Key points are not available for this paper at this time.
OBJECTIVE: Impact factor bias, in which positive results are preferentially published in higher-impact journals, may influence evidence visibility and interpretation. Whether such an impact factor bias exists in gynecologic oncology trials remains unclear. We evaluated the association between study outcome and journal impact factor in phase III gynecologic oncology randomized controlled trials. METHODS: A meta-epidemiologic analysis was conducted of phase III gynecologic oncology randomized controlled trials identified through ClinicalTrials.gov up to May 2024. Only completed, two-arm, superiority-design trials (as this allows for a clear and unambiguous interpretation of trial results) with published primary endpoint data were included. Data extracted for each publication included impact factor at the year of publication, number of enrolled patients, duration of follow-up, disease site, number of authors, country of first and senior author based on primary institutional affiliation, and primary outcome (classified as positive if the intervention arm demonstrated statistically significant superiority over the control arm). The primary outcome was the journal impact factor stratified by study results. Univariate and exploratory multivariable analyses were performed to identify factors associated with higher journal impact factors. RESULTS: A total of 36 eligible trials were identified, published between 2009 and 2024; 47.2% reported positive primary outcomes. The median journal impact factor across all studies was 35.1 (interquartile range; 18.3-51.6) (range; 3.7-158.5). Trials with positive outcomes were published in journals with significantly higher impact factors than those with negative outcomes (44.5 interquartile range; 18.7-71.5 vs 24.0 interquartile range; 5.3-44.5, p =.034). The number of authors was also greater in positive studies (median 21 vs 16, p =.032). In multivariable linear regression, a positive primary outcome was the only independently associated factor with a higher journal impact factor. Positive outcomes were published in journals with, on average, 32.6 (95% confidence interval 6.4-58.8) point higher impact factors, respectively. Follow-up duration and sample size were not significant predictors. CONCLUSIONS: Positive phase III gynecologic oncology trials were published in higher-impact journals, with outcome positivity independently associated with impact factor. This pattern may affect evidence visibility and highlights the need to support dissemination of high-quality negative studies.
Levin et al. (Sat,) studied this question.