Eligibility requirements in hypnotic trials

Forty-eight patients complaining of insomnia were studied at two sleep laboratories using an identical protocol to evaluate hypnotic efficacy. All met the screening requirement of a mean sleep latency of 30 min or greater on 3 laboratory nights following an adaptation night. Of these patients 34 still complaining of insomnia were screened a second time 2 to 6 months later. Sixteen of the 34 failed the second screen. Sleep parameters for the 34 on screen 1 compared with screen 2 were the same except for sleep latency (the eligibility criteria), which was significantly shorter. There was no evidence of a systematic difference between laboratories, a change in procedure from screen 1 to 2, or a systematic loss of patients from screen 1 to 2. The data show that the statistical phenomenon of regression toward the mean must be considered in designing hypnotic efficacy studies.