Automated Surveillance of Primary Outcome Reporting in Clinical Trials: Development and Validation of a Scalable Screening Pipeline

Version 2 (July 2026). Preprint of an automated pipeline that screens primary-outcome reporting across 46,365 completed interventional trials with posted results on ClinicalTrials.gov (2015-2025), comparing registered primary outcomes with published results and validating against manual review of 123 trial-publication pairs. This version corrects the screen's operating characteristics (sensitivity 16.7%, specificity 100%, overall agreement 95.9%), reframes automated flags as prompts for contextual review rather than confirmed discrepancies, adds the validation 2x2 table and a time-to-event registry-structure sensitivity analysis, and updates the figures. Flags denote prompts for review, not confirmed discrepancies. Code and data are available at https://github.com/hayden-farquhar/Outcome-Switching-Study