An Intelligent Stethoscope

Lead Research Organisation: University of Cambridge
Department Name: Engineering

Abstract

Valvular heart disease (VHD) has become an epidemic and nearly two million people in the UK suffer from it. This number is expected to double by 2040.
About half of those affected by VHD are unaware of their condition. Early diagnosis is the key to providing essential treatment and preventing untimely death. We are developing an intelligent stethoscope to automatically detect heart murmurs and the underlying pathology, which would help provide early diagnosis of VHD.
Currently, heart murmurs are picked up as a part of a physical exam by GP using a stethoscope. However, VHD can often become severe before symptoms develop, meaning many patients may only be picked up when they have already developed irreversible complications, such as heart failure, or post-mortem.
The AI-enabled stethoscope proposed could provide a much higher accuracy than could be achieved by trained practitioners via auscultation. It could also be used by less-skilled healthcare professionals in a screening program for those at risk or by patients for at-home monitoring. This would reduce the number of missed diagnoses, reduce the burden on GPs of detection, and support early intervention to minimise morbidity and mortality.

Technical Summary

Valvular heart disease (VHD) has become an epidemic that currently affects 1.3 million people in the UK. This number is expected to double by 2040 with ageing of the population. 800,000 of those affected by VHD are unaware of their condition. Early diagnosis is key to providing timely treatment, minimising morbidity and preventing premature death. We are developing an intelligent stethoscope to automatically detect heart murmurs and diagnose the underlying pathology, which would improve early diagnosis of VHD.
Currently, heart murmurs are picked up as a part of a physical exam by a General Practitioner (GP) using a stethoscope. However, VHD is often identified late, when patients are symptomatic and when many have already developed irreversible complications (such as heart failure) or post-mortem.
Our proof-of-concept system can detect 94% of the heart murmurs identified by doctors. With the collection of more clinical-data and matched echocardiograms, the system will be able to surpass the rate of identification and accuracy achieved by trained practitioners via auscultation. It could also be used by less-skilled healthcare professionals in a screening program for those at risk. This would reduce the number of missed diagnoses, reduce the burden on GPs of detection, and support early intervention to minimise morbidity and mortality.

Planned Impact

VHD has become an epidemic that currently affects 1.3 million people in the UK. This number is expected to double by 2040 with ageing of the population. 800,000 of those affected by VHD are unaware of their condition. Early diagnosis is key to providing timely treatment, minimising morbidity and preventing premature death. We are developing an intelligent stethoscope to automatically detect heart murmurs and diagnose the underlying pathology, which would improve early diagnosis of VHD.

The main beneficiaries are the patients who will be diagnosed with valve disease or congenital heart disease that could otherwise go undetected. These patients would then get access to the right treatment in a timely fashion and avoid late-stage heart failure. We have already involved patients in this project through our collaboration with the charity Heart Valve Voice. We held a focus group with patients affected by VHD, to try and understand how our technology could have helped them and to get feedback on the proposed research study.

This technology could also have a big impact in developing countries, where the burden of rheumatic heart disease (RHD) is high. RHD is estimated to be responsible for 319,400 deaths per year worldwide, with the highest mortality rates being in Oceania, South Asia, and central sub-Saharan Africa. RHD can be treated with antibiotics if it is detected early.

By extending the pool of healthcare professionals able to diagnose VHD, the electronic stethoscope will decrease the demand on GPs and will reduce the burden of managing late-stage heart failure. Reducing demand on GPs and new ways of working are recognised as critical by NHS England.

Publications

10 25 50
 
Description Queen Elizabeth Hospital, Birmingham 
Organisation Queen Elizabeth Hospital Birmingham
Country United Kingdom 
Sector Hospitals 
PI Contribution We are developing a software to automatically detect valvular heart disease based on the data collected by our collaborators at their hospital.
Collaborator Contribution The collaborators are collecting heart sounds (auscultation) and echocardiography data from patients who enrol for the study.
Impact We are in an early data-collection phase.
Start Year 2019
 
Description Royal Papworth Hospital 
Organisation Papworth Hospital
Country United Kingdom 
Sector Hospitals 
PI Contribution We are developing a software to automatically detect valvular heart disease based on the data collected by our collaborators at their hospital.
Collaborator Contribution The collaborators are providing clinical data-collection support (project, data and trial management) and are collecting heart sounds (auscultation) and echocardiography data from patients who enrol for the study.
Impact We are in an early data-collection phase.
Start Year 2019
 
Description Innovation in Valve Disease Meeting 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact The objectives of the meeting were to:
the objectives for the meeting, namely to:
• Identify existing best practice, including how innovation can improve this
• Recognise the barriers to uptake for innovative technology and discuss how these could be overcome
• Explore how the current pathway can be modernised to better suit patient needs
• Identify innovative technologies that can help the NHS streamline the referral process for heart valve disease and/or improve the identification of suspected heart valve disease
• Secure support for pilot schemes to test these innovations as specifically related to heart valve disease
Year(s) Of Engagement Activity 2020