Categories
Public
We focused on the use of machine learning applications for diagnostic purposes in healthcare services.
What we did
We worked with:
- healthcare providers
- people who use services
- clinicians
- technology suppliers
- other stakeholders
Together we agreed what care providers and tech suppliers need to do so that people receive high-quality care through these applications. We used this to identify and consider how to update our methods - to help us regulate these services better. We also considered how other parts of the healthcare system could support high-quality care.
What we found
Key findings:
- healthcare providers that use machine learning in diagnostic services need to have good governance for the clinical, information, technical and human aspects of the application. We still need to develop our approach with providers to get this right
- most suppliers of machine learning applications in diagnostics won’t need to register with CQC. Only those suppliers that deliver clinical activity themselves as part of a regulated activity need to register
- to regulate the few suppliers that do become registered providers, and assure the public that their services are safe and effective, CQC will need other national bodies to develop technical standards and assess against them
- there is a need for more assurance about the clinical aspects of the algorithms in machine learning, and more clarity on how hospitals should implement machine learning devices within clinical pathways to ensure high-quality care
- technology suppliers need to be clear what their products, solutions and devices do and how they perform. Suppliers do not always accurately state whether their products use machine learning. This makes it harder to implement devices safely and poses a risk to patients
Regulators' Pioneer Fund
This was made possible by a grant from the £10m Regulators’ Pioneer Fund launched by The Department for Business, Energy and Industrial Strategy (BEIS) and administered by Innovate UK.