Results from the WISDOM AI study

People recovering from surgery are increasingly being given the opportunity to send photos of their wounds from home using smartphone apps. This can help their healthcare teams identify infections or healing problems earlier, without the inconvenience and cost of patients needing to travel back to hospital. 

The original WISDOM study developed an artificial intelligence (AI) tool designed to help prioritise which wound images may need urgent clinical review. The updated WISDOM 1.1 model was developed using more than 17,000 surgical wound images, including thousands with clinician-assessed darker skin tone information. This is important because AI systems and clinical training have in the past performed less well for people with darker skin tones.

The study showed that WISDOM 1.1 improved detection of wounds needing urgent review, compared with the original model and reduced some of the performance differences between lighter and darker skin tones. The model performed particularly well as a safety-focused assessment tool, helping identify wounds that may need earlier attention from clinicians.

Although some differences in performance remained, especially for colour-based signs such as redness in darker skin tones, the results suggest the technology is ready for testing in larger real-world studies alongside normal clinical oversight. The long-term aim is to support faster identification of wound problems, reduce unnecessary hospital visits, and improve access to fair and equitable postoperative care for all patients.