Two-stage single-arm trials are rarely reported adequately

摘要: Purpose: Two-stage single-arm trial designs are commonly used in phase II oncology to infer treatment effects for a binary primary outcome (e.g., tumour response). It is imperative that such studies be designed, analysed, and reported effectively. However, there little available evidence on whether this the case, particularly key statistical considerations. We therefore comprehensively review trials, examining particular quality of reporting. Methods: Published trials utilised "Simon's two-stage design" over 5 year period were identified reviewed. Articles evaluated they sufficient design details, as required sample size, analysis confidence interval (CI). The articles did not adjust their inference incorporation an interim re-analysed evaluate impact point estimate CI. Results: Four hundred twenty five results single arm included. Of these, 47.5% provided components ensure reproducibility. Only 1.2% 2.1% adjusted or CI, respectively. Just 55.3% final stage rejection bound, indicating many test hypothesis constructed assess. Re-analysis suggests estimates underestimated CIs too narrow. Conclusion: Key details often unreported. Whilst regularly performed, it rarely done so way removes bias introduced by analysis. In order maximise value, future must improve analysed reported.