Risk Assessment as Loss Prevention: 7 Mistakes That Quietly Drive Up Workers’ Comp Costs

A risk assessment can pass every audit and still cause an employer to pay more for workers' compensation three renewals from now. 

The risk assessment document on file describes the job as someone wrote it down. Claims come from how the job is actually done, and every claim that gap produces feeds into the experience modification rate, the multiplier insurers apply to the premium.

The most expensive risk assessment mistakes are the ones that leave frequent, ordinary injuries uncontrolled. They show up as tasks nobody assessed, controls that depend on a worker wearing something, chemical data that went out of date, and assessments that stay the same after a claim proves them wrong. 

In this blog, we will discuss 7 risk assessment mistakes and how directly they feed claims into the premium, with the evidence behind each one.

1. Treating the Workplace Risk Assessment as Paperwork Instead of a Loss Forecast

A workplace risk assessment written mainly to satisfy an auditor will usually describe hazards in the language of the form. Entries like "slip hazard, low risk, housekeeping" are complete on paper and predict nothing. A forecast asks narrower questions: which task is most likely to produce the next claim, how often is that task done, and what would the claim cost?

The quickest test uses records you already have. Pull the loss runs, the carrier’s report of every claim filed, for the last three years and list the top claim causes by count and by dollars. 

If those causes don’t appear as high-priority hazards in the assessment, the assessment and the claims are describing two different workplaces, and only one of them is being priced. Keeping hazards, incidents, and claims in one place, as EHS risk management software does, turns that comparison into a standing check rather than a once-a-year exercise. 

2. Assessing the Normal Shift and Skipping Maintenance, Startup and Cleanup

Most risk assessments are built by watching a normal production shift, so they miss the work that happens when the line stops. OSHA's hazard identification guidance says tasks that are rare or aren't routine, "including maintenance and startup/shutdown activities, also present potential hazards." 

The same guidance names crushing, electrocution, arc flash, suffocation, and poisoning among the resulting injuries.

Clearing a jam, changing a blade, cleaning a mixer at the end of a shift and entering a tank are done less often and by fewer people. They also tend to happen under pressure to restart production, which is when guards come off and lockout steps get shortened. 

Non-routine work usually produces fewer claims than routine production, but the injuries on OSHA's list (crushing, electrocution, arc flash) sit at the severe end of the cost curve. Listing every non-routine task by name, then assessing each one separately, closes the gap a production-shift walk-through leaves open.

3. Writing Risk Assessments Without the People Who Do the Work

A risk assessment drafted in an office records the procedure as it was designed. Operators know the version that actually runs. For example, the step everyone skips because the guard blocks the view or the reach that is faster than walking around. 

None of that reaches the page unless someone who does the job is asked, which also makes this the cheapest of the seven risk assessment mistakes to correct.

OSHA's guidance on non-routine work puts it directly: "With your workers, think about what might go wrong or how people could get hurt". A practical version of a job hazard analysis is to walk each high-priority task with two people, one experienced and one fairly new. 

The experienced worker knows the shortcuts, and the new one still notices what feels wrong. Where their answers differ, the assessment has found a hazard the written procedure doesn't show.

Each of those unrecorded workarounds is a hazard repeated on every shift. Repetition is what turns a minor risk into a steady stream of small claims, the kind the primary-loss weighting counts at full value.

4. Leaning on PPE When Higher Controls Were Available

Personal protective equipment is often the first control an assessment reaches for because it is cheap, fast and easy to document. It is also the least reliable, which makes over-reliance on it one of the more common risk assessment mistakes. 

The hierarchy of controls published by the National Institute for Occupational Safety and Health (NIOSH) ranks elimination, substitution, engineering controls and administrative controls above PPE. PPE sits last because it works only when the right equipment is chosen, fitted and worn correctly every time.

An assessment that answers a lifting hazard with gloves and a reminder to "use proper technique" has accepted that the hazard stays in place. A lift table, a vacuum hoist or a change in how stock is delivered removes the load from the worker's back instead. 

Where PPE is the only control listed for a high-scoring hazard, the assessment should record why nothing higher on the hierarchy was feasible. If nobody can answer that question, the hazard has not been controlled yet. 

In claim terms, a PPE-only control tends to show up as the same injury type recurring in the loss runs, because the hazard is still there each time the equipment is skipped or worn wrong.

5. Underrating the Strains and Slips That Drive Claim Costs

Many scoring matrices treat strains and slips as low severity because a single one rarely looks catastrophic. The claim data says otherwise. The 2025 Liberty Mutual Workplace Safety Index draws on 2022 data and counts serious injuries, meaning those that keep an employee off work for more than five days. It puts their total cost at $58.78 billion, with overexertion involving outside sources at $13.7 billion and falls on the same level at $10.5 billion.

Together, those two causes make up about 41 percent of the cost of workplace injuries in the index. They are also exactly the frequent, moderate claims that the primary-loss weighting described above counts most heavily. 

A risk assessment that scores severity only by the worst plausible outcome will keep ranking the forklift above the loading dock ramp, even when the ramp is where the claims come from. Scoring each hazard on expected frequency and typical claim cost, alongside worst case, moves strains and slips up to where the loss runs already put them.

6. Assessing Chemical Hazards From Outdated Safety Data Sheets

Chemical risk assessments are only as current as the safety data sheet (SDS) behind each product. Section 8 matters most, because it sets out exposure limits and the protective equipment the manufacturer specifies. 

OSHA's Hazard Communication Standard, at 29 CFR 1910.1200(g)(5), requires manufacturers and importers to add significant new hazard information to an SDS within three months of becoming aware of it. A sheet filed years ago may no longer match the product on the shelf.

The claims that follow tend to be occupational illnesses such as contact dermatitis and respiratory sensitization, which can surface long after the exposure began and are harder to trace to a single task than an injury is. 

The fix is a scheduled check of the chemical inventory. For each product the assessment covers, compare the revision date on file with the manufacturer's current version, found through the supplier's website or an SDS search database. Where Section 2 or Section 8 has changed, the assessment for every task using that product needs to change with it.

7. Never Reopening the Assessment After a Near Miss or a Claim

An assessment that stays the same after an incident has been proven wrong and left in place. Every claim and every near miss is evidence about which hazard scores were too low. The cheapest of that evidence is the near miss, because nobody was hurt and nothing entered the loss runs.

The mod treats early, minor claims more gently than lost-time ones. In states that apply NCCI's Experience Rating Adjustment (ERA), medical-only claims are reduced by 70 percent before they enter the calculation. A hazard that produces a medical-only claim this year and is left uncorrected can produce a lost-time claim next year, and that one counts at full weight.

The practical answer is a written list of events that reopen the assessment: any recordable injury, any near miss on a high-scoring task, new equipment, a new chemical and any change to a process or layout. Teams that record each near miss against the assessment it tested can see a stale entry within days rather than at the next annual review. That link between reporting and review is how employers reduce workers' comp costs over time because the hazard is fixed while the evidence is still a near miss.

Fewer Risk Assessment Mistakes, Fewer Claims on the Mod

The seven mistakes share one cause: the assessment was written once and treated as finished, while the work, the chemicals and the claims kept changing. The benefits of risk assessment come from the part that keeps up with them. That means the comparison with loss runs, the walk-through with the people doing the task and the review after each near miss.

A useful first step this quarter is to set the last three years of loss runs beside the current assessment and mark every claim cause that the assessment scored as low. Each of those marks is a hazard the premium has already found. The assessment should find it next.

Share This Article: