Shift-work sleep fixes rest on thin evidence

Shift work is regularly associated with sleepiness on the job and disturbed sleep between shifts, which matters for workers' safety, well-being and long-term health. The usual remedies suggested are person-directed and non-pharmacological: timed exposure to bright light, planned naps, exercise, and education about sleep. These are cheap, widely recommended and easy for an individual to try, which is precisely why it is worth knowing how strong the evidence for them actually is. This review asked whether any of them measurably reduces sleepiness on shift or lengthens and improves sleep off shift.

The reviewers searched eight databases from inception to August 2015, screened reference lists and conference proceedings, searched a trial register and contacted experts for unpublished data. They included randomised controlled trials, including cross-over designs, testing any person-directed non-drug intervention in shift workers who also work nights. Seventeen relevant trials with 556 review-relevant participants were included and grouped into three types: various exposures to bright light (10 studies), various nap opportunities (4 studies), and other approaches such as physical exercise or sleep education (3 studies). At least two authors extracted data and assessed risk of bias, with meta-analysis where pooling was possible.

Combining two comparable studies with 184 participants, bright light during the night reduced sleepiness during a shift by a mean of 0.83 points on the Stanford Sleepiness Scale, rated very low quality evidence. Another trial in 16 participants found no significant difference on a different sleepiness scale, and bright light at night plus sunglasses at dawn produced no significant change against normal light in 17 participants. Bright light during a day shift did not significantly reduce daytime sleepiness in trials of 61 and 12 participants. A meta-analysis of a single nap opportunity found an 11.87 ms reduction in reaction time, again very low quality, and a two-nap opportunity produced a statistically non-significant increase in subjective sleepiness in 15 participants.

This is a case of the evidence being too weak to support confident advice rather than evidence that the advice is wrong. The included studies varied so much in intervention, setting and assessment tool, and reported so sparsely, that the reviewers concluded it is not possible to determine whether sleepiness can be reduced or sleep improved by these measures. They called for better, adequately powered trials of light and naps, alone and combined, that also take shift workers' chronobiology into account. The search closed in August 2015, so the picture may have moved since.

If you work nights, this is a reason to hold the standard tips loosely rather than to abandon them: light and naps remain plausible and low-risk, but the trials behind them are small and inconsistent. It also means that confident claims about the best light schedule or nap length are running well ahead of the published randomised evidence. Physical exercise and sleep education were flagged as promising, though with too little data to draw conclusions. Individual trial and error, rather than a prescribed protocol, is where most shift workers currently sit.