Steps in Literature Review: A 10-Stage Process Flowchart from Question to Gap Statement, with Owners per Stage

The steps in a literature review with a gap analysis run from a defined question to a written gap statement. Figure 1 shows the literature review process as ten stages with the usual owner of each. The order is a working model; the review types you choose set how formal each stage must be [1].

Table 1 adds the output each stage hands to the next. The effort ratings are our planning estimates, relative rather than in hours.

The ratings follow a scoping review of resource use in systematic review production, which found that study selection, data extraction, critical appraisal and project management used the most resources, and literature searching among the least [3]. Plan reviewer time around stages 6 to 8.

Grant and Booth describe review types by four steps, Search, AppraisaL, Synthesis and Analysis (SALSA), and show how each type handles them [1]. A thesis background, a grant application and a publishable systematic review need different levels of rigour, so agree the type before writing a single search strategy line.

Our guide to types of literature review compares more formats.

Build each concept as a block of subject headings and free-text synonyms, combined with Boolean operators. Before the final run, a second searcher checks the strings against the PRESS 2015 checklist [2]; errors found after the search cost a full re-run. Record every string, date and record count in the search log, and add grey literature sources where the question needs them.

Screening and full-text appraisal are done by two reviewers independently, with disagreements resolved by discussion or a third reviewer. Keep counts as you go: for a systematic review they feed the PRISMA 2020 flow diagram [5]. Pilot the extraction sheet on two or three studies before extracting the rest.

There is no single published time for a literature review, because review types vary so much in scope [1]. Time depends mainly on how many records the search returns and how many studies are included (Table 2). Agree a stage-by-stage estimate once the record count is known, and revisit it after the first screening batch.

A research gap is a question the evidence cannot yet answer. Robinson and colleagues’ framework classifies why evidence is insufficient: insufficient or imprecise information, biased information, inconsistent results, or not the right information. It then characterises each gap by population, intervention, comparison, outcome and setting [4].

Apply it during evidence synthesis at stage 9, then write the gap statement at stage 10 in plain terms: what is known, what is not, why it matters, and which studies show it. A gap statement that cites no studies is an opinion, not a finding about research gaps.

• An unfocused question. A broad question inflates every later stage.
• Unreviewed search strings. Errors found after the search cost a full re-run.
• Screening backlog. Selection is among the heaviest stages; book two reviewers early.
• Gaps asserted without evidence. A gap statement should point back to the studies that show it.

Before submitting a narrative review, score it with our SANRA self-assessment.