Time · 2.2

Shift patterns, and what they cost

Shift patterns, and what they cost. What is actually the case, and how it compares with what is repeated.

For a practical software perspective on this topic, this Monitask guide provides a concise companion guide.

A roster is a piece of engineering. Its parameters are shift length, direction of rotation, number of consecutive nights, and time off between shifts, and each of them has been studied.

The parameters

Length. Eight or twelve hours, mostly. Twelve gives more days off and concentrates fatigue; error and accident rates rise towards the end of long shifts, and the rise is steeper at night.

Direction. A rotation that moves forward, morning to afternoon to night, is generally better tolerated than one that moves backward. The body adjusts more readily to a later start than an earlier one, which is the same reason travelling west is easier than travelling east.

Consecutive nights. Partial adaptation takes several nights and is lost on the first days off, so a pattern of many consecutive nights and a pattern of isolated ones produce different problems rather than more and less of the same one.

Gap between shifts. A turnaround shorter than about eleven hours leaves too little time for a full sleep after travel, and short turnarounds are associated with worse outcomes across several measures.

What is established about health

The International Agency for Research on Cancer classifies night shift work involving circadian disruption in its group for probable carcinogens, a classification it reaffirmed after a further review.

Two things about that. The classification describes the strength of the evidence that a hazard exists, not the size of any risk, and the two are routinely confused in reporting. And it is a statement about a category of exposure rather than about any individual.

Nothing here is medical advice. Anybody concerned about their own shift pattern should raise it with a doctor rather than with a website about employment.

Who works these patterns

Health and social care, transport and logistics, manufacturing, hospitality, retail, security, emergency services and utilities. The deskless majority again, and the same population that the general conversation about work does not describe.

What the premium pays for

Unsocial hours premiums exist in most sectors by collective agreement or company policy. In the United States there is no general federal requirement to pay more for night work; in several European countries there is a statutory element.

Whatever the mechanism, the premium is compensation for a cost that falls outside the job: a social life scheduled against everybody else's, childcare that has to cover nights, and a sleep pattern that does not recover on a normal weekend.

The costs that are not paid for

Coordination with anybody on a standard week. School events, medical appointments, and the ordinary logistics of a household all assume daytime availability.

Those costs are real, they are borne by the worker and by the household, and they appear in no accounting of what shift work costs.

What good roster design looks like

Forward rotation. Limited runs of consecutive nights. Turnarounds of at least eleven hours. Predictability far enough ahead to plan around. And some worker input into the pattern, which the evidence associates with better outcomes independently of the pattern itself.

None of that is exotic and much of it costs nothing beyond the attention required to design the roster deliberately rather than by inheritance.

Why bad rosters persist

Because they are inherited, because changing one requires renegotiating with everybody it affects, and because the costs fall on people other than the person who would have to do the renegotiating.

The last is the general shape of this entry and of several others on this site: an arrangement is stable when the party who could change it is not the party paying for it.

The pattern that suits the worker

Where workers have been given real input into roster design, the patterns chosen are frequently different from the ones imposed and the outcomes are better. That result appears across several studies and it holds even where the resulting pattern is not the one the ergonomics literature would recommend.

Which suggests that control is doing work of its own, separately from the pattern, and that a roster designed for people without asking them is losing something no amount of design can replace.

The napping question

Short scheduled naps during night shifts improve alertness and reduce errors, and the finding is well replicated. Adoption is patchy and is resisted mainly on appearances rather than on evidence, since a member of staff asleep looks like a member of staff not working.

For broader context, consult NIOSH material on shift work.