Rethinking bias assessments for emergency setting trials

Posted on 7 September 2015

A new review of quasi-randomisation in trials in emergency care settings by Mark Corbett and CRD colleagues has just been published in Research Synthesis Methods.

In systematic reviews, quasi-randomised trials are normally classed as high risk for selection bias. This means their results are often deemed unreliable due to the possibilty that the trial treatment groups differed in key characteristics before receiving treatment. Mark Corbett and CRD colleagues undertook a review of quasi-randomisation (compared with true randomisation) in trials in emergency care settings where recruitment time restrictions may limit opportunities for selection bias.

The review did not show strong or consistent evidence that quasi-randomisation is associated with selection bias more often than true randomisation. Limited trial data meant it was not possible to adequately explore the possibility that true randomisation might result in slower trial recruitment rates or in the recruitment of less representative populations. High risk of bias judgements for quasi-randomised trials should not be assumed when such trials are included systematic reviews of emergency setting interventions.

Randomization methods in emergency setting trials: a descriptive review