enabled electrocardiogram. Nature Medicine 2019;25:70–74.

A summary of the NHS Long-Term Plan, which aims to leverage digital technologies to transform the health service.

NIHR Evidence Alert: Advance care plans improve quality of life for heart failure patients

Funding: This study was supported by the NIHR Artificial Intelligence in Health and Care Award.

Conflicts of Interest: The authors did not declare any conflicts of interest.

Disclaimer: NIHR Alerts are not a substitute for professional medical advice. They provide information about research which is funded or supported by the NIHR. Please note that views expressed in NIHR Alerts are those of the author(s) and reviewer(s) at the time of publication. They do not necessarily reflect the views of the NHS, the NIHR or the Department of Health and Social Care.

  • Share via:
  • Print article
Back to top