EPID-32. IS THERE ANY EVIDENCE OF REPORTING AND CITATION BIAS AMONG COMPLETED PHASE III INTERVENTIONAL BRAIN TUMOR TRIALS?

Abstract

Abstract OBJECTIVE Reporting and citation bias based on positive versus negative outcome among completed phase III interventional brain tumor trials (IBTTs) have not been previously reported. Thus, we aimed to assess the evidence. METHODS Clinicaltrials.gov was used to obtain all phase III IBTTs completed prior to December 31st 2016. Trials closed due to poor accrual or non-phase III were excluded. Data about the funding source, type of intervention, conducted at U.S or Non-U.S locations, patients enrolled, primary completion date, time to dissemination of results in months (either reported on clinicaltrials.gov or published in a journal), citations per year (from web of science) and outcome (positive or neutral/negative) were abstracted. Median time to dissemination was estimated using Kaplan-Meier estimates and a log rank test for statistical significance. The number of citations per year for positive and negative/neutral studies was compared using a t-test. RESULTS 65 studies were analyzed. The median time from primary completion to dissemination of results for trials with a positive outcome was 20 months (n = 20; 95% CI: 12–31) compared to 30 months for trials with a negative outcome (n=35; 95 % CI: 22 – 37) (p = 0.095). The average number of citations per year for trials with a positive outcome is 62.4 (n = 21; range 1.1 – 614.8) compared to 25.0 for trials with a negative outcome (n=35; range 0.14 - 158.8) (p = 0.213). There was no significant difference in time to dissemination based on the funding source, type of intervention or the location where the trial was conducted. CONCLUSION Positive phase III IBTTs results are disseminated earlier and are cited more frequently compared to negative trials. However, there is no statistically significant evidence of reporting or citation bias.