A blood test request in much of the NHS still begins with a paper form and the hope that it will meet its sample in the laboratory. Each stage invites delay. Forms go missing, handwriting defeats the person keying it into a laboratory system, and results reach a ward after the patient has gone home. At York and Scarborough Teaching Hospitals NHS Foundation Trust, that sequence has now been replaced across its sites. Electronic requests for pathology and radiology went live as the second phase of the trust's digital record upgrade, with no system downtime and no interruption to patient services.

The lack of disruption is notable in itself. Electronic patient record programmes in England have a mixed track record. Several trusts have reported slower clinic throughput, frustrated clinicians and lost activity in the months after go-live, at a time when every lost appointment adds to waiting lists that still run into the millions. A trust that moves test ordering onto a new platform without that dip has done something national programmes have often failed to reproduce. The phasing looks like the reason. Inpatient documentation, prescribing, emergency care tracking and bed management came first, and requesting followed once staff were used to the system.

The timing is politically significant. The government's ten year health plan commits to a shift from analogue to digital, built around a single patient record and wider use of AI in clinical and administrative work. Ministers have pressed trusts to have electronic records in place by fixed dates, after earlier deadlines passed. Those ambitions are set in Whitehall and delivered in places like York, where a board must find capital, clinical time and supplier capacity while running emergency departments through winter. National strategy depends on that local work being finished.

Diagnostics make a sound place to look for early returns. A cancer referral waits on imaging, an emergency admission on bloods, a discharge on a result. Electronic requests make each step visible and timed. Clinicians can see what has already been ordered, which reduces duplicate tests, and managers gain a record of how long each stage takes. The productivity case for digitisation is usually made in general terms. Here it can be checked against turnaround times, repeat testing and the length of diagnostic waits, all of which the trust can now measure far better than before.

The trust also says it is preparing its infrastructure for artificial intelligence, a claim that deserves disciplined reading. Algorithms that summarise records, flag deterioration or prioritise scans are only as good as the data beneath them. Free text, scanned documents and inconsistent coding limit what any model can do. Structured electronic orders produce clean, timestamped data as a by-product of ordinary work. If AI is to function at scale in the NHS, much of its raw material will come from a clinician choosing a test from a menu.

Caution is still warranted. Completing a phase is a measure of delivery, and zero downtime tells the public that the system stayed up. It says little about whether clinicians are faster or patients wait less. The later phases, covering core clinical administration, are harder because they touch outpatient booking, referral management and the data that feeds waiting list reporting. Workforce strain compounds the difficulty. Staff who have absorbed one change after another need training time that rotas rarely allow, and a record that clinicians work around generates poor data however sophisticated the platform.

Other trusts can draw practical lessons. Sequencing matters, as does clinical involvement in design and a willingness to delay a phase until the previous one has settled. Policymakers should attach funding to published evidence of use, such as the share of requests made electronically, repeat test rates and diagnostic turnaround, and avoid treating a go-live date as the achievement. Suppliers will find that credibility accrues to those who deliver change without loss of service. Patients may notice little at first. The benefit, if it arrives, will appear as fewer repeat tests and fewer lost results.

The national plan for a digital NHS will be settled in thousands of small decisions about where and how a test is ordered. York has shown that one of those decisions can be changed safely. The next task is to show that it shortens a clinician's working day or a patient's wait, and the trust should be asked to publish the evidence.

‍