Premium
Gender differences in randomised, controlled trials in intensive care units
Author(s) -
KRISTENSEN M. L.,
VESTERGAARD T. R.,
BÜLOW H.H.
Publication year - 2014
Publication title -
acta anaesthesiologica scandinavica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.738
H-Index - 107
eISSN - 1399-6576
pISSN - 0001-5172
DOI - 10.1111/aas.12337
Subject(s) - medicine , randomized controlled trial , intensive care unit , dominance (genetics) , population , mean difference , intensive care , significant difference , demography , intensive care medicine , confidence interval , surgery , biochemistry , chemistry , environmental health , sociology , gene
There is a male dominance among patients in intensive care units ( ICU s). Potentially, this will increase the risk of a skewed male/female distribution in randomised, controlled trials ( RCT s). We have evaluated if this has in fact happened when randomising and whether the authors have been aware of that. We performed a systematic search on P ub M ed from 1 January 2011 to 31 May 2012 using the mesh terms ‘randomized controlled trial’ and ‘intensive care unit’. Twenty‐five RCT s with a total of 12,788 patients met the inclusion criteria, with an overall male dominance of 63.6% ( P < 0.0001). Eighteen of the 25 papers had an individually statistically significant gender difference in their total trial population. None of the 18 trials with a significant gender difference in their overall trial population had calculated the P ‐value for this overall difference. In the randomised groups, there was a significant gender difference in five papers. Seventeen had no significant gender difference in the randomised groups, and three papers did not state gender in the randomised groups. This study show that there is a marked male dominance in RCT s conducted in ICU s. We recommend that when planning future RCT s, the authors contemplate if their results can be used indiscriminately among ICU patients if the distribution of males and females is much skewed. It is relevant to determine if ones endpoint can be influenced by gender differences and if there is a risk of gender influence on data, proportional allocation or stratification should be considered.