There is no single NHS evidence checklist that applies to every medical technology. The evidence required depends on the product, claim, care setting, current pathway, decision-maker, budget route and consequence of adoption.
The practical question is not simply, “Do we have evidence?” It is, “Do we have the evidence required for the next person or organisation to make a confident decision?”
Begin with the adoption decision
A technology may need several different decisions before it becomes routine practice: a clinician may decide to use it, a service leader may approve a pathway change, a trust may approve governance, procurement may select a route and a budget holder may agree to fund it.
Each decision has a different evidence threshold. A pilot sponsor may accept uncertainty that a procurement board cannot. A clinician may focus on diagnostic or therapeutic value, while an operational lead needs to understand staffing, training and workflow.
Five evidence domains usually matter
1. Product and regulatory evidence
The product must have an appropriate regulatory route, quality foundation and evidence supporting the intended purpose and claims. This is the essential baseline, but it does not by itself demonstrate that the NHS pathway should change.
2. Clinical evidence
Decision-makers need to understand performance in the relevant population and setting. The important endpoint may be diagnostic accuracy, treatment effect, safety, decision change, referral, escalation, admission or another clinically meaningful outcome.
Evidence is stronger when the study population, comparator, workflow and endpoint reflect the intended NHS use rather than a convenient research setting.
3. Workflow and usability evidence
A product can be clinically valuable but operationally unusable. Evidence may be needed on training, task time, failure modes, sample handling, result interpretation, digital integration and the effect on existing roles.
This is particularly important where use moves across settings or into the hands of staff with different training and competing priorities.
4. Economic and capacity evidence
Economic claims should show where costs and benefits occur, over what time period and for which organisation. A saving to the wider system may not create a usable business case for the budget holder expected to pay.
Relevant measures can include staff time, bed use, referrals, investigations, antibiotic use, avoided appointments, length of stay, conveyance, admissions or capacity released for other activity.
5. Implementation and equity evidence
Adoption also depends on whether the intervention can be introduced safely and consistently. Implementation evidence can cover training, governance, data, supply, maintenance, service ownership, access and variation between patient groups or sites.
Regulatory evidence and adoption evidence are connected
The two evidence systems should not be developed independently. Intended use and product claims influence the regulatory evidence plan. The pathway and adoption proposition influence which claims are commercially meaningful.
If the adoption case depends on an operational or economic outcome that is not connected to the product evidence, the company may reach approval without a persuasive route to use. Conversely, a broad market-access claim can create regulatory or study demands that are disproportionate to the product stage.
Design pilots around a next decision
A pilot should not be defined only by the number of products deployed or users trained. It should state:
- which adoption uncertainty the pilot will address;
- who owns the evaluation and the decision that follows;
- which outcomes and process measures will be collected;
- what result would support continuation, amendment or stop;
- how the evidence will inform procurement, commissioning or scale.
This changes a pilot from activity into a decision-producing programme.
Build an evidence ladder
Most companies cannot answer every question at once. A staged evidence ladder is often more useful: establish technical and clinical credibility, demonstrate usability and workflow fit, test the economic assumptions, then generate the implementation evidence required for wider use.
The sequence should reflect risk and the next access decision. This keeps evidence generation proportionate while ensuring that each study makes the next step more credible.