Many members might have been aware of the ongoing work towards a NICE early value assessment (EVA) around how AI can be used in skin cancer pathways in the NHS. This guidance has now been published and more information can be found here: Overview | Artificial intelligence (AI) technologies for assessing and triaging skin lesions referred to the urgent suspected skin cancer pathway: early value assessment | Guidance | NICE
With many of us being involved in teledermatology pathways, some involving AI, and with the NHS looking for ways to make patient care as efficient as possible, it’s worth being aware of this guidance and what it might mean for your services.
In summary, the EVA suggests that Skin Analytics’ DERM AI can be conditionally used within NHS pathways, as long as appropriate safety nets are in place during the 3-year evidence collection period. This is on the condition that the evidence from the evidence generation plan is being collected, and the appropriate regulatory approval (including NHS Digital Technology Assessment Criteria) is obtained.
The safety nets described are:
· Having AI decisions reviewed by a healthcare professional for patients with black and brown skin.
· Regular monitoring of DERM’s performance.
· Using additional protocols where necessary, like a healthcare professional review or a national governance framework to ensure local oversight.
The British Association of Dermatologists (BAD) have released a statement on the EVA, which can be found here: https://bad.org.uk/bad-statement-on-nice-early-value-assessment-on-ai-technologies-for-suspected-skin-cancer/
The BAD recommends that, until further evidence is generated, all patients should have some input from a dermatologist and that any AI implementation should have a clear underpinning of patient safety. They also outline that it is important to fully understand the benefits and limitations of AI.
When we look to the future, AI will likely be an important tool for healthcare professionals, however it will always be key to ensure full understanding of how this fits in with local pathways to make sure patients truly benefit and are not put at risk by the unsafe implementation of AI.
Having an awareness of this guidance as well as the BAD statement around it will help our members be part of discussions around local teledermatology pathways, and make sure that photographers, and their skills and knowledge, are key to any new implementations.