Preventing Work-Related Stress: What Is Yours to Change, and What Isn't

A three-hundred-person company notices a rise in sick leave in one department. It responds quickly and, on the face of it, well: a stress management workshop is arranged, a meditation app is offered to everyone, a talk on sleep is scheduled for October. Six months later, sick leave has not fallen. Nobody lied, nobody did their job badly, and nothing moved.

What did not move is the department. Headcount is the same, volume is the same, priorities still change mid-week, the team lead still runs two teams, and no one can say any more who decides what. Employees now know how to breathe properly in an environment that has not changed by a millimetre.

It isn't that these measures are useless. It's that they act on the capacity to absorb, not on what there is to absorb. And you cannot honestly split what depends on you from what doesn't until you have looked at where the lever actually sits.

What the models say, and why it matters here

Three models from occupational psychology are enough to frame the problem properly.

Karasek: demand, latitude, support

Robert Karasek and Töres Theorell describe job strain through three variables. Psychological demand: volume, pace, interruptions, deadlines, contradictory requirements. Decision latitude: room to decide how the work is done, in what order, with what priorities, and whether your skills get used and developed. Social support: from colleagues and from management.

The core of the model is a combination: high demand paired with low latitude is the most damaging configuration. Support moderates it — softening it when present, worsening it when absent.

Look at those three one by one and ask which an employee can change alone. Volume is handed to them. Latitude is granted or withheld. Support depends on a team and a management culture. None of the three is a personality trait, and none is corrected by an individual exercise.

JD-R: why two people in the same job don't break in the same place

The Job Demands-Resources model of Demerouti and Bakker adds a useful distinction. Every job carries demands (what costs energy) and resources (what restores or saves it: autonomy, feedback, support, role clarity, prospects). It is not the level of demand in itself that damages people, but a lasting imbalance between the two.

This explains why two people in the same role do not deteriorate in the same place or at the same rate: they do not have the same resources, inside or outside work. It is also what makes comparisons between colleagues so misleading — "he manages" is not an argument, it is an observation about different resources.

Siegrist: when effort isn't repaid

Johannes Siegrist describes a second imbalance, distinct from the first: between the effort put in and the reward received. Reward is not only salary; it includes esteem, career prospects and job security.

High effort with low reward wears people down even when the workload is objectively manageable. This is the configuration of people who "don't understand why they're so tired when the work isn't that hard": it isn't the volume that costs, it's the ratio.

To which one should add organisational justice, as described by Jerald Greenberg: how outcomes are distributed, by what procedures decisions are made, and with what respect they are communicated. An unfavourable decision delivered with reasons and consideration does not produce the same effect as the identical decision dropped without explanation.

What genuinely depends on you

The split is not "nothing depends on you." It is more precise than that, and the list is short — but each of these five has a measurable effect.

The scope you accept

The most direct lever, and the least used. A large share of workload was never decided: it accumulated through individual acceptances, each reasonable on its own. A one-off favour becomes a habit, a habit becomes an expectation, an expectation becomes an obvious fact.

Taking back control does not mean refusing everything. It means reintroducing a decision where there was an automatism: every new out-of-scope request deserves one question — instead of what? Asked calmly, that question shifts the subject from yes/no to prioritisation. And prioritisation is your manager's job, not yours.

What you make visible

Much over-delivery is never measured, therefore never acknowledged, therefore never reflected in how a job is sized. It is a closed loop: invisible work carries no weight in decisions about work.

Making it visible is not complaining. It is producing a count: what the role contained a year ago, what it contains now, what was added and by what route. A written count turns a feeling into something discussable, in an annual review as anywhere else.

What you document

Documenting means recording facts as they happen, with dates, without characterising them. It serves three purposes: seeing for yourself a trend that memory smooths out, having usable material if you approach occupational health or employee representatives, and establishing the date on which a difficulty was raised.

It is a thankless exercise, because it means stripping the adjectives out of a situation saturated with them. ScanMyJob is built for this step: you describe the situation in your own words, and it returns a dated, neutral factual log.

How you phrase a request

A complaint describes a state and gets filed. A request calls for a decision and produces an answer. "I'm overwhelmed" calls for nothing; "I'm asking that a priority be set between X and Y before 30 June" calls for a yes, a no or a counter-proposal — and all three are information you do not have until you ask.

A refusal is not a failure. It tells you in seconds the actual room for manoeuvre of the person opposite, which months of waiting would not have revealed.

Recovery, which is not a reward

This is the only item on the list that really belongs to personal habits, and it deserves to be stated without inflation: sleep, physical activity and time genuinely offline do not compensate for a failing organisation, but their absence makes everything else worse. They determine how fast you recover from what you absorb.

Recovery is not what you grant yourself once the work is done. It is a condition for doing it. Which of those two sentences you believe decides the place recovery actually occupies in a week.

And there is an honest limit to set: fatigue is repaired by rest, wear is not. When three weeks off restore nothing, it is no longer a question of recovery, and the relevant person becomes occupational health or your GP. A test on work-related stress can give a measured reference point to discuss with a professional — a reference point, not a conclusion.

What does not depend on you

Everything else. And everything else is the bulk of it.

Workload. Volume, deadlines, number of files, contradictory requirements, headcount. No individual exercise changes a workload. Only an organisational decision does. Latitude. The ability to choose the order, the method, the moment. It is the most frequently overlooked variable, although in Karasek's model it is the one that decides whether high demand becomes a challenge or a strain. It is granted; it cannot be taken. Support. A manager who is genuinely available, a team that closes ranks, the possibility of saying you are struggling without it being used against you. That belongs to a culture and to a line of management. Recognition. The reward side of Siegrist's model: esteem, prospects, security. No one can recognise themselves. Role clarity. Knowing who decides what, knowing what is expected of you, and not receiving contradictory instructions. Role ambiguity and role conflict are among the best-documented demands, and they are properties of the org chart. Real work versus prescribed work. Christophe Dejours described the permanent gap between what is prescribed and what actually has to be done for the work to come out. That gap is closed, silently, by the ingenuity of the people doing the job — and when it is neither acknowledged nor discussable, it becomes one of the most constant sources of suffering at work. Whether it can be spoken about depends on the organisation, not on the person closing it.

One more point, and it is not a psychological variable: in most legal systems the duty to provide a safe workplace, including protection from foreseeable psychological harm, rests with the employer. It is not an optional good practice. Which means prevention is not, in the first instance, a personal development topic.

The transfer of burden

It is worth naming what purely individual prevention produces.

When an organisation offers resilience workshops without touching workload, latitude or support, it does more than pick the wrong lever: it transfers responsibility for the outcome onto the person carrying the constraint. The implicit message is that the difficulty comes from insufficient adaptation. The employee who still feels bad after the workshop draws the expected conclusion: it's me.

That is false, and knowing it has a concrete effect. Karasek's three variables describe a job. They say nothing about the worth or the sturdiness of the person doing it. Remembering that does not change the job, but it avoids the double penalty of enduring an organisation and then blaming yourself for it.

Which of course does not excuse skipping what does depend on you. The two are not opposed: they simply do not carry the same weight, and it is more honest to say so.

Key takeaways

Preventing work-related stress turns on three variables — demand, latitude, support — that belong to how work is organised, not to the person. A fourth imbalance, between effort and reward, wears people down even when the workload is manageable.

Five things genuinely depend on the employee: the scope they accept, what they make visible, what they document, how they phrase a request, and recovery. These are real levers, and they are limited.

Everything else — workload, latitude, support, recognition, role clarity — comes from organisational decisions, and the duty to keep the workplace safe rests with the employer. Learning to breathe inside a high-demand job with no room for manoeuvre does not change the job. Framing prevention as a matter of personal resilience puts the constraint on the person already carrying it.

---

Going through this? Our AI assistant, trained on 14 psychotherapy models, supports you confidentially. Try the assistant →

FAQ

How do you prevent stress at work when you have no decision-making power?

By acting on the five things that genuinely belong to you: the scope you accept, what you make visible, what you document, how you phrase a request, and recovery. That is little, and it is real. But the absence of room for manoeuvre is precisely one of the three variables in Karasek's model: it is a feature of the job, to be raised as such, not a personal shortcoming to be offset by extra effort.

Do meditation and breathing exercises actually help?

They act on recovery and on physiological reactivity, which is not negligible. They do not change workload, decision latitude or support — none of the variables that produce strain. Using them as a supplement is reasonable; presenting them as the answer to an organisational problem is not, because it implies that whoever still feels bad simply didn't practise hard enough.

Who is responsible for preventing psychosocial risk?

In most legal systems the duty to keep the workplace safe, including protection from foreseeable psychological harm, rests with the employer. Employee representatives have means to put the subject on the record, prevention or health and safety services act on risk assessment and organisation, and occupational health can recommend adjustments. Arrangements differ substantially between countries — check what applies where you work, and take advice from a union representative or an employment lawyer.

When should I see someone about it?

When rest no longer restores. A holiday that returns no energy, sleep that doesn't repair, physical signs that settle in, withdrawal that spreads from work into the rest of life: these are observable facts, and they justify a professional opinion. Occupational health is a good first port of call, because the consultation is covered by medical confidentiality and your employer receives only an opinion or recommendations. Your GP is the other door.