Propensity matching cannot recreate all the protections of contemporaneous randomization.
What the supplied source reports
The study prospectively observed dexmedetomidine-assisted procedures and compared them with earlier midazolam-plus-pentazocine records. After matching, ninety-one procedures per group were analyzed. Body movement and respiratory depression were less frequent in the dexmedetomidine group; hypotension and bradycardia did not differ significantly.
Limits of the evidence
The groups came from different calendar periods, and some patients had repeated procedures. Matching adjusts recorded factors but cannot guarantee balance in undocumented practice changes, operator behavior or patient selection. This was not a randomized drug-only comparison.
Implications for expert analysis
For an expert account, identify the whole regimen and the unit of analysis. Procedures and individual patients are not interchangeable denominators. A lower observed event rate supports an association under the studied practice conditions without proving that substituting one molecule caused every difference.
Source and related reading
Source: Publication PMC5588802, read from the supplied PDF. Applications to expert analysis are interpretation.
Related reading: A Hospital Fall Series Without Nonfalling Controls Cannot Rank the Risk of Individual Sleeping Pills.
Leave a Reply