Interventions:
The comparators were relevant to the study setting. Their dose and duration of therapy were not measured, so the variation in these is unknown. The authors reported that opioid agonists were often given with propofol, and were associated with respiratory depression; no further information on the regimens was presented.
Effectiveness/benefits:
The effectiveness data were from a retrospective cohort study; this design is susceptible to bias due to confounding factors. The study's random selection method did not appear to control for baseline differences in patient characteristics between the two groups. Propensity score matching between the two groups would have allowed for greater study power and more similar groups. There appear to have been some differences between the groups; propofol patients were in a more severe state at time of surgery (p=0.001), and they were more likely to be obese (p=0.079). The clinical outcomes appear to have either captured the relevant health outcomes or were proxies. The authors indicated that both drugs were associated with adverse events, but other than mortality, these were not analysed.
Costs:
The costs were from the same retrospective cohort study, and susceptible to bias from the same confounding factors. Few details of the cost analysis were reported; it was only stated that hospital and pharmacy charges were included. The study perspective appears to have been that of a third-party payer, as hospital charges were reported.
Analysis and results:
The analysis and results were adequately reported. The authors compared their results with those of other studies. One of which was a randomised controlled trial, which is generally considered to be a more rigorous design than retrospective cohorts. This study found no statistically significant difference in the time to extubation between the two drugs (see Other Publications of Related Interest). The authors acknowledged some limitations to their study: its single-centre retrospective design, the lack of measurement of duration of opioid use, and differences in patient characteristics between groups. They acknowledged that none of the subgroup analyses were powered to detect statistical significance.
Concluding remarks:
The study methods and results were well reported, apart from the cost details. The results were susceptible to bias, and the authors' conclusions did not adequately account for this. It is unclear if these conclusions were appropriate.