Contents
Academy · Research Data, Analytics & AI · Systematic Reviews of Prospective Studies
Conducting a Systematic Review of Prospective Studies: Methods and Reporting
A systematic review of prospective studies follows a fixed sequence of question framing, protocol, search, screening, extraction, risk-of-bias assessment and synthesis, and PICO has been shown to be more sensitive than SPIDER for this type of quantitative search.
A systematic review of prospective studies combines and critically appraises findings from research that followed participants forward in time from a defined exposure or intervention to a measured outcome, with the aim of identifying cause-and-effect associations, risk factors, or prognostic variables more reliably than any single study can on its own. Because prospective designs are the main source of evidence for questions about risk and prognosis, a review built on them carries particular weight in clinical and policy decisions, which is also why its methodology should follow a defined, auditable sequence rather than resemble an informal literature summary presented as a review.
Learning objectives
- The fixed sequence of steps that a systematic review of prospective studies should follow
- Why PICO is more sensitive than SPIDER for most searches for prospective studies
- How risk-of-bias assessment for non-randomized prospective studies differs from assessing randomized trials
Common misconceptions
- “Any structured question framework works equally well for finding prospective studies.” A direct comparison of search frameworks found PICO (Population, Intervention, Comparator, Outcome) more sensitive than SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research type) for quantitative and mixed-methods searches, while SPIDER was better suited to smaller, more precise qualitative searches, meaning the right framework depends on the kind of prospective evidence being sought.
Question framing and protocol development
The first step is a precise, structured question; a direct comparison of common frameworks found PICO more sensitive than SPIDER for quantitative and mixed-methods literature searches, while SPIDER performed better for smaller, more targeted qualitative searches, which means that a review built mainly on prospective cohort or trial data should generally frame its question with PICO rather than SPIDER3. Five steps to conducting a systematic review, a widely used methods paper, presents protocol-writing as the step that should immediately follow question-framing, covering objectives, eligibility criteria, search strategy, data extraction plan, and synthesis method, written and, where possible, registered before the search is run2. Eligibility criteria should specify study design (prospective cohort, prospective case-control, or prospective registry-based), population, exposure or intervention, outcome, and a publication date range, set in advance specifically to reduce the risk that criteria are adjusted later according to which studies the search happens to retrieve1.
Searching and screening against pre-specified criteria
A comprehensive search covers multiple databases — typically PubMed, Embase, and Web of Science at minimum — together with a check of the reference list of each included study, since database coverage of the same literature varies sufficiently that single-database searches routinely miss eligible studies12. Screening proceeds in two stages — titles and abstracts first, against the pre-set eligibility criteria, then the full texts of the remaining records — with common exclusions at either stage including irrelevant study designs, duplicate publications, studies with unavailable full texts, and abstract-only conference records that have not undergone full peer review1. The updated PRISMA 2020 flow diagram, revised specifically to accommodate reviews searching multiple source types including trial registers and organizational websites, is the standard means of reporting how many records were identified, screened, and excluded at each stage, and completing it accurately as screening proceeds, rather than reconstructing it afterward, substantially reduces the risk of a reporting error at submission4.
Extracting data and assessing risk of bias in prospective designs
Data should be extracted using a standardized form covering design, population, sample size, exposure definition and measurement, outcome definition and measurement, follow-up duration, and reported effect estimates with confidence intervals, piloted on a small sample of included studies before full extraction begins1. Risk-of-bias assessment for prospective, non-randomized studies should use a tool built for that design rather than one built for randomized trials: ROBINS-I, the Cochrane-endorsed tool for assessing risk of bias in non-randomized studies of interventions, evaluates confounding, participant selection, exposure measurement, and outcome ascertainment specifically, while the Newcastle-Ottawa Scale offers a star-rating alternative widely used for cohort studies in particular9. A review that applies a randomized-trial risk-of-bias tool, such as RoB 2, to a set of prospective observational cohort studies is using an instrument inappropriate to the design, since the domains of RoB 2 — randomization process and deviation from intended interventions chief among them — do not map onto a design that was never randomized.
Synthesis of findings and reporting against PRISMA
Synthesis proceeds either through meta-analysis, in which effect estimates are statistically pooled, or through narrative synthesis, in which findings are summarized qualitatively because pooling is not appropriate owing to high heterogeneity or incompatible outcome measures across studies; the choice between the two should be stated and justified rather than determined by which approach is more familiar to the review team1. The final report should interpret findings in light of the strengths and limitations of the included studies, state implications for practice, policy, and future research, and follow the full 27-item PRISMA 2020 checklist for reporting, since adherence to that checklist is independently associated with more complete and reproducible review reporting across the literature45. Because evidence accumulates over time, reviews addressing rapidly developing questions should also state a planned update interval or adopt a living systematic review format, rather than being treated as a permanent, one-time answer to a question that new prospective studies will continue to inform10.
Common pitfalls in reviewing evidence from prospective studies
A recurring pitfall is pooling or narratively combining prospective cohort studies that measured the same nominal outcome over substantially different follow-up durations without accounting for that difference, since a cohort followed for one year and a cohort followed for ten years are not estimating the same quantity even when both report, for example, an “incidence rate” for the same condition1. A second common pitfall is treating confounder adjustment as binary — adjusted versus unadjusted — rather than recording exactly which confounders each study adjusted for, since two “adjusted” estimates controlling for different confounder sets are not directly comparable, and a review that does not record this detail cannot adequately explain heterogeneity it later observes between studies6. A third pitfall specific to prospective designs is underestimating the impact of loss to follow-up: a cohort study with high attrition can produce a biased effect estimate if participants who withdrew differed systematically from those who remained, and a systematic review should extract and report the attrition rate of each included study and, where available, how the original study handled or adjusted for it, rather than treating attrition as a minor methodological footnote9. Reviews that build these three checks — matched follow-up duration, recorded confounder sets, and documented attrition — into their extraction and risk-of-bias process from the outset produce substantially more defensible syntheses than reviews that identify these gaps only once pooling has already been attempted.
A fourth, more subtle pitfall is reverse causation in prospective data that is nonetheless treated as settling a causal question outright: even a well-conducted prospective cohort can be affected by an exposure and outcome sharing a common, earlier cause, or by subclinical disease at baseline that influenced both the measured exposure and the later outcome, and a systematic review should extract and report how, or whether, each included study addressed this possibility, for example through a run-in period excluding early events or through sensitivity analyses lagging the exposure measurement9. Reviews that summarize prospective findings as straightforwardly causal without checking for and reporting this kind of design safeguard risk overstating what the underlying evidence supports, a distinction that is particularly important when the conclusions of a review are intended to inform clinical guidance or policy.
A fifth pitfall that merits separate consideration is combining prospective studies that measured an exposure at a single baseline point with studies that measured exposure repeatedly across follow-up, since a single-baseline measurement can misclassify participants whose exposure status changed during the study period, generally biasing an estimated association toward the null relative to a design that captured exposure changes over time1. Extracting and reporting the exposure-measurement frequency of each included study — single baseline versus repeated measures — allows a review to assess whether this design difference plausibly explains part of any heterogeneity observed between pooled estimates, rather than leaving that heterogeneity unexplained or incorrectly attributing it to clinical differences between study populations.
Taken together, these five pitfalls — mismatched follow-up duration, inconsistently recorded confounder sets, unreported attrition, unaddressed reverse causation, and mixed exposure-measurement frequency — share a common root cause: each is a structural feature of the underlying prospective studies that would be missed by an examination of the final effect estimate alone, which is why a disciplined, field-level extraction process is as important as the statistical synthesis that follows it.
Key points
Key points
A systematic review of prospective studies follows a fixed sequence: a PICO-framed question and registered protocol, a multi-database search with two-stage screening, standardized extraction with a design-appropriate risk-of-bias tool such as ROBINS-I, and synthesis reported against the full PRISMA 2020 checklist.
Practical application
Practical application
- Questions for quantitative reviews of prospective studies should be framed with PICO rather than SPIDER. PICO has been shown to be more sensitive for this type of search; SPIDER is better reserved for smaller, more targeted qualitative reviews.
- ROBINS-I or the Newcastle-Ottawa Scale, rather than a randomized-trial tool, should be used for non-randomized prospective studies. A randomized-trial risk-of-bias tool's domains do not map onto a design that was never randomized.
- The extraction form should be piloted, and the PRISMA flow diagram completed as screening proceeds. Reconstructing either retrospectively is considerably more error-prone than completing both concurrently.
Frequently asked questions
Not always. A direct comparison found PICO more sensitive for quantitative and mixed-methods searches, but SPIDER performed better for smaller, more targeted qualitative searches, so the choice depends on the kind of evidence being sought.
Yes. It is common for a review to meta-analyze a subset of studies with compatible outcome measures while narratively synthesizing studies that cannot be statistically pooled due to heterogeneity or incompatible measures.
ROBINS-I or the Newcastle-Ottawa Scale, both designed specifically for non-randomized study designs, rather than tools such as RoB 2 that are designed for randomized trials.
Guidance generally recommends at least PubMed, Embase, and Web of Science as a minimum, supplemented by reference-list checks of included studies, since no single database fully covers the relevant literature.
Fast-moving topics benefit from a stated update interval or a living-review format, since new prospective studies will continue to be published after the review's search date.
References
- Tawfik GM, Dila KAS, Mohamed MYF, et al. A step by step guide for conducting a systematic review and meta-analysis with simulation data. Tropical Medicine and Health. 2019;47:46. read 10 October 2026. View on Tropical Medicine and Health
- Khan KS, Kunz R, Kleijnen J, Antes G. Five steps to conducting a systematic review. Journal of the Royal Society of Medicine. 2003;96(3):118-121. read 10 October 2026. View on JRSM
- Methley AM, Campbell S, Chew-Graham C, McNulty A, Cheraghi-Sohi S. PICO, PICOS and SPIDER: a comparison study of specificity and sensitivity in three search tools for qualitative systematic reviews. BMC Health Services Research. 2014;14:579. read 10 October 2026. View on BMC Health Services Research
- Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. read 10 October 2026. View on The BMJ
- Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Medicine. 2009;6(7):e1000097. read 10 October 2026. View on PLOS Medicine
- Li T, Higgins JPT, Deeks JJ, eds. Chapter 5: Collecting data. In: Cochrane Handbook for Systematic Reviews of Interventions. read 10 October 2026. View on Cochrane Training
- Munn Z, Moola S, Lisy K, Riitano D, Tufanaru C. Chapter 7: Systematic reviews of etiology and risk. In: JBI Manual for Evidence Synthesis. read 10 October 2026. View on JBI Global
- Wells GA, Shea B, O'Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Hospital Research Institute. read 10 October 2026. View on OHRI.ca
- von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. The Lancet. 2007;370(9596):1453-1457. read 10 October 2026. View on The Lancet
- Sterne JA, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919. read 10 October 2026. View on The BMJ
Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.
How to cite this article
Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.
Pubrica Academy. (2026, October 10). Conducting a Systematic Review of Prospective Studies: Methods and Reporting. Pubrica. https://pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/
Pubrica Academy. “Conducting a Systematic Review of Prospective Studies: Methods and Reporting.” Pubrica, 10 Oct. 2026, pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/.
Pubrica Academy (2026) Conducting a Systematic Review of Prospective Studies: Methods and Reporting. Pubrica. Available at: https://pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/ (Accessed: 15 October 2026).
Pubrica Academy. “Conducting a Systematic Review of Prospective Studies: Methods and Reporting.” Pubrica. October 10, 2026. https://pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/
Pubrica Academy, ‘Conducting a Systematic Review of Prospective Studies: Methods and Reporting’ (Pubrica, 10 October 2026) <https://pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/> accessed 15 October 2026.
Pubrica Academy. Conducting a Systematic Review of Prospective Studies: Methods and Reporting. [Internet]. Pubrica; 2026 Oct 10 [cited 2026 Oct 15]. Available from: https://pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/
Pubrica Academy, “Conducting a Systematic Review of Prospective Studies: Methods and Reporting,” Pubrica, Oct. 10, 2026. [Online]. Available: https://pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/
@misc{/la_systematic_review_prospective_studies_steps_2026,
author = {{Pubrica Academy}},
title = {Conducting a Systematic Review of Prospective Studies: Methods and Reporting},
howpublished = {Pubrica},
year = {2026},
month = {oct},
url = {https://pubrica.com/academy/manuscript-editing/conduct-a-systematic-review-for-prospective-studies/},
note = {Accessed: <span data-cite-acc="iso">2026-10-15</span>}
}This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.
Systematic Review of Prospective Studies Checklist
This sequence should be followed when reviewing prospective-study evidence.
[ ] Framed the question with PICO (or SPIDER, if primarily qualitative) [ ] Wrote and registered the protocol before searching [ ] Searched at least 3 databases plus reference lists of included studies [ ] Screened titles/abstracts, then full texts, against pre-set criteria [ ] Extracted data with a piloted, standardized form [ ] Assessed risk of bias with ROBINS-I or Newcastle-Ottawa Scale (not a randomized-trial tool) [ ] Chose and justified meta-analysis vs. narrative synthesis [ ] Reported against the full PRISMA 2020 checklist
Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.
How Pubrica can help
Pubrica’s evidence synthesis team conducts this full sequence for reviews of prospective-study evidence, from PICO question framing to PRISMA-compliant reporting, with design-appropriate risk-of-bias assessment throughout.