Objectives: The aim of this study was to identify and quantify the characteristics of studies associated with the likelihood of publication.
Study design and setting: We searched for manuscripts that tracked cohorts of clinical studies ("cohorts") that from launch to publication. We explored the association of study characteristics with the probability of publication via traditional meta-analyses and meta-regression using random effects models.
Results: The literature review identified 85 cohorts of studies that met our inclusion criteria. The probability of publication was significantly higher for studies whose characteristics were favorable (odds ratio [OR] = 2.04; 95% confidence interval [CI]: 1.62, 2.57) or statistically significant (OR = 2.07; 95% CI: 1.52, 2.81), had a multicenter design (OR = 1.32; 95% CI: 1.16, 1.45), and were of later regulatory phase (3/4 vs. 1/2, OR = 1.34; 95% CI: 1.14, 1.49). Industry funding was modestly associated with lower (OR = 0.81; 95% CI: 0.67, 0.99) probability of publication. An exploratory analysis of effect modification revealed that the effect of the study characteristic "favorable results" on likelihood for publication was stronger for industry-funded studies.
Conclusion: The study characteristics of favorable and significant results were associated with greater probability of publication.
Keywords: Meta-analysis; Meta-epidemiology; Meta-regression; Publication bias; Result favorability; Statistical significance.
Copyright © 2017 Elsevier Inc. All rights reserved.