Occupational ill health does none of that. Hearing loss develops over years. Hand-arm vibration syndrome begins as intermittent tingling that operators dismiss. Occupational asthma emerges after months of exposure that felt entirely tolerable at the time. By the point anyone notices, the damage is frequently permanent.
This is the reason the HSE has rebalanced its inspection effort towards health, and it is the area where manufacturers are most exposed.
The scale of it
In 2024/25, an estimated 1.9 million working people in Great Britain were suffering from a work-related illness. Of those, around 964,000 had work-related stress, depression or anxiety, and 511,000 had a work-related musculoskeletal disorder. Work-related illness and workplace injury together accounted for 40.1 million lost working days.
For comparison, provisional figures for 2025/26 recorded 126 workers killed in work-related accidents — the lowest annual total the HSE has recorded. Meanwhile, mesothelioma alone caused 2,146 deaths in Great Britain in 2024, the overwhelming majority attributable to past asbestos exposure at work.
The estimated annual cost of injuries and ill health arising from current working conditions was £22.9 billion in 2023/24.
Of the 13,200-plus inspections the HSE completed in 2024/25, more than 7,000 focused on work-related health risks. The three priorities it has identified for manufacturing are noise, musculoskeletal disorders and respiratory risk.
Noise
What the law requires
The Control of Noise at Work Regulations 2005 set three thresholds based on daily or weekly personal noise exposure:
- Lower exposure action value: 80 dB(A) (peak sound pressure 135 dB(C)). You must assess the risk, provide information and training, and make hearing protection available on request.
- Upper exposure action value: 85 dB(A) (peak 137 dB(C)). You must reduce exposure so far as is reasonably practicable by means other than hearing protection, mark hearing protection zones, enforce the use of hearing protection, and provide health surveillance.
- Exposure limit value: 87 dB(A) (peak 140 dB(C)). This must not be exceeded, and it is measured taking the attenuation of hearing protection into account.
Where manufacturers go wrong
Defaulting straight to hearing protection. The regulations require noise to be reduced at source so far as is reasonably practicable before hearing protection is relied upon. Damping panels on sheet metal, resilient mountings, enclosure of a specific noisy component, changing an air nozzle for a low-noise design, replacing an impact process with a progressive one — these are the controls an inspector will ask about. Issuing ear defenders and moving on is a documented failure to apply the hierarchy.
Assessing the room instead of the person. Exposure is personal and time-weighted. An operator who spends two hours a day at a 95 dB(A) machine and six hours at 78 dB(A) has a different exposure from a colleague who does the reverse. Assessments need to reflect actual work patterns.
Over-protection. Selecting the highest-attenuation defender available can push effective exposure below about 70 dB(A), leaving the wearer isolated from warning signals, machine noise that indicates a fault, and colleagues. Over-protected workers remove their protection. The aim is to bring exposure to a safe level, not to the lowest achievable one.
Health surveillance in name only. Audiometry is not a box-tick. It produces results that require action — referral, review of controls, and in some cases a change of role.
Dust, fume and hazardous substances
The COSHH duty
The Control of Substances Hazardous to Health Regulations 2002 require you to assess risk, prevent exposure where reasonably practicable, and otherwise adequately control it — applying the principles of good control practice set out in Schedule 2A.
The point manufacturers most often miss is that COSHH applies to substances generated by your processes, not only to those you purchase. Welding fume, wood dust, flour dust, metalworking fluid mist, rubber fume, silica from cutting or grinding, and solvent vapour all fall within scope even though none of them arrives with a safety data sheet.
Welding fume deserves specific mention: the HSE’s position, following international reclassification of welding fume as a human carcinogen, is that effective engineering controls are required for all welding indoors, with RPE where engineering controls alone cannot achieve adequate control.
Local exhaust ventilation
LEV is the workhorse control in most manufacturing environments, and it is also where the most common findings arise.
Under COSHH Regulation 9, LEV must undergo a thorough examination and test at least every 14 months — more frequently for certain processes specified in the regulations. That examination should verify performance against the system’s original design specification, which means you need to have that specification. A commissioning report retained from installation is worth a great deal at inspection; a system with no baseline data cannot be meaningfully tested against anything.
Practical failure points: hoods moved further from the source, flexible ducting kinked or replaced with an unsuitable substitute, filters loaded and never changed, dampers adjusted to help one workstation at the expense of another, and extraction switched off because it makes the area cold in winter.
Respiratory protective equipment
Where RPE forms part of the control strategy, tight-fitting facepieces require face fit testing for each individual wearer and each make, model and size of facepiece. Facial hair in the region of the seal prevents an adequate fit — no amount of tightening compensates for it.
RPE without face fit testing is one of the most frequently identified failures on health-focused inspections, and it is entirely avoidable.
Hand-arm vibration
The Control of Vibration at Work Regulations 2005 set, for hand-arm vibration:
- Exposure action value: 2.5 m/s² A(8). Above this you must reduce exposure to as low a level as is reasonably practicable and provide health surveillance.
- Exposure limit value: 5 m/s² A(8). This must not be exceeded.
The single most common error is calculating exposure from manufacturers’ declared vibration emission values. Those figures come from standardised laboratory tests and routinely understate real-world exposure, sometimes substantially, because they do not reflect a worn tool, a blunt consumable, the workpiece material, or how the operator grips and applies force.
Exposure should be calculated from realistic trigger times — actual time the tool is in contact and vibrating, not shift length — using representative data for the tool as used.
Hand-arm vibration syndrome is preventable and irreversible. Health surveillance exists to catch it at Stage 1, when removing the exposure can halt progression. A tiered programme typically runs from an initial questionnaire through annual screening to assessment by a qualified occupational health professional where symptoms are reported.
Musculoskeletal disorders
MSDs remain the largest single category of work-related ill health after stress. The HSE’s stated expectation is that preventative controls are “embedded into day-to-day operations, not just documented.”
That framing matters. A manual handling risk assessment and a training record will not satisfy an inspector who can watch an operator repeatedly reaching across a conveyor to retrieve a component. The controls that count are the ones designed into the task: bench heights matched to the work, components delivered at waist level, lift assists and manipulators, job rotation that genuinely varies posture and load rather than moving people between similar tasks, and tool selection that reduces grip force.
Building a health surveillance programme that works
Health surveillance is a legal requirement in defined circumstances — under COSHH, the Noise Regulations, the Vibration Regulations, the Control of Lead at Work Regulations and the Control of Asbestos Regulations among others — but it is not a substitute for control. It is a check on whether your controls are working.
A functional programme needs:
A defined scope. Which employees, based on which exposures, at what frequency. Derived from your risk assessments, not from job titles.
A baseline. Assessment on appointment or on first exposure, so that later results have something to be compared against.
A feedback loop. This is the component most often missing. Where surveillance identifies a case, it should trigger a review of the controls for that person, that task, and everyone else doing similar work. Surveillance results that go into a file and change nothing are an expense, not a control.
Records kept for the required period. Health records must generally be kept for 40 years from the date of the last entry under COSHH, and equivalent long retention periods apply under the asbestos and lead regulations. Given that latency, these records outlast most IT systems — make sure yours will survive a system migration or a change of occupational health provider.
Soteria (UK) Ltd helps manufacturers across the UK and Ireland get occupational health risk under proper control — assessing exposure, specifying practical controls, and building health surveillance programmes that actually inform decisions. Our consultants are Chartered Health and Safety Practitioners registered on the OSHCR.
Call 01925 626283 or email info@soteriauk.co.uk.
Recent Comments