How are you taking part in this consultation?

You will not be able to change how you comment later.

You must be signed in to answer questions

  • Question on Consultation

    Has all of the relevant evidence been taken into account?

  • Question on Consultation

    Are the recommendations sound and a suitable basis for guidance to the NHS?

  • Question on Consultation

    Are the summaries of clinical and cost effectiveness reasonable interpretations of the evidence?

  • Question on Consultation

    Are there any aspects of the recommendations that need particular consideration to ensure we avoid unlawful discrimination against any group of people on the grounds of race, gender, disability, religion or belief, sexual orientation, age, gender reassignment, pregnancy and maternity?

HealthTech draft guidance

Open for comments Open until   Request commenting lead permission

We have updated this service so that members of the same organisation can now collaborate on a joint online response.

Read our blog to learn more.

1 Recommendations

More research is needed

1.1

More research is needed on the following artificial intelligence (AI)-assisted echocardiography analysis and reporting technologies to support the diagnosis and monitoring of heart failure, before they can be funded by the NHS:

  • EchoConfidence

  • EchoGo Heart Failure

  • Ligence Heart

  • US2.ai

What this means in practice

There is not enough evidence to support funding AI-assisted echocardiography analysis and reporting technologies to support the diagnosis and monitoring of heart failure in the NHS.

Access to the technologies should be through company, research or non-core NHS funding, and clinical or financial risks should be managed appropriately.

What research is needed

More research is needed on:

  • the impact on service capacity, including:

    • time saved during echocardiography procedures when the AI technologies are used in NHS secondary, community or primary care settings

    • the impact on number of appointments per day and waiting times

    • time taken for human review of AI findings

  • reliability and failure rate (for example, performance with echocardiograms of varying quality)

  • diagnostic accuracy and performance when the AI technologies are used by operators with varying levels of experience outside of secondary care

  • patient selection and cohorts used in the AI training and validation datasets, to assess whether these reflect the diverse population seen in NHS clinical practice

  • healthcare professional-reported ease of use and acceptability

  • any harm caused by using the technologies

  • for technologies with the additional functionality, accuracy in detecting or distinguishing between heart failure and other less common cardiac conditions (which may co-present) such as amyloidosis and heart valve disease.

Why the committee made these recommendations

Limited evidence suggests that using EchoConfidence or Us2.ai could reduce the time taken for echocardiography analysis and reporting. But there is no evidence to suggest that the time saving results in more people being seen or reduced waiting times.

Most of the clinical-effectiveness evidence for the AI technologies relates to diagnostic accuracy. For all 4 technologies the evidence is uncertain and may not reflect clinical practice in the NHS. This is because many of the studies:

  • are retrospective and observational

  • were done in single centres with single operators

  • were done outside of the UK

  • excluded complex cases and poor-quality echocardiography scan images.

Because of the uncertainties in the evidence, it is not certain whether the AI technologies could lead to delays and longer waiting times if healthcare professionals need to check and review the AI findings. It is also uncertain whether EchoGo Heart Failure, Ligence Heart and Us2.ai were developed using population data that reflects the NHS. So, using these AI technologies in the NHS could pose clinical risks.

The results of the economic model suggest that EchoConfidence and Us2.ai could be cost effective. But this is uncertain because of the uncertainty in the clinical evidence. It is not possible to assess the cost effectiveness of EchoGo Heart Failure and Ligence Heart because of limited evidence. So, more research is needed on all 4 technologies.