Two things drive most of this, and neither is simply the amount of work. The first is demand paired with little say over how the work gets done. Researchers named that combination job strain in 1979 and it still holds up.

The second is effort that keeps outrunning what you get back. Low reward, thin recognition and shaky security all count, which is why a busy job you feel good about wears you down less than a quiet one you feel stuck in.

Either way your body reacts the same. A fast nerve signal lifts your heart rate and muscle tone in seconds, and a slower hormone signal keeps you in a raised state for the rest of the morning.

That system is superb for a short, sharp problem. It was never designed to be held open from Monday to Friday, which is where the wear comes from.

The clearest sign that your body treats work as a signal is the start of the day. Your main stress hormone rises further in the first half hour after waking on a work day than on a weekend day.

That happens before you have done anything. The body is responding to what it expects, which is also why Sunday evening can feel worse than Tuesday afternoon.

The other half of the problem is what does not happen. Recovery needs a real gap, and pooling 91 samples covering 38,124 people, being unable to switch off is what links a hard job to feeling worn out.

So the loop is not only about pressure going in. It is also about a body that never quite gets to the part where the pressure comes back out again.

The response your job triggers is the same one built for physical danger. It moves blood to your muscles, sharpens your attention, raises your heart rate and holds your shoulders and jaw a little tighter than usual.

All of that is useful for a few minutes. The design assumes an end point, a burst of effort and then a return to rest, and every part of the recovery depends on that return actually happening.

A job does not offer one. The demands arrive in a steady stream, five days a week, for years, so the state that was meant to be brief becomes close to your normal setting instead.

This is why work stress shows up as so many different complaints. It is one raised state touching sleep, muscles, digestion, mood and attention at the same time, rather than a separate problem in each of those places. Nothing in the body has gone wrong; it is doing exactly what it was asked to do, for far too long.

Nothing physical has to happen for this to begin. The pressure is a thought about what is expected of you, and your body treats that thought as the real event.

Researchers show this with a simple test. People give a short unrehearsed talk and do mental arithmetic in front of a panel, with no physical effort at all, and their stress hormones and heart rate climb reliably.

Two paths carry the signal. A fast nerve path lifts heart rate and muscle tone within seconds, and a slower hormone path keeps the raised state running for the next half hour or more.

The waiting counts too. In that test, ten minutes of simply knowing the talk was coming did much of the work, which is the laboratory version of the meeting you are dreading at four. Your body is not reacting to the job; it is preparing for it.

The most common objection is that the job is not that bad. The evidence does not need it to be. What matters is how long the raised state runs, not how dramatic any one day was.

Hours show this clearly. Following 2,960 full-time civil servants for five years, those working the longest hours had more anxiety and low-mood symptoms later, and the pattern tracked the hours rather than any single crisis.

Sleep follows the same graded shape. In a large survey of working adults, higher demands and longer hours came with more disturbed sleep, step by step, and one prospective study found a poor work environment roughly doubled the chance of a new run of bad nights.

So a mild version still adds up. Slightly too many hours, slightly too little say and slightly short nights, repeated for a year, does more than one terrible week.

Most people notice work stress outside work hours, which makes it easy to blame something else. The evening tension, the short fuse at dinner and the 3am replay are all still the job.

Ordinary habits keep the state open. Checking messages after dinner, a commute spent rehearsing tomorrow, screens late, and coffee used to get through the afternoon all extend the alert period rather than ending it.

Caffeine deserves its own line, because it is the most adjustable. Across trials and population studies, coffee reliably delays sleep and shortens it, and a tense week is usually the week people drink the most of it.

None of this makes the job the only cause. Short sleep, long static sitting, alcohol used to wind down, some medications and hormone shifts produce the same tiredness and tension, and a doctor can rule those in or out quickly. Stress is a very common explanation, not an automatic one.

Repetition teaches the pattern. After enough weeks your body starts the working day already braced, which is measurable: the morning hormone rise is bigger on a work day than on a rest day, before anything has happened.

The good news is that this is trainable in the other direction. Pooling 36 experimental studies with 2,847 people, structured stress-management programmes produced a solid, consistent improvement in strain, with an average course of about seven weeks.

The cognitive and behavioural versions did best. Learning to leave the work day behind mentally, rather than only resting physically, is the single most useful skill the recovery research points to.

A large review of 58 studies in healthcare staff found the same direction: small to moderate benefits from both individual and workplace changes, with real gains still present months later, though the quality of the evidence was low.

One popular idea deserves retiring. Years of work stress do not exhaust your glands and leave them unable to make stress hormone. A review of 58 studies found no support for that, and no valid test for it, so the tiredness is not adrenal fatigue.

The research also argues with itself in useful places. The famous idea that having more say protects you from high demands is very weakly supported: across 83 studies, the demands and the control each mattered, but the buffering effect mostly did not appear.

Direction is unsettled too. Our headline figure comes from six published studies of 27,461 people, and the same paper’s larger unpublished set gave a smaller number, which shrank further once earlier mood symptoms were accounted for.

And fixes are not guaranteed. A 2026 review of programmes where staff redesign their own work found no reliable improvement, so treat any single change as worth trying rather than proven.