CW Innovation Blog Series: Mike Wright

For the first in a series of blog posts showcasing the work of the CW Innovation Team, we spoke to Innovation Business Partner Mike Wright, who leads the team and supports the delivery of key innovation projects across the Trust. He talked to us about the CW Innovation programme and his work on Skin Analytics at Chelsea and Westminster Hospital.

What is CW Innovation?

CW Innovation is a partnership between CW+ and Chelsea and Westminster Hospital NHS Foundation Trust. It cements a mutual commitment to drive innovation at the Trust, whether that’s about benefiting patients and staff with new products or way of working, or supporting long-term sustainability by creating a step-change in financial and operational performance. Since its inception in 2019, the programme has gone from strength to strength.

Can you tell us about Skin Analytics?

For the past three years I have worked alongside colleagues in the dermatology service at the Trust on DERM (Deep Ensemble for the Recognition of Malignancy), an Artificial Intelligence (AI) tool, developed by Skin Analytics, to support our urgent skin cancer pathway at Chelsea and Westminster Hospital. Its primary purpose is to allow clinical teams to triage patients, quickly identifying those at high risk of skin cancer and prioritising them for consultant review. It also identifies low-risk lesions, potentially enabling swift discharge and reassurance.

The suspected skin cancer pathway accounts for the highest volume of urgent referrals of any suspected cancer pathway. Nationally, the volume of referrals has risen by almost 200% since COVID. There is also a national shortage of consultant dermatologists, with almost 24% of vacancies remaining unfilled, which has put dermatology services under enormous pressure. The advent of new technologies such as AI has the power to transform ways of working, helping to meet the rising challenges of delivering healthcare in a timely fashion.

How does the AI DERM technology work?

Patients are referred to the Trust by their GP. Their first appointment is with a medical photographer, who takes images of their skin lesion – for example, a mole – using a mobile smartphone with a dermatoscope (magnifying lens) attachment. Images are captured and uploaded to the Cloud, where they are analysed by DERM AI and categorised as high risk or low risk. Based on the outcome of the assessment, patients are then directed to the right level of care.

Hear about a patient’s experience of Skin Analytics.

How have you been involved in the project?

CW Innovation has supported this project from its earliest stage, including supporting the sourcing of funding, and working closely with Lucy Thomas, Consultant Dermatologist and Innovator at the Trust. I have provided expert guidance, including preparing the data, privacy impact assessments, clinical safety case reports, looking at qualities and equalities, health impact assessments (essential in ensuring that all patients benefit from new technology, not just a limited few) and navigating the governance. I’ve also locally reviewed contracts and agreements with Skin Analytics.

What impact has Skin Analytics had on the dermatology service?

The Skin Analytics project has been very successful to date. A service overview (April 2022 to January 2024) shows that 6,320 cases have come through the pathway and 8,321 lesions have been assessed (sometimes patients are referred with more than one lesion). 3,196 cases have been discharged – 1,543 by Skin Analytics (without relying on Trust capacity) and 1,653 by the Trust directly.

We had a number of goals for the programme when we set out. The first goal was to discharge more than 20% of urgent skin cancer patients without relying on Trust clinician capacity – it’s currently running at about 24%. The second goal was to decrease the amount of time Trust clinicians spend processing urgent teledermatology cases. Since introducing Skin Analytics, the service now sees 650 additional patients a year using the same number of dermatologists. The third goal was to maintain consistent success in meeting the national standard for urgent skin cancer referrals, which we have exceeded.

In summary, the use of AI DERM has allowed us to see more patients, and to give a quicker diagnosis and reassurance to patients. It also frees up clinician time for non-urgent skin cancer pathways such as severe eczema or psoriasis, where patients can experience delays in getting an appointment, which can negatively impact their care.

How else has it made a difference?

This work has prompted a much broader discussion about the use of AI in the Trust, and it has paved the way for other AI projects including the Dora AI pilot, which supports cataract patients at Chelsea and Westminster Hospital. Another example is the Chest X-ray AI in radiology, which is used for early detection of lung cancer. It has opened up a world of possibilities.

What are the next steps for Skin Analytics?

One of the next steps is to roll out Skin Analytics at West Middlesex University Hospital, which will enable us to reach even more people. It also means that whether patients are referred to West Middlesex or Chelsea on the urgent skin cancer pathway, they will get exactly the same service.

We would also like to remove the second-read safety net. Low-risk lesions currently go to a second panel of dermatologists employed by Skin Analytics, who will either confirm they are low risk – in which case the patient will be scheduled for discharge – or disagree with the AI, in which case the patient will go back into the skin cancer pathway for review by Trust clinicians.

By removing the second-read safety net, we could discharge about 41% of patients without clinician review, compared to our current figure of 24%. This would be a major step and it would free up a lot of resource that we can use to support other patient groups on other pathways. Such a step will only be taken once we are confident that patient safety is not compromised.

Any longer-term developments in the pipeline?

The work we’ve done and the track record we’ve established potentially allows us to innovate around how we deliver dermatology services in the Trust. There is no reason why the initial Skin Analytics medical photography appointment couldn’t be done out in the community – in a community diagnostic hub, for example.

We intend to use our experience of working with a frontline clinical AI, provided by a reputable innovative company, to identify further opportunities to innovate in clinical pathways. Improving the service we offer to patients, both in terms of how long they have to wait for an appointment and the time taken to get a diagnosis through to successful treatment, will remain at the heart of all our decision making.

Skin Analytics is just one example of the ways in which CW Innovation has been able to make a significant improvement to the experience of staff and patients at the Trust. But innovation doesn’t have to mean using cutting-edge technology – sometimes just developing new ways of working can deliver equally important results. As CW Innovation heads towards its fifth anniversary, we’re keen to reach as many staff as possible, in every role and division. The best ideas are always going to come from the people at the frontline of delivering care to patients, and it’s our role to help them bring those ideas to life.