In 2025, 188 people died in workplace accidents in Kazakhstan, out of 1,345 recorded incidents. The numbers are marginally better than the year before — 202 deaths and 1,408 incidents in 2024, a drop of 7.9% and 4.5% respectively. Then you read the causes and the mild optimism evaporates. Poor organisation of work: 34.2%. Carelessness of the injured person: 33.6%. Violation of occupational safety rules: 8.5%. Roughly two thirds of everything that goes wrong on industrial sites has nothing to do with equipment failure or bad luck. It comes down to how people are organised and how people behave — which is precisely why safety culture in hiring deserves far more attention than it gets. Whatever you fail to filter at the gate, you will end up catching on the platform, and the currency there is somebody's health.
Operators and contractors in oil and gas usually invest elsewhere: in protective equipment, permits to work, monitoring systems, monthly toolbox talks. All of that is necessary. But at FSSA, running high-volume recruitment for industrial projects in the Atyrau region, we keep seeing the same pattern — a company spends millions on barriers and then walks people onto site who have no valid certifications, no relevant experience, and no habit of stopping work when something looks wrong. Barriers do not protect those people, because barriers only work on someone who does not route around them.
What Kazakhstan law already demands before day one
The formal framework here is stricter than most people assume. It is simply rarely read in full.
Article 182 of the Labour Code of Kazakhstan obliges the employer to provide training, instruction and knowledge testing on occupational safety and health — at the employer's own expense. The key wording: persons hired must undergo training followed by mandatory knowledge testing, and employees who have not passed training, instruction and knowledge testing are not admitted to work. Not "should preferably not be admitted." Not admitted. For managers and those responsible for safety, training and knowledge testing must be repeated at least once every three years at accredited professional development organisations.
Article 185 runs in parallel: mandatory medical examinations, split into pre-employment and periodic. A candidate applying for heavy work or work with harmful and hazardous conditions discloses chronic conditions that may be contraindications during the pre-employment examination. The examination produces a formal conclusion on whether the person's health matches the job, and those found fit receive a medical certificate. The list of hazardous production factors and occupations, the target groups and the examination frequency are set out in a separate Ministry of Health order. The employer organises and pays for all of it; the worker must not bear the cost.
There is also a national professional standard for occupational safety and health. It defines uniform requirements for the profession and — the part HR teams tend to miss — serves as a basis for hiring and certification criteria. The state has already handed you the yardstick.
Put it together and the law describes recruitment as a sequence of mandatory gates: medical check, training, knowledge test, admission to work. On paper, every company has them. In practice, each gate collapses into a rubber stamp with remarkable ease.
The medical examination happens at a clinic that issues certificates in half a day and is not, let us be gentle, especially demanding. Induction training is delivered to forty men in a site cabin, half of whom are half-listening and some of whom do not fully follow the language it is delivered in. Knowledge testing is recorded as a signature in a logbook. Formally, nothing is wrong — the paperwork is clean and an inspection will find no fault. Factually, a person nobody knows anything about has just walked onto a live site.
That is how "poor organisation of work" earns its 34.2%.
Where safety culture in hiring actually breaks down
The interesting failures do not live in the legislation. They live in the gap between hiring and safety — the gap that exists because one department does the recruiting, another owns the injury rate, and the two rarely talk.
Gap one is urgency. A contractor wins a tender, the site must be staffed in three weeks, 150 blue-collar workers are needed. What happens to the candidate requirements by week three? They drop. First you accept people without the right grade. Then people whose industrial safety certificates have expired — "we'll retrain them on site." Then whoever turns up. Funnel speed eats selection quality, and this is not malice, it is arithmetic: when the headcount deadline is rigid and the labour market is thin, rejection rates fall. For context, the Atyrau region sits among the top five regions for workplace injuries — 48 people in the first nine months of 2025 — while mining and metals accounts for 16% of the injured nationally, construction 10%, and oil and gas around 5%.
Gap two: nobody assesses attitude. Verifying a trade grade is easy — there is a certificate. Verifying experience is harder but doable; a call to the previous employer settles most of it. Attitude to risk, though, goes almost entirely unexamined, because it is not obvious how to examine it. And attitude is exactly what produces the second third of the statistics, the "carelessness" bucket. A man can hold every certification in existence and still sincerely believe that a hard hat in 40-degree heat is for worriers and that a harness slows the job down. That will never surface in an interview unless someone asks. Few do: interviews for trades run about fifteen minutes and cover whether the candidate can do the thing he is paid to do.
Gap three is language and comprehension. Large projects in Atyrau run mixed crews, documentation often exists in several languages, and the induction is delivered in one. The worker signs to confirm he understood. Whether he did is an open question — and this is not academic. If a person cannot restate in his own words what to do during an emergency shutdown, the induction did not happen, however immaculate the protocol.
Gap four is the uncomfortable one: contractor chains. The client demands safety from the general contractor, who demands it from the subcontractor, who hires through an intermediary, who takes whoever is available. Every tier has formally complied, and a worker whose competence nobody genuinely verified is now on site. Article 181-1 of the Labour Code separately addresses the safety rights and duties of workers of the sending party, so the legislator sees the problem. But the provision works only to the extent that the client is willing to audit people rather than documents.
The limits deserve honesty too. Screening at the gate does not solve everything and does not replace engineering controls: if a pump sits unguarded, the perfect candidate still gets hurt. Filtering on "attitude to risk" is subjective and slides easily into filtering out inconvenient people — the ones who ask questions. And tightening requirements in a thin labour market runs into physics: if the region does not contain 200 certified welders, no amount of strictness will create them. It will create vacancies and overtime for those who did show up, and overtime is itself an injury factor. So safety culture in hiring is not "filter harder." It is "filter smarter, and earlier."
What actually works
Start with the dull one: verify documents before the offer, not after. An expired industrial safety certificate discovered on the worker's first morning means either a missed deadline or a quiet admission that bypasses the rule. Guess which one gets chosen when the site is already standing idle.
Second: put a safety specialist into the hiring process — literally, as an interviewer for high-risk roles. Twenty minutes of their time per candidate costs incomparably less than one incident, the investigation, the downtime and the fine.
Third: ask about behaviour, not about rules. "Tell me about a time you stopped work over a safety concern" filters better than any written test. A candidate with twenty years on industrial sites and not a single such episode either does not notice hazards or does not believe he is allowed to mention them. Both are bad. A good answer sounds mundane: stopped, told the supervisor, sorted it, carried on. Watch how he narrates past incidents, too — if someone else is at fault in every story, that is a diagnosis.
Fourth: test comprehension, not signatures. After the induction, ask the person to restate the key points in his own words, in whichever language he is comfortable with. It adds ten minutes and removes the central illusion — that "signed" equals "understood."
Fifth: share the metrics. If recruitment is measured on time-to-fill and HSE is measured on injury rate, the two functions will pull in opposite directions. Tie them together: track the share of incidents involving workers in their first 90 days, the percentage of candidates rejected on medical or competence grounds, and the share who fail the knowledge test on the first attempt. That last figure is unusually revealing. If it is zero, your knowledge testing is not testing anything; it is a ritual.
Sixth, and this one is about money: price safety into the cost of recruitment and into the tender. When a subcontractor wins on a number that physically cannot fund qualified people and their training, he will hire unqualified people — not because he is a villain, but because of arithmetic. A client who selects on price alone is buying injuries on deferred payment terms.
The difference between a company where safety is real and one where it is decorative usually shows up in something small. Ask a rigger what happens if he stops the job over a hazard. If he says they will sort it out and thank him, the system is alive. If he smirks and says he will lose his bonus, then your inductions, your permits and your handsome posters are worth precisely nothing, and you will find that out on the day somebody decides not to stop. A safety culture is not created by decree and does not arrive with the hard hats. It is assembled out of the people you brought onto the site — one at a time, at the gate, long before anything goes wrong.
Common questions
Can a candidate be rejected on the basis of a medical examination?
If the pre-employment examination concludes that the person's health does not match the job and identifies contraindications to work in harmful or hazardous conditions, admitting them to that work is not permitted — this is a legal requirement, not employer discretion. The refusal must rest on the conclusion for that specific role with its specific production factors, not on a general impression of the candidate's health. The reverse situation, where someone with contraindications is waved through "at their own risk," is a direct liability for the worker and the company alike.
Who pays for training and medical examinations when hiring?
The employer, from its own funds. Article 182 of the Labour Code says so explicitly for training, instruction and knowledge testing on occupational safety and health. The same principle governs periodic medical examinations for workers in heavy, harmful or hazardous conditions: the employer organises and funds them, and the worker must not bear costs connected to that monitoring. Arrangements where a candidate is asked to "pay for the medical yourself and we'll reimburse you later" are not thrift — they are a violation that surfaces at the first inspection.