Thread by @raphaels7

#sleep thread

1/ In 2014, a meta-analysis of randomized controlled parallel-group trials comparing placebos to sleep drugs for primary insomnia showed statistically significant small-to-moderate effect size improvements that favored drugs ncbi.nlm.nih.gov/pubmed/25168785

2/ However, the small number of polysomnography studies between 1992 to 2012 were low quality, resulting in “these effects sizes to be non-robust” according to the fail-safe N statistic. If “causal interpretations from non-robust models should be avoided” ncbi.nlm.nih.gov/pubmed/33781862

3/ then our confidence in non-robust small-to-moderate effect sizes should wane. A year later in '15 the same authors conducted a meta-analysis on the placebo effect in primary insomnia from randomized controlled parallel-group trials using polysomnography ncbi.nlm.nih.gov/pubmed/25515108

4/ Like drugs, placebos ALSO showed statistically significant small-to-moderate effect size improvements for primary insomnia. Unlike drugs however, the fail-safe N statistic indicated the effect sizes from placebos were robust.

5/ Furthermore, 63.56% (SD = 20.92) of sleep drug effects can be ACCOUNTED FOR by placebo effects. In one outcome variable called subjective wake-time after sleep onset, placebos performed even better than the sleep drugs.

6/ So if the effect sizes for primary insomnia from placebos and sleep drugs are similar, and if only the placebo effect size is statistically robust, costs close to nothing and is virtually side-effect free - what are the advantages of sleep drugs?

🤔

Let me know