Clinical trial registries are of minimal use for identifying selective outcome and analysis reporting

Susan L Norris; Haley K Holmer; Rongwei Fu; Lauren A Ogden; Meera S Viswanathan; Ahmed M Abou-Setta

doi:10.1002/jrsm.1113

Clinical trial registries are of minimal use for identifying selective outcome and analysis reporting

Res Synth Methods. 2014 Sep;5(3):273-84. doi: 10.1002/jrsm.1113. Epub 2014 Mar 11.

Authors

Susan L Norris¹, Haley K Holmer¹, Rongwei Fu¹, Lauren A Ogden¹, Meera S Viswanathan², Ahmed M Abou-Setta³

Affiliations

¹ Department of Medical Informatics and Clinical Epidemiology, Oregon Health & Science University, Portland, OR, 97239, USA.
² RTI International 3040 Cornwallis Road, PO Box 12194 Research Triangle Park, NC, 27709, USA.
³ George & Fay Yee Centre for Healthcare Innovation, University of Manitoba, 820 Sherbrook Street, Winnipeg, MB R3T 2N2, Canada.

PMID: 26052852
DOI: 10.1002/jrsm.1113

Abstract

Objective: This study aimed to examine selective outcome reporting (SOR) and selective analysis reporting (SAR) in randomized controlled trials (RCTs) and to explore the usefulness of trial registries for identifying SOR and SAR.

Study design and setting: We selected one "index outcome" for each of three comparative effectiveness reviews (CERs) of pharmacotherapy and extracted data on this outcome from trial registries and from study publications.

Results: Among 50 RCTs published since 2005 and reporting the index outcome, only 50% were listed in registries; 90% of RCTs were assessed as having SOR or SAR. The index outcome in the registry was different from that in the publication in 75% of trials in two CERs, and not specified at all in the third. Reported outcomes and analyses were not consistent between the publication's methods section and the results section in 33% and 46% of the two CERs where the index outcome was a benefit. There were no statistically significant predictors of SOR and SAR in our small sample where some predictors lacked variability.

Conclusion: The SOR and SAR were frequent in this pilot study, and the most common type of SOR was the publication of outcomes that were not pre-specified. Trial registries were of little use in identifying SOR and of no use in identifying SAR.

Keywords: bias; randomized controlled trials; selective analysis reporting; selective outcome reporting; systematic reviews; trial registries.

Publication types

Research Support, U.S. Gov't, Non-P.H.S.

MeSH terms

Bias
Outcome Assessment, Health Care / methods*
Outcome Assessment, Health Care / statistics & numerical data*
Pilot Projects
Randomized Controlled Trials as Topic / statistics & numerical data*
Registries / statistics & numerical data*
Reproducibility of Results
Research Design / statistics & numerical data*
Research Report*
Sensitivity and Specificity