We agree with Drs. Petersen, Braithwaite, and Torgeson1 that the CI values reported in our recent article provide valuable information to guide interpretation of our findings. As we state in the article, our results exclude large effects of anesthesia type on long-term outcomes but do not rule out more subtle effects.2
We also agree with Petersen et al.1 that avoiding selective enrollment is important for ensuring the generalizability of pragmatic trials. Within Regional versus General Anesthesia for Promoting Independence after Hip Fracture (REGAIN), we took extensive efforts to limit selective enrollment through site auditing and feedback and completed an ancillary study to understand reasons underlying individual clinicians’ decisions to permit or decline randomization for a given patient.3 Ultimately, 6% of exclusions in REGAIN were due to clinician discretion.4 We are reassured by this low rate; however, because we did not collect detailed data on excluded...
