CQC has published its findings from its first ‘regulatory sandboxing’ exercise, the output of which includes a change in the criteria which providers of digital triage tools in healthcare services are inspected against, as well as a number of recommendations and insight for the wider health and care system.
What is a regulatory sandbox?
According to CQC, ‘Regulatory sandboxing is a way of working proactively and collaboratively to understand new types of health and social care service, agree what good quality looks like, and develop our approach to regulation.’ The concept of sandboxing has been used in many sectors to allow safe space for innovation, and whilst this is relatively new in the context of the health sector, we have seen the Information Commissioner’s Office also set up a sandbox for a number of projects, with NHS Digital being one of the chosen organisations to participate last year.
In terms of the CQC sandbox, there are currently three pilot rounds, the first of which focused on the use of digital triage in healthcare services, and involved healthcare providers using such services, technology suppliers, government partners and patients. The other two regulatory sandbox pilots will:
• Consider how machine learning algorithms support diagnostic and imaging services; and • Examine providers delivering community care at home such as personal assistants and other micro-providers.
The reports of these sandboxing exercises are expected later this year.
What is digital triage and why is it important?
Clinical triage is an important part of how healthcare is delivered as it directs people to the health and care services that best meet their needs. However it is not without risk; if triage is not carried out properly, it can delay or deny timely care leading to potentially significant consequences for both the patient and triage provider. The use of digital technologies to carry out triage is developing rapidly and CQC recognised that it needed to fully understand how such services are being delivered, in order to work out the best way to regulate them. Digital triage can take many forms and the following were considered as part of the sandboxing exercise:-
• Patient facing tools i.e. tools that ask people questions about their symptoms and direct them to services based on their answers. • Tools for call handlers i.e. tools using standardised algorithms to support non-clinical staff to make triage decisions by phone. • Clinical decision support tools i.e. tools that make suggestions and guide clinicians in making decisions about how to triage or treat people.
Establishing what ‘Good’ looks like for digital triage
One of CQC’s primary aims for this regulatory sandbox was to establish what ‘Good’ looks like for digital triage. CQC admit in the report that ‘we have struggled to delineate clear boundaries between the suppliers of technology solutions and the health services that use them’. This is often due to the fact that ‘contracts may specify that either the tech supplier or the health service or commissioning body carries out critical activities such as pathway development, testing, and training staff to use the tool.’
To address this ambiguity, CQC have confirmed that going forward, what is suggested in the report as good practice will apply to both the supplier and care provider (or commissioners where they are procuring on behalf of providers). Section 7 of the report sets out a detailed statement of what ‘Good’ will look like for the ‘Safe’, ‘Effective’, ‘Caring’, ‘Responsive’, and ‘Well-led’ domains, for both ‘digital triage tool suppliers, and the providers that use them’.
It is helpful that CQC acknowledges the important point that the performance of such digital tools can be affected by the acts and omissions of several parties, particularly where the product is subject to updates by the original supplier or third parties, or has the capacity to learn independently from the data on which it works. The report suggests that technology suppliers who remain responsible for certain aspects of the digital triage tool (such as developing pathways, testing and staff training), could require registration with CQC in their own right. However, later in the report CQC explicitly state that they do not regulate technology suppliers that deliver digital triage tools, and so the report is somewhat ambiguous in this respect.
The report states that
“Suppliers of digital technology need not register unless they are in a small minority that carry out regulated activity. However, where a healthcare provider is using these tools, we need to understand how well they are working for patients and, if necessary, CQC has the powers to review key third-party technology suppliers responsible for an activity ancillary to regulated activity”.
How CQC moves forward in this respect and what is involved in “review” in this respect will be keenly observed by all parties (and their legal advisors). Ultimately, since safety and quality are likely to be goals shared by both suppliers and registered providers, it would be wise for both to reflect on and aim to deliver, for their part, what “good” looks like according to CQC. The principal focus of CQC will inevitably be on registered providers and their governance systems in procuring and deploying products and services to support their delivery of regulated activities. However, any party retaining substantial responsibility (whether by contract or the reality of how the tool is maintained and works “on the ground”) for the safety and quality of the digital triage tool may receive attention from the CQC.
A new approach to inspection
CQC have concluded from the sandboxing exercise that they are not currently capturing all of the information they need to in order to form a judgment as to whether digital triage services are ‘Good’ (according to the new definition set out in the report) at inspection, and this is impacting their ability to assess the quality of care in this area.
CQC will therefore be revising their current inspection framework for digital triage services published in February 2019 (‘Additional prompts for triage apps supporting healthcare services’) to ensure that the following is present:-
• Clinician-facing decision support tools allow clinicians to override their suggestions where it is in the best interest of people who need care. • There is a clear, workable and consistently applied way for patients to access care if the tool is unsuitable or unavailable for any reason. • The tool is accessible, user-friendly and does not frustrate patients and staff. • The information sent from the triage tool to healthcare services is reliable and provides sufficient clinical detail to meet their needs when accepting patients. • Safety is managed well and follows NHS Digital’s Clinical Risk Assessment guidance. • The development of clinical pathways is governed effectively, the pathways are evidence based, testing happens before they are put into use, and clinical input is secured. The organisation responsible also audits its pathways regularly.
CQC also make clear in the report that providers that use digital triage tools must be able to provide assurance that their chosen tool is ‘appropriate for their purposes and supports safe, high-quality care and this will usually mean requiring evidence of compliance with NHS Digital’s guidance DCB0129 and product certification of the appropriate class before procuring’. Commissioners, providers and suppliers will need to consider ever more carefully the application of DCB0129 and DCB0160 in any procurement exercise. If a non-compliant system is deployed, it is essential that a defensible reason is given following due scrutiny of the standards and accompanying guidance.
The CQC acknowledge the patchwork of existing standards and guidance applicable to such products, and that due to their nature, some triage tools may not fall within the ambit of existing medical device regulations enforced by the MHRA. It is worth nothing that in Evaluating Digital Health Products, Public Health England has very recently published extensive guidance online as to digital product “evaluation”. This is a helpful reference tool pointing readers to a range of standards and obligations that could apply instead of or in addition to those mentioned by the CQC.
Impact for the wider healthcare system
In addition to achieving their main aim of defining what good quality care looks like for digital triage tools, the sandboxing exercise identified a number of issues for wider consideration by other organisations in the healthcare system. The report lists 5 recommendations directed at specific bodies such as NICE, the Department of Health, NHS England and NHSX.
These recommendations flow from a number of observations concerning the following:-
• Commercial and contracting gaps when procuring digital triage systems and the reliance on an NHS Standard Contract, which is not fit for this purpose. Whilst the NHS Standard Contract can be adapted, it has many terms (including Service Conditions and General Conditions) that cannot be amended and are not designed for contracting for digital health solutions. • Digital triage tools are not independently externally clinically validated and there is significant variation in their clinical performance. We note that there are a number of organisations in the market that are offering services to assess digital health tool compliance (including a focus on clinical effectiveness) however there is no “kite mark” recognised as yet. • People use digital triage tools differently and there is concern that patients are not clear about what they are using and how they should treat it. • The information needed to be captured by digital triage systems to identify patients in vulnerable positions, such as those at risk of abuse and neglect. • How digital triage tools can be used to improve care pathways, integration and efficiency.