Define the answer before choosing the pilot.

NICE's evidence standards framework helps developers plan evidence for digital health technologies (NICE, n.d.). NHS England's clinical safety strategy describes manufacturer and deployment responsibilities (NHS England, n.d.). Neither removes the need to define the actual task and assess the requirements that apply to it. An organisational preparation aid and a patient-specific treatment function are different propositions. The study should say what participants will do, what the technology produces and who reviews the output. Without that definition, claims about evidence or assurance have no stable object.

A small study may establish that participants can use the workflow and that the required inputs are available. It may not be able to show a reliable performance difference or downstream clinical benefit. State those limits in the plan. Early development can use synthetic or appropriately de-identified examples where suitable; later operational work requires its own approvals and design. The evidence programme should progress because an earlier result supports the next question, not because the word pilot implies that the product is already ready.

A proposed first study could measure quality, traceability, total effort and the kinds of errors reviewers must correct. Define critical failures separately from minor edits. Decide who reviews the outputs, whether reviewers can be blinded to the preparation method and how disagreement will be resolved. The analysis plan should also describe missing observations and participant withdrawal. These choices make a modest study more interpretable and help prevent a striking anecdote from becoming the public claim for the entire programme.

Stop for the agreed critical quality failure, inappropriate use, unavailable inputs or a checking burden that removes the proposed benefit. Set a practical limit on setup and support effort. Define who can halt the study and what happens to unfinished work. A negative result is valuable when it redirects development or prevents a poor investment. The programme should preserve that result, including disagreement and limitations, rather than repeatedly changing the measure until one favourable figure can be advertised.

The most useful conversation involves the people doing the work, an accountable clinical or organisational sponsor and someone who can decide what happens after the study. Bring the recurring problem, the current approach and the result that would change your view. We can discuss whether PHARMORIS offers a testable development proposition for that task. This is an invitation to define evidence with prospective partners; it is not a claim of an approved study, completed pilot, validated product or guaranteed route to procurement.