COSHEM study materials
COSHEM study resources

COSHEM: Hazard identification, risk assessment and risk control

COSHEM · Work through your exam, one topic at a time.

All guide sections

Hazard identification, risk assessment and risk control

What this domain tests

This domain - hazard identification, risk assessment and risk control - tests whether you can identify hazards before they cause harm, assess risk consistently, and select controls in the right order. Scenarios often ask for the next step, the residual risk, or the weakness in a proposed control.

Core rules

  • Hazard and risk are different. A hazard is a source of harm; risk is the combination of likelihood and severity of that harm. A risk assessment records both the exposed people and how exposure occurs.
  • Techniques. Use HAZID or brainstorming for new systems, HAZOP for process deviations, JSA for task steps, FMEA for failure modes, and bow-tie analysis for major hazards with barriers. The technique must match the decision.
  • Risk matrix. A 5x5 matrix needs defined severity and likelihood criteria. Scores are not absolute; they prioritise action and must be applied consistently across similar tasks.
  • Hierarchy of controls. Eliminate, substitute, engineer, administrate, then PPE. PPE is last because it depends on fit, behaviour and maintenance.
  • Residual risk. Re-score after controls and record whether the residual risk is tolerable. If not, add controls or stop the activity.
  • Change and non-routine work. Maintenance, start-up, shutdown, lone work and new equipment need assessment, not only routine production.

Worked example

A warehouse task is scored severity 4 and likelihood 3, giving 12 on a 5x5 matrix, above the acceptable threshold. The first proposal is back belts and manual handling training. That relies on administrative control and PPE and does not reduce the source of risk. Better reasoning: eliminate the lift by installing a vacuum lifter, or reduce the load and lift height; then add training only for the remaining handling. Re-score: severity 4, likelihood 1, giving 4, now tolerable with monitoring. The assessment should record the workers consulted, the loads involved and the review date.

Traps that cost marks

  • Using inconsistent matrix definitions so scores cannot be compared.
  • Jumping to PPE or training before elimination and engineering.
  • Forgetting residual risk and the review date.
  • Ignoring cleaners, contractors, pregnant workers and lone workers.
  • Assessing only routine operation and missing maintenance or breakdown.
  • Producing a generic assessment with no task-specific exposure detail.