TL;DR

Barts Health NHS Trust is starting an 18-month evaluation of AI decision support for its breast cancer multidisciplinary teams (MDTs), Digital Health reported on 4 November. The tool, Cancer Pathways IQ from Deontics, aims to identify which patients most need full team discussion. This is a test with stated measures rather than a rollout with results, and a drug company’s involvement is worth keeping in view.

MDTs bring specialists together to agree a patient’s treatment, and consultant oncologist Dr Melissa Phillips says the demands on them keep rising. The trust wants to learn whether the software can trim “unnecessary workload and discussion” and release capacity, while still supporting safe decisions. Deontics, a spin-out linked to Oxford University, UCL and Cancer Research UK, combines patient information with published evidence. The pilot will also try a newer feature that lifts structured data out of radiology and pathology reports and referral letters.

What will be measured

The useful detail is in the evaluation plan. The project will track preparation time, discussion time, admin, repeat discussions and time per patient, adding those into capacity released each month. It will also look at performance on the 62-day cancer waiting time and the number of patients seen through the service.

That target gives the work its urgency. Digital Health notes the NHS has not hit its 85% standard for starting treatment inside 62 days since 2014. February’s National Cancer Plan promises the standards will be met by March 2029.

The AstraZeneca question

The deployment sits under a joint working agreement linking the trust and AstraZeneca. AstraZeneca’s UK oncology head, Anna Arent, speaks of “enabling more patients to access the right treatment sooner”. The report gives no breakdown of who funds what, or how the evaluation is overseen.

In our view, those are the questions any NHS board should ask before relying on the results. The measures listed are mostly about time and throughput, so a good result would show capacity gained rather than better care in itself. The list named in the report does not include clinical outcomes, or how often a patient triaged away from discussion should have been seen, although Phillips does say the team will keep supporting safe decision-making. A safety measure of that kind would make the findings far more useful to other trusts, which face the same pressure on these meetings and will want to know what was lost as well as what was saved.