We thank Zerr for the opportunity to discuss our process.1 We agree that the issue of participant selection in a trial is very important, in part for the reasons he stated. We expect authors to justify the selection of participants, to defend their eligibility criteria, and to discuss the generalizability of their findings. These issues are thoroughly addressed in CONSORT guidelines for nonpharmacological interventions,2–4 which PTJ requires authors to use for manuscripts reporting on RCTs. Indeed, subject variability is one of the many reasons why we rely extensively on CONSORT to aid in judging the credibility of a trial. Use of CONSORT—combined with a strong reviewer team of Editorial Board members and content reviewers—helps PTJ assure that the trial has a defensible internal and external validity. PTJ's RCT methods prescreen is accompanied by a content review (often by a different Editorial Board member), so that concerns related to eligibility criteria and participant selection will be considered from both angles.
All that said, we encourage Zerr to consider the “checks and balances” that are built into the RCT design itself. Assuming that randomization is done appropriately and there is ample distribution of baseline characteristics (age and others), the subject variability should largely be cancelled out. Undoubtedly, “tissue structure difference” and other differences may exist between age groups, but the random assignment of treatment should obviate most of these concerns.
We also note that the argument that the treatment effects may be more difficult to generalize in an RCT with a wide age range can work both ways. If there is a positive treatment effect in a wide age range, we might not know specifically whether to expect the same effects throughout the age range, but we do know that there is an effect. In contrast, with a very narrow age range, we have no data on the ages excluded from the trial.
- © 2012 American Physical Therapy Association
