A consultant sees a patient in the clinic, completes the review and discharges them. Nothing in the hospital record shows that the patient is still waiting for cancer to be ruled out. Elsewhere in the building, a colleague requested a follow-up within six weeks. The patient deferred once, the clinic was full, and the new date landed eight or nine months away. Nobody was positioned to notice.

These accounts come from a discovery project that DrDoctor ran with Chelsea and Westminster Hospital NHS Foundation Trust and the North West London Integrated Care System. Before proposing any software, the company held a stakeholder workshop in November 2020 and then interviewed more than 25 people, among them consultants, nurse specialists, general managers and digital leads. It listened first, and the resulting picture was more candid than most internal reviews manage. More than five years on, with the waiting list still counted in millions, the findings describe the present as much as the past.

DrDoctor reduced them to a single sentence. Clinicians and administrators could not reliably see where patients stood on their pathway, which made it hard to secure appointments when they were needed. The 18-week referral to treatment clock was closely watched, and breaches at 18 and 52 weeks were plain to see. Beyond it, visibility thinned. Follow-up and chronic patients sat across several lists, some necessary data was never collected, and the main reporting tool refreshed once or twice a day and was distrusted by some staff. Clinicians kept personal spreadsheets. Booking teams relied on meetings to decide whom to book first. The e-referral service did not talk to the main patient record, and cancer pathways were keyed in by hand. One multidisciplinary decision to operate was printed and never uploaded, so the patient arrived for surgery and the surgeon knew nothing of the appointment.

Current policy asks the system to perform in exactly this terrain. Ministers have committed to restoring the 18-week standard, which requires management of the whole pathway and a record that can be trusted. Patient-initiated follow-up, which lets patients trigger an appointment when symptoms change, is a national lever for releasing outpatient capacity, and it works only if the list shows who can safely be left off it. Productivity arguments tend to begin with theatre utilisation and clinic throughput. DrDoctor's research went to the prior question of whether the list is true. A wrongly populated list cannot be cut efficiently, however many sessions are added.

What followed shows sound judgement. DrDoctor advised that scope should be small enough to deliver and show results. When the needs of the ICS shifted and Palantir was brought in to provide end-to-end data visibility, DrDoctor took the patient-facing layer, where trusts in the area already used it for reminders and rescheduling. Its design covers booking from waiting lists, self-service rescheduling, secure messaging and outcome questionnaires, all of which feed back into the record. Patients are the only participants who know whether they deferred, whether symptoms have changed and whether they still want the appointment. A channel that lets them say so corrects the record at source, which no integration engine can do alone.

The limits deserve plain statement. The discovery project closed after its first phase, and the available material contains no outcome data on waiting times or missed appointments. The evidence comes from a single trust under pandemic strain. Palantir's role in NHS data remains politically contested, and a patient-facing supplier sits close to that argument, because patients will act on a message only if they trust what stands behind it.

The implications are practical. Trust leaders should audit the lists outside the 18-week window first and treat clinician spreadsheets as a map of where provision has failed. Policymakers should fund data quality and integration with the same enthusiasm they show for new tools. Suppliers should expect to be judged on whether fewer patients are lost. Patients should be treated as a source of information as well as recipients of it.

The backlog will not clear on capacity alone. Knowing where each patient stands is the unglamorous precondition, and DrDoctor identified it, scoped it sensibly and took on the part of the problem it was best placed to solve. The next test is evidence that fewer people disappear between appointments.

‍