Issues relating to selective reporting when including non‐randomized studies in systematic reviews on the effects of healthcare interventions
Abstract
Background: Selective outcome and analysis reporting (SOR and SAR) occur when only a subset of outcomes measured and analyzed in a study is fully reported, and are an important source of potential bias. Key methodological issues: We describe what is known about the prevalence and effects of SOR and SAR in both randomized controlled trials (RCTs) and non‐randomized studies (NRS), and the effects of SOR and SAR on summary effect estimates and conclusions in systematic reviews of the effectiveness of healthcare interventions. Guidance: Review authors should always suspect SOR and SAR in reviews that include NRS, assess primary studies for the risk of bias, and make reasonable attempts to retrieve study protocols or other documentation developed before study recruitment began. There are clues that may suggest SOR or SAR in NRS, including differences between the methods and results sections of the publication, study funder, and differences between study protocol or registration information and the study report. Conclusion: Existing evidence about reporting biases in primary studies comes almost exclusively from methodological reviews of RCTs. The prevalence and impact of SOR and SAR in NRS are likely even greater than in RCTs but it is difficult to identify and confirm selective reporting in NRS. Copyright © 2012 John Wiley & Sons, Ltd.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Sources of heterogeneity between studies can relate to design, risk of bias, or clinical characteristics of included study populations.
“Sources of heterogeneity between studies can relate to design (e.g. case-control studies vs control studies, different follow-up periods), to risk of bias (e.g. blinded vs unblinded outcome assessment) or clinical characteristics of included study populations (e.g. different age distributions between studies) [71, 98].”
From Methods to systematically review and meta-analyse observational studies: a systematic scoping review of recommendations (2018) · cites this paper as [71]
Many authoritative articles providing sound advice exist, such as those by Greenland, Egger et al., and a series in Research Synthesis Methods.
“Although many authoritative articles providing sound advice exist, such as those included in this review by Greenland [85], Egger et al. [29, 97], and a series in Research Synthesis Methods [37, 40, 65, 71, 75], the inexperienced reviewer may find them difficult to identify among the many other papers with contradictory advice.”
From Methods to systematically review and meta-analyse observational studies: a systematic scoping review of recommendations (2018) · cites this paper as [71]
The citation lists haven’t been fetched yet
References 39 · Cited by 50
- Authors
- Norris, S. L., Moher, D., Reeves, B. C., Shea, B., Loke, Y., Garner, S., Anderson, L., Tugwell, P., & Wells, G.
- Affiliations
- Oregon Health & Science University (US)
- Ottawa Hospital Research Institute (CA)
- Bristol Royal Infirmary (GB)
- Institute of Population and Public Health (CA)
- University of East Anglia (GB)
- National Institute for Health and Clinical Excellence London UK
- University of Washington (US)
- Corresponding author
- Norris, Susan L
- ORCID
- Venue
- Research Synthesis Methods
- Publisher
- Wiley
- Published
- 2012-11-13
- Volume
- 4
- Issue
- 1
- Pages
- 36-47
- ISSN
- 1759-2879 (ISSN-L)
- 1759-2887
- OpenAlex
- W1984170021
- Cited by
- 50
- References
- 39
- Citations by year
- 2026: 1 · 2023: 1 · 2022: 3 · 2021: 11 · 2020: 3 · 2019: 1 · 2018: 4 · 2017: 3 · 2016: 2 · 2015: 6 · 2014: 11 · 2013: 4
- FWCI
- 1.88
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Research Synthesis Methods (publishedVersion)
- PubMed (publishedVersion)
- Bristol Research (University of Bristol) (submittedVersion)
- Topics
- Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Keywords
- selective outcome reporting, reporting bias, non-randomized studies, systematic reviews, health interventions, risk of bias assessment, research registration
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
