How to Verify a Research Gap Before Starting a Review

A published priority is a starting point

A guideline can document that a question mattered to its committee. It cannot establish that the question remains unanswered years later. Before starting a review, separate three claims: a research priority was published; uncertainty remains after more recent evidence; and your proposed project will reduce that uncertainty. Each requires different evidence.

Our Research Opportunity Observatory makes this distinction explicit. Its source checks establish specific published facts. They do not certify novelty or replace a systematic search.

Write the question so overlap can be tested

Consider a fictional team proposing a review of rehabilitation after a respiratory infection. “Does rehabilitation work?” leaves the population, intervention, comparator, outcome and timing unspecified. One existing review may study hospital survivors, another community patients, and another a specific symptom group. Similar titles can conceal different questions; different titles can cover nearly identical evidence.

Write the intended population and setting, the intervention or exposure, comparator, design, outcome, and follow-up window. For a diagnostic or prognostic question, adapt the framework rather than forcing it into an intervention question. Record why each restriction is clinically or methodologically justified.

Build an overlap table

Use one row for each relevant review or protocol. Include its objective, eligible designs, populations, interventions, outcomes, last search date, included studies, registration and publication status. Read the methods and supplements before relying on its title or conclusion.

A fictional review published in 2026 could have searched only through 2023. That creates an interval to investigate, not automatic justification for an update. Conversely, an older publication may belong to a living review with a much newer version. Record the version you actually read.

The PRISMA 2020 checklist calls for clear reporting of information sources, search dates, search strategies and registration information. Those reporting principles help make an overlap check auditable; completing a checklist does not itself make the search comprehensive.

Check work that has not reached a journal

Search relevant protocol and study registries as well as bibliographic databases. Save the query, platform, date, filters, and records assessed. A protocol is evidence that someone planned work, not proof that it was completed or abandoned. Check linked publications and updates, and contact investigators when clarification is essential and feasible.

Do not interpret a search-engine result count, an absent exact-title match, or a single database search as proof that a question is new. Important records may use different terminology or be poorly indexed.

Make a decision before extracting everything

Proceed when a defensible contribution and feasible evidence base are documented. Narrow the question when existing work leaves a meaningful population, outcome or methodological issue unresolved. Stop when the project would merely reproduce a current, adequate synthesis. If reporting is too sparse, an evidence map or primary study may be more useful than a pooled estimate.

Write down both the decision and its limits. “No directly matching review identified in the sources searched on this date” is a bounded statement. “No one has studied this” usually exceeds the evidence.

Download the verification worksheet, use the pooling planner, or request a research idea feasibility check.