Clinical evidence gap analysis for medical devices: find the weak link first

A gap is always relative to a claim

‘We need more clinical evidence’ is too vague to guide work. The missing evidence depends on the device’s intended purpose, target population, user, clinical setting, risk profile and the claims being made. Change one of those and the gap can change with it.

The first task is therefore alignment. Put the intended purpose, indications, contraindications, user groups, clinical benefits, safety claims and performance claims in one place. Then ask what evidence each statement requires and whether the current sources answer that exact question.

A focused literature review can organise the existing evidence and expose indirect support; it does not replace a formal clinical evaluation or specialist sign-off.

Directness matters as much as volume

A large literature can still be indirect. Studies may concern an earlier device generation, a related product with meaningful technical differences, a different population, a surrogate endpoint, or a setting unlike intended use. Each difference needs an argument, not an assumption.

European guidance on sufficient clinical evidence and FDA guidance on real-world evidence both make relevance central. Evidence has to be fit for the regulatory question. Reliability matters too: data provenance, completeness, measurement, analysis and bias affect whether a source can carry weight.

Create a claim-evidence matrix

For every material claim, record:

• The required clinical benefit, safety or performance proposition.
• The available clinical, preclinical, usability, post-market and real-world sources.
• Directness to the current device and intended use.
• Methodological strengths and limitations.
• Whether the source supports, contradicts or does not address the claim.
• The residual uncertainty and consequence if unresolved.

This matrix stops a common failure: discussing evidence by document rather than by claim. A polished clinical report may contain many sources while leaving one commercially important claim unsupported.

Separate evidence gaps from documentation gaps

Sometimes the evidence exists but the chain is not documented. The rationale linking an equivalent device, a technical feature or a clinical outcome may be scattered across teams. That is a documentation gap. Other times no suitable data answer the question. That is an evidence gap.

The remedies differ. Better synthesis and traceability may fix the first. The second may require a focused literature update, usability work, post-market data, a clinical investigation, or a narrower claim. Calling both problems ‘more evidence’ can lead to unnecessary studies or false reassurance.

Prioritise by decision impact

Not every gap deserves the same response. Rank gaps by the importance of the claim, the device risk, the directness and quality of current support, the likelihood that new evidence changes the conclusion, and the time and cost of obtaining it.

A high-priority gap is not simply a blank cell. It is a dependency: if the claim fails, the intended purpose, benefit-risk case, adoption story or valuation may change. Making that dependency explicit helps leadership decide whether to generate evidence, narrow the scope or stop.

What the analysis should and should not promise

A rapid gap analysis can clarify the landscape, identify weak links, organise existing evidence and define the next research questions. It can support planning, diligence and preparation for specialist regulatory work.

It should not present itself as a completed clinical evaluation or regulatory determination unless it follows the required process and is produced by people with the relevant authority and expertise. The honest output is a map: what appears supported, what remains indirect, what is missing, and what to do next.

The next step should be specific

End with an evidence action plan, not ‘conduct more research.’ Name the claim, the uncertainty, the source or study design most likely to reduce it, and the decision that result would unlock. Good gap analysis reduces work by aiming the next piece of evidence at the actual weak link.