 |
 |
Changing subject for different discussion thread. If intention-to-treat analysis (with various assumptions) and as-treated analysis all give consistent results then the degree of crossover may not bias conclusions. If different analyses give different results, then it takes more interpretation to sort out what we think is really happening and to be cautious with respect to checking our assumptions in our explanations for the results. Others on the list may have more sophisticated (statistical) answers for this. Brian S. Alper, MD, MSPH, FAAFP Editor-in-Chief, DynaMed (dynamed.ebscohost.com) -----Original Message----- From: Djulbegovic, Benjamin [mailto:[log in to unmask]] Sent: Thursday, July 04, 2013 10:30 AM To: Brian Alper MD Cc: [log in to unmask] Subject: Re: Critical appraisal criteria for randomized crossover trials Indeed, but how should we manage "crossover that occurs in randomized parallel trials", which has become so rampant in cancer trials...I realize that your question is a narrow one in the context of "true" cross-over trials, and while it is always good to be more precise in our definitions, I would also be interested in hearing about the broader views on cross-over issue...for one, when the cross-over occurs in parallel RCT,is there anything else that can be done to minimize bias outside of the intention-to-treat analysis ( and the standard adjustments for confounders employed in traditional observational studies)? b Sent from my siPhone (Please excuse typos & brevity) On Jul 4, 2013, at 10:05 AM, "Brian Alper MD" <[log in to unmask]> wrote: > how to manage crossover that occurs in randomized parallel trials