Dear all,
I am involved in a systematic review which happens to include only
crossover studies.
I have three problems:
i) A carryover effect is unlikely, since the topic is on the effect of a
short-acting drug (inhaled bronchodilators). However, I don't know how I
could assess the carryover effect when the authors failed to report
about it.
ii) We don't have individual data, just means and SD, so we cannot use
paired statistics in computing the SD of the difference. As a result,
the treatment effect for some studies results to be not-significant,
even if the authors claim an highly significant difference.
iii) I could not find very much to read on this topic, except some
studies on fertility (where the carryover problem is probably much more
relevant) or papers on the analysis of individual studies (rather than
systematic reviews). These problems should be fairly common, though, so
maybe I failed to look in the right places.
Can somebody help me help me sorting out these issues, especially the
latter?
Many thanks,
Piersante Sestini
Institute of Respiratory Diseases
University of Siena, Italy
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|