7 Common COSHH Risk Assessment Mistakes UK Teams Make

The most common COSHH risk assessment mistakes are treating the supplier's safety data sheet as the assessment itself, assessing the chemical instead of the task, and never revisiting the assessment after something changes. Under the Control of Substances Hazardous to Health Regulations 2002 (COSHH), the duty is to assess the risk to health created by the work, not simply to file the hazard information a supplier sent you. The seven mistakes below are the ones that most reliably turn a well maintained COSHH folder into paperwork that no longer describes what happens on the floor.

Is a safety data sheet the same as a COSHH assessment?

No. A safety data sheet describes the substance as the supplier placed it on the market, while a COSHH risk assessment describes the risk to health created by the way your organisation uses it. The 16 section SDS format tells you the classification (Section 2), the composition (Section 3), the recommended exposure controls (Section 8) and the toxicological information (Section 11), but the supplier has no knowledge of your ventilation, your batch sizes or how long an operator stands over an open container. The 16 section SDS structure is set out in Annex II of GB REACH, retained following EU exit.

The practical test is simple. If someone removed every chemical name from your assessment, would anything left on the page tell a reader what your people actually do? If not, you have transcribed an SDS rather than assessed a risk.

What the SDS is genuinely good for

• Confirming the hazard classification and hazard statements you are assessing against
• Flagging respiratory or skin sensitisers, carcinogens and mutagens, which change what "adequate control" means
• Giving a starting point for control and PPE selection in Section 8, to be tested against your own conditions
• Identifying incompatibilities and storage constraints in Section 7

Should a COSHH assessment cover the substance or the task?

COSHH assessments should be built around the task, because exposure is a product of the work, not of the label. The same solvent decanted into a 20 ml vial in a fume cupboard and sprayed across a workshop floor by hand produces entirely different exposures, and therefore entirely different control requirements, from an identical safety data sheet.

A task based assessment records the quantity used, the physical form, the method (brushing, spraying, pouring, mixing, heating), the duration and frequency, the number of people exposed, the routes of exposure and the ventilation conditions. Where many low hazard products are used in an identical way, grouping them into a single activity assessment is usually more useful than producing one thin document per product. Our walkthrough of how an SDS becomes a task based assessment shows where each of those inputs comes from.

What about hazardous substances that arrive without a safety data sheet?

COSHH applies to substances generated by the work as well as substances you buy, and this is where registers most often fall short. Welding fume, wood dust, respirable crystalline silica from cutting or grinding, flour dust in bakeries, rubber fume, and mist from metalworking fluids all fall within scope, and none of them come with a supplier SDS.

Welding fume is the clearest example of why this matters. HSE strengthened its expectations for welding fume control after the International Agency for Research on Cancer classified welding fume as carcinogenic to humans, and now expects effective engineering control for all indoor welding, with respiratory protection where extraction alone is not sufficient.

A useful check is to walk the process rather than the stock cupboard. Anything that is heated, cut, ground, sanded, sprayed, decanted or allowed to evaporate is a candidate for the register.

Why is PPE the wrong place to start?

COSHH requires exposure to be prevented where reasonably practicable, and adequately controlled where prevention is not, with personal protective equipment used only where other measures do not achieve adequate control. An assessment whose control column reads "gloves, goggles, mask" has usually skipped elimination, substitution, process change, enclosure and local exhaust ventilation without recording why.

COSHH Schedule 2A sets out the principles of good control practice, and Regulation 7 governs the prevent or adequately control duty.

Recording why a higher control was rejected matters as much as recording the control you chose. "Substitution considered, no compliant water based alternative available at the required cure time" is defensible. Silence is not.

Do you need to know the workplace exposure limit?

If your substance has a workplace exposure limit, the assessment should say so and should explain how you concluded that exposure is below it. Workplace exposure limits for Great Britain are published by HSE in EH40, expressed as an 8 hour time weighted average and, for some substances, a 15 minute short term exposure limit.

Two nuances are commonly missed. First, an absence of a WEL is not evidence of safety, and many substances with serious health effects have no assigned limit. Second, a WEL applies to airborne concentration and says nothing about skin absorption or sensitisation, so substances carrying a skin notation need control measures aimed at contact, not just at air.

Air monitoring is not required for every assessment. It becomes relevant where exposure is likely to be significant, where the substance carries a WEL and control is uncertain, or where you need evidence that engineering controls are performing as designed.

What keeps control measures working after the assessment is signed off?

Controls degrade, and COSHH treats their upkeep as a continuing duty rather than a one off decision. Three checks catch most of the drift:

• Local exhaust ventilation must be examined and tested at intervals not exceeding 14 months for most processes, with shorter intervals specified for certain listed processes.
• Tight fitting respirators rely on a face seal, so they need fit testing for the specific make, model and size issued to each wearer, and that seal is defeated by beard growth.
• Health surveillance is required where exposure to certain substances, including respiratory and skin sensitisers, is linked to an identifiable disease and there are valid techniques for detecting it. Deciding whether it applies is part of the assessment, not a separate exercise.

Workers also need to know what the assessment concluded. An assessment that lives in a filing cabinet has not discharged the duty to provide information, instruction and training, and in practice it is the fastest way for good controls to stop being used.

How often should a COSHH assessment be reviewed?

COSHH does not set a fixed review interval. It requires assessments to be reviewed regularly, and immediately where there is reason to suspect the assessment is no longer valid or where there has been a significant change in the work. In practice, most organisations set a periodic review cycle and layer event triggers on top of it.

The event triggers that matter most:

• The supplier reissues the safety data sheet, particularly where the classification or hazard statements change
• You switch supplier for a product you consider equivalent, and the formulation differs
• The process, quantity, frequency or location changes
• Monitoring, health surveillance or an LEV examination produces an unexpected result
• An incident, near miss or health complaint points at the substance

The reissued SDS trigger is the one most often missed, because a revised sheet usually arrives as an email attachment rather than as a change request, and nothing in the process forces a comparison against the previous version. Keeping a dated record of which SDS version each assessment was built from is a low cost way to make that comparison possible. Safe Foundry was built around this problem, generating a draft COSHH assessment from the SDS and keeping the link between the two so that a version change surfaces as a review prompt rather than an unread attachment. You can see what that covers on the features overview, and the help centre has more on how assessments are structured.

What should you check before signing off a COSHH assessment?

Run the assessment against these questions before it is approved:

1. Does it describe a task, with quantity, method, duration and frequency, rather than only a substance?
2. Does the register include substances the process creates, not just substances you purchase?
3. Have higher controls been considered and, where rejected, has the reason been recorded?
4. If a workplace exposure limit exists, is it named, and is there a stated basis for believing exposure is below it?
5. Are the maintenance, examination, fit testing and health surveillance duties assigned to someone with a date?
6. Do the people doing the work know what it says?
7. Is it clear which version of the safety data sheet it was based on?

This article is general guidance on common assessment failures, not legal or compliance advice on your specific duties. Where an assessment involves carcinogens, mutagens, respiratory sensitisers or asbestos, lead or radioactive substances (which sit under their own regulations rather than COSHH), competent specialist input is worth obtaining. Our pricing page and FAQ set out what the platform does and does not do.