Should have been triple-blind... (Score 5, Informative)

The patients did not know they were being monitored (blind.) The doctors/nurses who entered the charts didn't know their patients' data would be used for this research (double-blind.) The people who analyzed the data, however, had everything upfront to poll and draw whatever conclusions they were looking for. "Using a query into the EHR..." "A further query of this subset..." "For each hypnotic user, we attempted to identify two controls with no record of a hypnotic prescription..." Sounds like they need a triple-blind experimental design.