Sleep restriction therapy for insomnia – a new rubbish trial in THE LANCET

A ‘pragmatic, superiority, open-label, randomised controlled trial’ of sleep restriction therapy versus sleep hygiene has just been published in THE LANCET. Adults with insomnia disorder were recruited from 35 general practices across England and randomly assigned (1:1) using a web-based randomisation programme to either four sessions of nurse-delivered sleep restriction therapy plus a sleep hygiene booklet or a sleep hygiene booklet only. There was no restriction on usual care for either group. Outcomes were assessed at 3 months, 6 months, and 12 months. The primary endpoint was self-reported insomnia severity at 6 months measured with the insomnia severity index (ISI). The primary analysis included participants according to their allocated group and who contributed at least one outcome measurement. Cost-effectiveness was evaluated from the UK National Health Service and personal social services perspective and expressed in terms of incremental cost per quality-adjusted life year (QALY) gained. The trial was prospectively registered (ISRCTN42499563).

Between Aug 29, 2018, and March 23, 2020 the researchers randomly assigned 642 participants to sleep restriction therapy (n=321) or sleep hygiene (n=321). Mean age was 55·4 years (range 19–88), with 489 (76·2%) participants being female and 153 (23·8%) being male. 580 (90·3%) participants provided data for at least one outcome measurement. At 6 months, mean ISI score was 10·9 (SD 5·5) for sleep restriction therapy and 13·9 (5·2) for sleep hygiene (adjusted mean difference –3·05, 95% CI –3·83 to –2·28; p<0·0001; Cohen’s d –0·74), indicating that participants in the sleep restriction therapy group reported lower insomnia severity than the sleep hygiene group. The incremental cost per QALY gained was £2076, giving a 95·3% probability that treatment was cost-effective at a cost-effectiveness threshold of £20 000. Eight participants in each group had serious adverse events, none of which were judged to be related to intervention.

The authors concluded that brief nurse-delivered sleep restriction therapy in primary care reduces insomnia symptoms, is likely to be cost-effective, and has the potential to be widely implemented as a first-line treatment for insomnia disorder.

I am frankly amazed that this paper was published in a top journal, like THE LANCET. Let me explain why:

The verum treatment was delivered over four consecutive weeks, involving one brief session per week (two in-person sessions and two sessions over the phone). Session 1 introduced the rationale for sleep restriction therapy alongside a review of sleep diaries, helped participants to select bed and rise times, advised on management of daytime sleepiness (including implications for driving), and discussed barriers to and facilitators of implementation. Session 2, session 3, and session 4 involved reviewing progress, discussion of difficulties with implementation, and titration of the sleep schedule according to a sleep efficiency algorithm.

This means that the verum group received fairly extensive attention, while the control group did not. In other words, a host of non-specific effects are likely to have significantly influenced or even entirely determined the outcome. Despite this rather obvious limitation, the authors fail to discuss any of it. On the contrary, that claim that “we did a definitive test of whether brief sleep restriction therapy delivered in primary care is clinically effective and cost-effective.” This is, in my view, highly misleading and unworthy of THE LANCET. I suggest the conclusions of this trial should be re-formulated as follows:

The brief nurse-delivered sleep restriction, or the additional attention provided exclusively to the patients in the verum group, or a placebo-effect or some other non-specific effect reduced insomnia symptoms.

Alternatively, one could just conclude from this study that poor science can make it even into the best medical journals – a problem only too well known in the realm of so-called alternative medicine (SCAM).