The method should support judgment—not replace it.

Hazard identification

Start by understanding:

  • task;
  • people exposed;
  • equipment;
  • materials;
  • energy sources;
  • environment;
  • abnormal conditions;
  • changes.

Use sources such as:

  • observation;
  • procedures;
  • incident history;
  • worker input;
  • design information;
  • inspections.

Risk concepts

Risk usually reflects some combination of:

  • severity of harm;
  • likelihood or probability;
  • exposure/frequency;
  • effectiveness of existing controls.

Different organizations use different models.

Do not assume one scoring matrix is universally correct.

Qualitative and semi-quantitative methods

A simple matrix can help prioritize routine hazards.

More complex technical hazards may require specialist methods.

The important question is whether the chosen method is appropriate to the decision.

Hierarchy of controls

NIOSH's hierarchy prioritizes:

  1. elimination;
  2. substitution;
  3. engineering controls;
  4. administrative controls;
  5. PPE.

Do not jump directly to training and PPE if the hazard can be designed out or physically controlled.

Action ownership

A risk assessment without action ownership is incomplete.

Record:

  • action;
  • responsible owner;
  • due date;
  • verification method;
  • escalation if overdue.

Residual risk

After controls are proposed, reassess whether meaningful risk remains.

Do not automatically lower a score because an action is written down. The control must be implemented and effective.

Review triggers

Review may be needed after:

  • incident;
  • change;
  • new equipment;
  • new material;
  • process change;
  • regulatory update;
  • evidence that controls are ineffective.

Common mistakes

Treating the matrix as truth

The score does not know the process.

Generic hazards

"Slips/trips" is too broad if the actual exposure is an unprotected edge or chemical spill.

PPE-first thinking

Higher-order controls may provide more reliable protection.

Copy-paste assessments

A risk assessment should reflect the real task.

No worker input

People who perform the work often know failure modes that paperwork misses.

Evidence project

Create a fictionalized task risk assessment.

Show:

  • task;
  • hazards;
  • existing controls;
  • improvement options;
  • hierarchy-of-controls reasoning;
  • owners;
  • verification;
  • review trigger.

Strong risk assessment is about better decisions, not higher-quality spreadsheets.

Related content

Sources