On 24 October the Journal of Clinical Sleep Medicine published a clinical practice guideline for shift work disorder, alongside the systematic review that supports it. It is the American Academy of Sleep Medicine’s update of its earlier guidance, and it runs to 19 numbered recommendations covering three problems: sleepiness during the shift, insomnia when trying to sleep afterwards, and the project of adapting the body clock to night work at all.
Almost every recommendation begins with the same two words. Not “we recommend” but “we suggest.” Not one of the 19 is graded strong. Eighteen are conditional; the strength vocabulary exists in the document, and the guideline notes what a strong recommendation would mean — something a clinician should follow for nearly every patient, potentially a quality measure. Shift work disorder did not earn any of those.
That is not a scandal. It is a fairly precise description of how thin the trial base is, and it is more useful to a reader than a confident list would have been.
What the diagnosis describes
The guideline defines shift work disorder as the consequence of a schedule that routinely overlaps a person’s usual sleep timing for at least three months. It shows up in two forms, which the document treats as separate treatment problems: excessive sleepiness during the major waking period, and sleep disturbance or insomnia when sleep is finally possible. Not everyone who works nights develops it.
How common is it? The guideline cites a meta-analysis of 29 prevalence studies that pooled a figure of 26.5% of adult shift workers, with a 95% confidence interval of 21.0 to 32.8. The same analysis reported heterogeneity of 98.5%, meaning the studies disagreed enormously with one another, partly on the basis of which diagnostic criteria they applied and how large their samples were. Read that as “common, and measured inconsistently” rather than as a precise rate. Our earlier piece on why sleep researchers now treat timing as its own variable covers some of the same ground.
What the guideline suggests for the shift itself
The recommendations are listed alphabetically within each category, and the task force says explicitly that the order is not a ranking. Certainty, though, varies enough to be worth naming.
Bright light during the night shift drew the largest evidence pool of the non-drug options: 18 randomised trials and 6 non-randomised studies, with exposures ranging from 361 lux to 10,000 lux and from 20 minutes to nine hours, in continuous or intermittent doses. The pooled estimates pointed to meaningful reductions in sleepiness and better performance on vigilance testing, and the recommendation is still graded very low certainty — the trials were mostly in healthy volunteers in laboratories or in shift workers without a formal diagnosis. Possible downsides named in the document are headache and eye strain, and it notes that bright light is not practical in every job, driving being the obvious example.
Caffeine before or early in the night shift rests on 8 randomised trials and is also very low certainty, with a remark that caffeine close to bedtime can disrupt sleep onset and quality, which is the very thing the guideline is trying to fix. A nap before the shift is suggested too, with a specific caution: sleep inertia after waking can temporarily increase sleepiness and reduce performance, so the document points to allowing more than 20 minutes to recover before driving or doing safety-sensitive work.
Two recommendations are aimed at employers rather than individuals. One suggests clockwise rotating schedules over counterclockwise rotation, drawn from 10 non-randomised studies of planned schedules at very low certainty. The other declines to prefer 8-hour over 12-hour night shifts, noting that the reviewed studies covered 36 to 52 hours of work per week and did not look at shifts longer than 12 hours.
There is also a small nutrition note. For sleepiness during the night shift, the guideline suggests either not eating or eating a snack rather than a full meal — a conditional recommendation of low certainty, and the only place in the document where food appears.
The one moderate-certainty finding is a prescription drug
Armodafinil and modafinil, taken 30 to 60 minutes before a night shift, carry the guideline’s highest certainty grade: moderate. The supporting trials measured objective sleepiness on the multiple sleep latency test in people diagnosed with shift work disorder, plus self-rated sleepiness and vigilance performance, and found clinically meaningful improvements. Even here the recommendation is conditional, and the safety text is not incidental. The task force lists rare but serious events, including suicidal ideation in a patient with a history of depression, serious or life-threatening rash, and multi-organ hypersensitivity reactions; it also notes reported risk of abuse and dependence, and common side effects of headache, nausea, dizziness, sleep disturbance and anxiety.
In other words, the strongest evidence in this guideline points to something a clinician prescribes and monitors, not something a person picks up on the way home from a shift.
Melatonin: three suggestions and a lot of asterisks
Melatonin appears in three separate recommendations — for daytime sleep after a night shift (very low certainty, 10 randomised trials), for the transition back to night-time sleep when a stretch of shifts ends (low certainty, 3 trials), and for falling asleep before a nap ahead of a first night shift (very low certainty, 4 trials). Ramelteon and other melatonin receptor agonists get one more, graded low certainty on the strength of a single 8 mg ramelteon trial.
The document is candid about the mess underneath. Doses across the included studies ranged from 0.1 mg to 40 mg, taken before a daytime sleep opportunity, with no settled dose or timing emerging. It also cites work finding that the actual melatonin content of over-the-counter products ranged from 83% below to 478% above the labelled amount, with chewable and capsule forms varying most. Undesirable effects were judged trivial — headache, dizziness, nausea, drowsiness — but the guideline is a synthesis of short trials, not a long-term safety record.
That distinction matters because the surrounding federal advice has not moved as far as the new suggestion might imply. The National Center for Complementary and Integrative Health’s consumer page on melatonin describes the shift-work evidence as generally small or inconclusive: one 2014 review of 7 studies and 263 participants suggested people taking melatonin slept about 24 minutes longer during daytime sleep, on low-quality evidence, while a second review of 8 studies and 300 participants reached no recommendation. The same page notes that information on long-term safety is lacking, that the 2015 sleep-medicine guidance advised against melatonin for people with dementia, and that anyone with epilepsy or taking blood thinners should be medically supervised. A conditional “we suggest” from a task force reading simulated-shift laboratory data is a different kind of statement from “this works.”
What the guideline declines to do
Cognitive behavioural therapy for insomnia is suggested for the daytime sleep problems, at very low certainty, on the basis of a single randomised trial and three non-randomised studies using protocols modified for shift workers; the document flags access as a real barrier, since trained providers are scarce. Orexin receptor antagonists such as suvorexant are suggested at low certainty. Benzodiazepines are the one place the task force suggests against: triazolam improved daytime sleep in four trials, but falls, abuse, dependence and withdrawal outweighed the benefit in the group’s judgement.
The most portable sentence in the whole document may be a remark rather than a recommendation. The light-timing and reduced-light-transmittance-glasses strategies for circadian adaptation are qualified by the observation that for people on rotating or frequently changing shifts, circadian adaptation is not a realistic treatment goal at all — the evidence behind those recommendations came from workers who reached partial adaptation on stable night schedules.
What the evidence base actually contains
The systematic review screened 3,045 records and found 118 studies with data suitable for meta-analysis. Much of that is laboratory work in which healthy volunteers were put on simulated night schedules, plus field studies of shift workers who were never formally diagnosed. The task force says plainly that very few trials enrolled people with a diagnosis of shift work disorder, that little research addresses the long-term health consequences of shift work or circadian misalignment, and that early-morning shifts — which also require working through hours a person would normally be asleep — are barely studied at all.
Where this leaves a shift worker
The NIOSH fatigue page frames the same problem from the workplace side: fatigue is commonly associated with nonstandard schedules that disrupt or shorten sleep, and it slows reaction times, reduces attention and concentration, limits short-term memory and impairs judgement. NIOSH’s position is that employers and workers should address it together, which is a fair description of what this guideline can and cannot do. It gives clinicians a defensible menu and an explicit account of how weak each item on it is. It does not give anyone a self-treatment protocol, and its most confident entry is a prescription medicine with a serious-adverse-event list.
If the three-month definition above sounds like your working life, the reasonable next step is a conversation with a healthcare professional about sleepiness and sleep timing — not a supplement order. Nothing here is medical advice; our medical disclaimer sets out what this site does and does not do. For the ordinary evening routines that sit alongside any schedule, the habits worth avoiding before bed are a separate and much better-evidenced subject.