How the comparison was built
Oshima and colleagues compared double-balloon enteroscopy procedures using a dexmedetomidine regimen with records from an earlier period using midazolam and pentazocine. They used propensity-score matching and analyzed 91 procedures in each group. In the matched comparison, body movement occurred in 12.1% and 34.1% of procedures, respectively. The study’s respiratory-depression endpoint was also less frequent in the dexmedetomidine group.
What matching can balance
Matching can make groups more similar on the measured characteristics used to construct the score. A historical comparator can still differ in unrecorded patient characteristics, staffing, procedural practice or monitoring. The comparison was between care regimens from different periods, and some patients contributed more than one procedure.
How to assess a causation opinion
Ask which variables were matched, how the outcome was defined and whether the records would detect it equally in both periods. In an individual case, compare the medication sequence, oxygen and respiratory observations with procedural events. The group-level comparison can inform that review while leaving the contribution of individual drugs unresolved.
Leave a Reply