Blinding and bias in a hypnotic clinical trial.

OBJECTIVES: Information is lacking regarding how commonly unblinding of treatment assignment occurs in hypnotic randomized clinic trials (RCTs). We now report the "best guesses" of clinical trial participants, versus study coordinators, versus study physicians in the study Reducing Suicidal Ideation Through Insomnia Treatment (REST-IT). METHODS: REST-IT, a, 8-week double-blind RCT, compared zolpidem extended-release (ER) versus placebo at bedtime in 103 adults with major depressive disorder with insomnia and suicidal ideation, and who received open label selective serotonin reuptake inhibitors. At the conclusion of study participation, 89 of the participants in this study, the study coordinators, and the study physicians each independently recorded their "best guess" of the treatment assigned. RESULTS: Patients guessed correctly 58.4% of the time, coordinators 53.9% of the time, and physicians 49.4% of the time, and none were different from chance alone. Agreement between patient/coordinator, patient/doctor, and coordinator/doctor dyads were 75%-78% with no significant differences in agreement between the dyads. CONCLUSIONS: "Best guesses" of all parties were not different from chance, suggesting that the blind was maintained and that assessment bias was minimized in this RCT of zolpidem ER versus placebo. Our results may not apply to other hypnotics or other RCT designs.

Duke Scholars

Author Andrew Darrell Krystal Psychiatry & Behavioral Sciences, Behavioral Medicine & Neur ...