Why did two analyses give different answers about midazolam and emergence delirium?
An observational anesthesia study can give different estimates when researchers build the comparison group in different ways. A pediatric emergence-delirium study illustrates why the prespecified primary analysis matters.
How the results differed
The study compared children who received clonidine or midazolam before anesthesia with children who received neither premedication. The overall frequency of documented emergence delirium was about 6%. In the primary propensity-matched analysis, the estimate for midazolam was an odds ratio of 1.23 with a 95% confidence interval of 0.79–1.92. A secondary adjusted analysis reported an association for midazolam.
What to look for in the anesthesia record
Compare the definition and timing of delirium with the child’s actual observations during emergence. Factors such as the operation, pain and other anesthetic medications can affect recovery. The differing model results make the association sensitive to the analysis. They do not establish that midazolam caused a specific child’s postoperative behavior.
Read the original study · Full text
