Do Placebos Primarily Affect Subjective as Opposed to Objective Measures? A Meta-Analysis of Placebo Responses in Insomnia RCTs.

Introduction: Little is known about the relative magnitude of placebo responses on objective and subjective measures of sleep continuity. To address this issue, the pre-post effects of placebos on objective and subjective measures (i.e., polysomnography [PSG] and sleep diaries) were evaluated meta-analytically. The guiding hypothesis was that large responses would be observed on sleep diary measures and small responses would be observed on PSG measures.

Methods: PubMed searches, 1967-2016, yielded 329 possible articles, 17 of which met the inclusion and exclusion criteria for the present analysis (including 879 subjects with PSG data, 1,209 subjects with diary data, and six studies with both PSG and sleep diary data). Average change and weighted effect sizes (ESs) were computed via modeling for sleep latency (SL), wake after sleep onset (WASO) and total sleep time (TST).

Results: Pre-to-post change on PSG measures were: SL -13.7 min., ES = -0.37; WASO -14.3 min., ES = -0.36; and TST 29.8 min., ES = 0.50. Pre-to-post change on sleep diary measures were: SL -13.5 min., ES = -0.36; WASO -13.3 min., ES = -0.20; and TST 25.5 min., ES = 0.36. The modeled average objective subjective difference per sleep continuity measure was less than 5 minutes. The modeled average objective subjective difference per sleep continuity measure (in effect sizes) was less than 0.17.

Discussion: The observed outcomes of this analysis suggest that placebos produce comparable effects on objective and subjective measures of sleep continuity. Thus, objective measures do not appear to protect against placebo responses. This being the case and given the importance of the subjective experience of illness severity and recovery, such data suggests that prospectively sampled sleep continuity data (sleep diaries) may be the optimal data for clinical trials, particularly when only one measure is possible.