Why real shift work doesn't look like a lab misalignment protocol
A controlled misalignment protocol is designed to produce a clean, predictable phase shift: researchers impose a shifted schedule and measure the internal clock's response under tightly controlled light, sleep, and timing conditions. Real shift work is messier. Workers rotate between day, evening, and night shifts, and the light they see on and around night shifts is often dim rather than bright enough to drive a strong reset [4]. The result is that the internal clock may barely move even when behavior moves a lot.
The police-officer field study makes this concrete. Before day, evening, and night shifts, most officers' 6-sulfatoxymelatonin (a breakdown product of melatonin measured in urine) and cortisol rhythms were still set to a day schedule [4]. After a median of seven consecutive night shifts, only about 30% of officers had adapted to a night-oriented schedule by the melatonin marker and about 24% by the cortisol marker [4]. That is a very different picture from a lab protocol, where the point is to shift the clock and the schedule is built to do so.
Adaptation, when it happens, is a large phase delay
The officers who did adapt showed a big move: their melatonin rhythm shifted by roughly 7.3 hours and their cortisol rhythm by roughly 6.3 hours, while non-adapters shifted by less than an hour on average [4]. So the real-world divergence is not a subtle drift. It is a split population: some workers' clocks follow the night schedule substantially, and many others stay essentially on day time.
That split matters for interpreting any single phase marker. A melatonin or cortisol midpoint taken from a real rotating worker is not a stable property of that person's schedule; it depends on whether that individual happened to be an adapter or a non-adapter over the shifts measured [4]. The same nominal night-shift exposure can therefore correspond to very different internal phases across workers.
Can wearables give a usable phase estimate outside the lab?
Directly measuring dim light melatonin onset (DLMO) — the rise in melatonin under dim light that marks the internal clock's phase — normally requires controlled sampling in a lab. A 2024 study tested whether wrist-worn actigraphy, which passively records light and movement, could predict DLMO in 45 fixed night shift workers [1]. Model predictions agreed reasonably well with in-lab DLMO, with a concordance coefficient of 0.70, about twice the agreement obtained from using average sleep timing on non-workdays as a proxy [1].
The accuracy is useful but not precise. The average absolute error was 2.88 hours, with 76% of predictions within 2 hours and 91% within 4 hours of measured DLMO [1]. In practical terms, a wearable estimate can tell you whether a worker is broadly day-oriented or night-oriented, but it can be off by several hours for an individual. That is consistent with the field picture: real-world phase is variable enough that a model built on light and movement captures the general pattern without nailing every person.
What the divergence means for real workers' health and mood
The real-world consequence of staying misaligned is not just inconvenient timing. In more than 800 first-year training physicians, wearable-derived markers of circadian disruption — measured both between internal rhythms and between those rhythms and the sleep–wake cycle — were linked bidirectionally with mood before and after the physicians began shift work, and the association was accompanied by clinically relevant changes on the 9-item Patient Health Questionnaire [2]. That study did not compare its markers to controlled lab misalignment predictions, so it supports a link between real-world disruption and mood rather than a specific divergence from lab protocols.
Inflammatory markers add another angle: a study of female healthcare workers found a cytokine signature associated with night-shift work exposure, consistent with circadian disruption having downstream immune effects in real-world shift workers [3]. Taken together, the evidence suggests that the real-world divergence from lab predictions is not a measurement artifact. Many shift workers remain internally misaligned, and that persistent misalignment is the state most plausibly tied to mood and inflammatory outcomes [2][3][4]. The open question is how much of an individual's risk depends on their particular phase position, which wearable estimates can currently only approximate [1].
Sources used in this answer
Predicting circadian misalignment with wearable technology: validation of wrist-worn actigraphy and photometry in night shift workers.
Wrist actigraphy plus a mathematical model predicted in-lab DLMO in fixed night shift workers with a concordance coefficient of 0.70 and mean absolute error of 2.88 hours, about twice as good as using average non-workday sleep timing as a proxy.
The real-world association between digital markers of circadian disruption and mental health risks
Wearable-derived digital markers of circadian disruption in over 800 first-year physicians were bidirectionally linked with mood before and after shift work began, with clinically relevant changes in PHQ-9 scores.
Multi-level inflammatory profiling reveals a cytokine signature associated with night-shift work exposure in female healthcare workers
A cytokine signature was associated with night-shift work exposure in female healthcare workers, consistent with circadian disruption having immune correlates in real-world shift workers.
Circadian Adaptation of Melatonin and Cortisol in Police Officers Working Rotating Shifts.
After a median of seven consecutive night shifts, only about 30% of police officers adapted to a night schedule by urinary aMT6s and about 24% by cortisol, with adapters showing roughly 6–7 hour phase delays and non-adapters shifting less than an hour.
Disturbance of the circadian system in shift work and its health impact
Review evidence indicates that shift work disturbs circadian phase markers such as melatonin and cortisol, with limited substantial phase shifts over multiple days in centrally controlled rhythms.