Secondary endpoint utilization and publication rate among phase III oncology trials.

Secondary endpoints (SEPs) provide crucial information in the interpretation of clinical trials, but their features are not yet well understood. Thus, we sought to empirically characterize the scope and publication rate of SEPs among late-phase oncology trials. We assessed SEPs for each randomized, published phase III oncology trial across all publications and ClinicalTrials.gov, performing logistic regressions to evaluate associations between trial characteristics and SEP publication rates. After screening, a total of 280 trials enrolling 244,576 patients and containing 2,562 SEPs met inclusion criteria. Only 22% of trials (62/280) listed all SEPs consistently between ClinicalTrials.gov and the trial protocol. The absolute number of SEPs per trial increased over time, and trials sponsored by industry had a greater number of SEPs (median 9 vs 5 SEPs per trial, P < 0.0001). In total, 69% of SEPs (1770/2562) were published. The publication rate significantly varied by SEP category Χ2(5, N = 2562) = 245.86, p < 0.001. SEPs that place the greatest burden on patients, patient-reported outcomes and translational correlatives, were published at 63% (246/393) and 44% (39/88) respectively. Trials with more SEPs were associated with lower publication rates. Overall, our findings are that SEP publication rates in late-phase oncology trials are highly variable based on type of SEP. To avoid undue burden on patients and promote transparency of findings, trialists should weigh the biologic and clinical relevance of each SEP with its feasibility at the time of trial design.