Examples of Search Terms from Published Systematic Reviews: The Structure of Multi-Concept Search Strategies

Reading a few real, published search strategies is an efficient way to understand how search terms are constructed for a systematic review, and is more concrete than an abstract description of Boolean logic. Three large reviews with clearly reported methods, covering physical activity and chronic disease, greenspace exposure and health, and sedentary behavior and mortality, show the same underlying pattern even though they cover different topics: each combines a small number of core concepts rather than trying to capture every possible synonym in a single flat list. Close examination of how each review's concept structure maps onto its specific research question also shows why that structure cannot simply be copied wholesale from one review to another.

Example one: physical activity and chronic disease

Kyu and colleagues' Global Burden of Disease dose-response meta-analysis pooled 174 prospective cohort studies examining physical activity against five chronic-disease outcomes: breast cancer, colon cancer, diabetes, ischemic heart disease, and ischemic stroke1. Structurally, a review of this kind needs its search to combine an exposure concept built around physical activity and its synonyms (exercise, leisure-time activity, physical activity level) with a study-design concept restricting results to prospective cohort studies, and an outcome concept covering each disease separately, since searching for "physical activity" alone across a general database would return an unmanageable and largely irrelevant volume of results2. The review's scale, 174 included cohorts, is itself evidence that the underlying search had to be built broadly enough to capture studies using different terminology for the same exposure and outcome constructs, across several decades of published cohort literature using inconsistent activity-measurement instruments.

Example two: greenspace exposure and health outcomes

Twohig-Bennett and Jones's systematic review and meta-analysis of greenspace exposure, published in Environmental Research, pooled 143 studies covering a wide range of health outcomes associated with exposure to green or natural environments3. A search strategy supporting this kind of review has to combine an exposure concept (greenspace, green space, natural environment, parks) with a broad outcome concept spanning multiple health domains, because the review's own scope, "wide-ranging health benefits," meant the outcome side of the search could not be narrowed to a single condition the way a disease-specific review's outcome concept can3. This illustrates a design trade-off: a broader outcome concept increases sensitivity but also increases the volume of studies that need screening by hand afterward, which this review's own scale, 143 included studies drawn from a far larger pool of initial search results, reflects directly.

Example three: sedentary behavior and mortality

Ekelund and colleagues' harmonised meta-analysis, published in The Lancet, pooled individual-level data from more than one million men and women to examine whether physical activity attenuates the association between sitting time and mortality4. Because this review examined an interaction between two exposures (sedentary time and physical activity level) against a single outcome (all-cause mortality), its underlying search concepts needed to capture both exposure constructs together, sitting, sedentary behavior, screen time on one side, and physical activity, exercise on the other, combined with a mortality outcome concept and, again, a prospective cohort design filter4. This is a more complex concept structure than either of the first two examples, reflecting the review's more complex research question, and it required the authors to negotiate individual participant data sharing agreements across dozens of cohort studies rather than relying solely on published summary estimates.

Features common to the three examples

Across all three reviews, the number of distinct concepts combined stayed small, generally three or four, even though each concept itself expanded into several synonyms and controlled-vocabulary terms25. This matches the general guidance in the literature-search methodology published by Aromataris and Riitano, which recommends identifying a small number of key concepts drawn directly from the research question before expanding each one into a list of synonyms and subject headings, rather than trying to search the whole question as a single phrase5. A study-design filter, restricting to prospective cohorts, randomized trials, or another defined design, also appeared as one of the core concepts in two of the three examples, indicating that methodology terms are not optional but constitute one of the standard building blocks of a systematic search, and one that disproportionately improves precision relative to the effort of adding it. None of the three reviews appears to have relied on a single database either; each drew on the standard combination of core biomedical sources supplemented by discipline-specific ones where the topic called for it, which is itself a practice worth adopting even when the exact source list needs to differ for a new topic.

Adaptation of concept structures between reviews

Although all three examples share the same general shape, a small number of combined concepts, the specific concepts and their synonym lists are tied tightly to each review's own question, and simply transplanting one review's string onto a different topic produces a search that misses the new topic's relevant terminology entirely. Bramer and colleagues' systematic approach to search development makes this point explicitly: concept identification has to happen fresh for each new question, because even superficially similar topics, physical activity and chronic disease versus physical activity and depression, for instance, draw on partly different outcome terminology and partly different relevant databases6. What can be reused across reviews is the general method, break the question into a small number of concepts, build a synonym list for each, add a design filter where relevant, not the specific strings themselves.

Interpreting the search appendix of a published review

Reading one of these three reviews' published search appendices with a new project in mind requires looking beyond the final string to the decisions behind it: which concepts were treated as mandatory versus advisory, which controlled-vocabulary terms were included alongside free-text synonyms, and which databases were searched beyond the obvious core set6. The PRISMA-S reporting standard is useful here from the reader's side as well as the author's, since a review that follows it will document exactly these decisions in enough detail to reconstruct the logic, rather than leaving a reader to infer it from a single collapsed search string7. When a published appendix does not meet this standard, cross-referencing the Cochrane Handbook's own worked examples alongside it is a reasonable way to fill the gaps, since the Handbook documents the general logic that a well-built strategy in the same subject area would be expected to follow8.

Limitations of published search structures as templates

A published review's search strategy reflects decisions made for that specific review's scope, resources, and timeline, and copying its concept structure wholesale can mislead a new reviewer in at least two ways. First, a large, well-resourced review like Kyu and colleagues' Global Burden of Disease analysis could afford to search broadly and screen a very large volume of results by hand, a trade-off that may not be realistic for a smaller team working to a tighter timeline, so the transferable lesson is the concept-identification method rather than the exact breadth of the search itself1. Second, a review's published search strategy reflects the state of controlled vocabulary and indexing at the time it was run; MeSH terms and database interfaces are periodically revised, so a term that worked well in a strategy published several years ago may need checking against the current version of the relevant thesaurus before being reused today2. A new reviewer borrowing the general shape of one of these three examples should therefore treat it as a starting hypothesis about which concepts matter, to be tested against a small number of pilot searches in the target databases, rather than as a finished template to fill in with substituted keywords and run unchanged. Treating a published search appendix this way, as a hypothesis to test rather than an answer to copy, also protects against inheriting a predecessor review's blind spots, since any gap in the original search carries forward silently into a new project that reuses its structure without independently re-examining it.

Key points

Practical application

Frequently asked questions

@misc{/la_search_terms_real_examples_2026,
  author       = {{Pubrica Academy}},
  title        = {Examples of Search Terms from Published Systematic Reviews: The Structure of Multi-Concept Search Strategies},
  howpublished = {Pubrica},
  year         = {2026},
  month        = {oct},
  url          = {https://pubrica.com/academy/publication-support/examples-of-search-terms-from-published-studies/},
  note         = {Accessed: <span data-cite-acc="iso">2026-10-15</span>}
}

Research question: ____________________
Concept 1 (exposure/intervention) + synonyms: ____________________
Concept 2 (population, if distinct) + synonyms: ____________________
Concept 3 (outcome) + synonyms: ____________________
Concept 4 (study design filter, if needed): ____________________
Boolean combination plan: (Concept1 OR synonyms) AND (Concept2...) AND (Concept3...) AND (Concept4...)