Availability of study protocols for randomized trials published in high-impact medical journals: A cross-sectional analysis
Abstract
Background: To improve reporting transparency and research integrity, some journals have begun publishing study protocols and statistical analysis plans alongside trial publications. We sought to assess the overall availability and characteristics of protocols and statistical analysis plans of randomized clinical trials published in the top five (by impact factor) general medicine journals. Methods: All randomized clinical trials published in Annals of Internal Medicine, BMJ, JAMA, Lancet, and NEJM in 2016 were identified. For each randomized clinical trial, we searched for protocols and statistical analysis plans on journal websites (including supplementary material) and in the article, for example, a referenced publication or link to trial or institutional website. Characteristics of randomized clinical trials were extracted from the publication and clinical trial registry. A detailed assessment of protocols and statistical analysis plans was conducted in a 20% random sample of randomized clinical trials. Results: Protocols were available for 299 (82%) trials, ranging from 50% in BMJ to >95% in NEJM and JAMA. Statistical analysis plans were available for 182 (50%) trials and varied from <10% for Annals of Internal Medicine, BMJ, and Lancet to 92% for NEJM. Of the 76 randomized clinical trials in the 20% random sample, 63 (83%) had a protocol but less than half (31; 44%) included an a priori (dated prior to patient enrollment) version of the protocol. Statistical analysis pl…
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?
We identified evidence of lack of clear specification of the primary outcome, primary assessment time and primary analysis method, again similar to that identified in individual randomized trials.
“We also identified evidence of lack of clear specification of the primary outcome, primary assessment time and primary analysis method, again similar to that identified in individual randomized trials [29].”
From Cluster randomized trials of individual-level interventions were at high risk of bias (2021) · cites this paper as [29]
The citation lists haven’t been fetched yet
References 27 · Cited by 37
- Authors
- Spence, O., Hong, K., Onwuchekwa Uba, R., & Doshi, P.
- Affiliations
- University of Maryland, Baltimore (US)
- Notre Dame of Maryland University, Baltimore, MD, USA
- Corresponding author
- Doshi, Peter
- Venue
- Clinical Trials
- Publisher
- SAGE Publications
- Published
- 2019-08-26
- Volume
- 17
- Issue
- 1
- Pages
- 99-105
- ISSN
- 1740-7745 (ISSN-L)
- 1740-7753
- OpenAlex
- W2970416360
- Cited by
- 37
- References
- 27
- Citations by year
- 2025: 10 · 2024: 4 · 2023: 5 · 2022: 6 · 2021: 4 · 2020: 8
- FWCI
- 3.02
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Clinical Trials (publishedVersion)
- PubMed (publishedVersion)
- Funders
- Laura and John Arnold Foundation
- Topics
- Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
- Social Sciences › Decision Sciences › Information Systems and Management › Academic Publishing and Open Access
- Life Sciences › Pharmacology, Toxicology and Pharmaceutics › Pharmacology › Pharmaceutical industry and healthcare
- Keywords
- statistical analysis plan, randomized controlled trials, research transparency, research integrity, clinical trial registries
- MeSH terms
- Clinical Trial Protocols as Topic, Cross-Sectional Studies, Humans, Periodicals as Topic, Publishing, Research Design, Models, Statistical, Randomized Controlled Trials as Topic, Journal Impact Factor
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
