Safety Data: 4 Distortions That Hide Changing Exposure Risk

A dashboard can show fewer recordable cases while the operation is carrying more high-energy exposure. That contradiction is why safety data deserves a work-based review before it reaches the executive agenda. This article gives EHS managers and C-level leaders four ways to detect when a denominator is hiding risk rather than explaining it.

The central thesis is simple. A safety number is only as credible as the relationship between its outcome, its exposure, and the controls that were operating during the period. ISO 45001, published in 2018, requires organizations to evaluate performance, but it does not make a convenient denominator a faithful picture of risk. The work still has to be understood.

Why does a clean safety rate fail to describe changing work?

A clean rate fails when the work behind the denominator changes materially. If a plant records events per 200,000 hours, the rate may remain comparable only when those hours represent a reasonably similar mix of tasks, hazards, staffing, and control conditions.

The U.S. Bureau of Labor Statistics separates counts, rates, and measures of hours worked in its occupational injury and illness reporting. That separation matters because a rate is a relationship, not a direct observation of risk. When leaders treat it as a complete description, they may compare unlike periods with great confidence.

In projects supported by Andreza Araujo, the practical question is rarely whether a dashboard contains enough color-coded tiles. The stronger question is whether each tile helps a supervisor, plant manager, or board member decide which exposure needs attention before the next serious event.

Four distortions are especially common when the work mix changes. They affect the numerator, the denominator, the reporting boundary, or the link between an outcome and the control that should have prevented it.

Distortion 1: The denominator counts time instead of exposure

Hours worked are useful, but hours alone can conceal the intensity and distribution of work. A maintenance shutdown, a warehouse expansion, and a normal production week may generate the same number of hours while creating very different contact with stored energy, mobile equipment, lifting operations, or process hazards.

The Occupational Safety and Health Administration explains how hours worked support recordkeeping and rate contexts, which makes the measure practical for comparison. OSHA?s guidance does not mean that hours are the best denominator for every operational question. It means the organization must understand what the rate can and cannot answer.

A better dashboard keeps hours for continuity, then adds an exposure measure tied to the risk decision. Examples include critical lifts completed, energized interventions, vehicle miles, line breaks, hot-work permits, or confined-space entries. The measure should be stable enough to trend and specific enough to change a decision.

When a denominator changes from hours to tasks, document the break in the series rather than presenting the new rate as a continuous continuation. The change may improve visibility, but it also changes what the number means.

Distortion 2: Production growth makes performance look safer

A growing operation can produce more work, more exposure, and fewer events per hour at the same time. The rate may improve while the absolute number of high-risk tasks increases, which creates a leadership problem that a single downward line cannot solve.

The National Institute for Occupational Safety and Health emphasizes exposure assessment as a way to connect hazards, contact, and prevention decisions. That logic applies beyond industrial hygiene. A metric should help leaders see how much opportunity exists for a serious event and whether the controls matched that opportunity.

Consider a distribution center that doubles vehicle movements in twelve months while keeping the same recordable rate. The rate may support a favorable headline, but the safety review still needs to examine traffic separation, reversing controls, pedestrian interfaces, and the quality of field verification. A larger exposure base can demand more control capacity even when the rate is stable.

Exposure-based safety metrics help move that conversation from performance theater toward the question of which serious exposures remain active.

Distortion 3: Improved reporting makes the trend look worse

Better reporting can increase the number of concerns, near misses, or low-severity cases without increasing the underlying danger. That pattern is not automatically good news, because a higher count may reflect better access to reporting, a new classification rule, or a temporary campaign rather than a sudden change in field conditions.

The Health and Safety Executive explains that incident data should be interpreted with care because reporting systems capture only what enters the system. A rise in reports can therefore signal more events, better reporting, a broader definition, or several changes at once.

The executive review should record the reporting boundary beside the trend. Note whether the organization changed its app, hotline, supervisor expectation, contractor inclusion, classification rule, or follow-up process during the period. Without that context, leaders may punish the team for making weak signals visible.

Andreza Araujo?s book Safety Culture: From Theory to Practice supports a similar discipline in cultural diagnosis. Stated values matter less than the decisions that follow bad news. If a more open reporting channel produces discomfort, the response should examine the work and the controls before it attacks the signal.

Distortion 4: A good outcome is credited to the wrong control

A favorable outcome does not prove that the most visible safety activity caused the improvement. Training completion, observation counts, and meeting attendance can rise while exposure remains uncontrolled, because activity is easier to count than control effectiveness.

OSHA?s hierarchy of controls places elimination, substitution, engineering controls, administrative controls, and personal protective equipment in an order that reflects expected reliability. The hierarchy does not turn a dashboard into a causal model, but it gives leaders a practical test. Ask whether the metric shows that a stronger control was installed and remained effective.

The same test applies to leading indicators. A completed inspection is an activity. A verified interlock, a closed design gap, or a documented isolation that survives field sampling is evidence about control condition. The difference is material when the hazard can produce a serious or fatal outcome.

Critical-control dashboards are useful when they show the gap between an expected barrier and the evidence that the barrier worked during real operations.

How can leaders rebuild the dashboard without losing continuity?

Leaders can rebuild a dashboard by preserving historical outcome measures while adding a documented exposure layer and a control-effectiveness layer. The goal is not to discard the past; it is to stop asking one rate to answer three different questions.

Start with a metric dictionary that records the numerator, denominator, data owner, collection frequency, inclusion rules, and decision supported. A metric with six fields that nobody owns is not mature enough for executive use, even if it has a polished visual design.

Then create a monthly exception review with three questions. Which serious exposure increased? Which critical control failed or lacked evidence? Which decision owner accepted the residual risk, and until when? These questions keep the conversation close to work that can still be changed.

The dashboard should also show data quality. Mark missing denominators, late submissions, scope changes, and classification breaks rather than smoothing them away, because a visible limitation tells decision-makers where confidence ends and where field verification must begin before funding or escalation occurs. When a number is uncertain, visible uncertainty is more useful than false precision.

Which measures belong together in an executive safety review?

An executive safety review needs a small set of complementary measures, each with a defined job. Outcome measures show harm, exposure measures show opportunity, control measures show prevention strength, and decision measures show whether risk has an owner.

TRIR, SIF exposure, and control effectiveness should be read as complementary evidence, not as competing dashboard winners. The right combination depends on the decision, the hazard, and the operating context.

What should an EHS manager test in the next 30 days?

An EHS manager can test denominator quality in 30 days by selecting one high-consequence exposure and tracing its data from field activity to executive decision. The test should be narrow enough to complete and serious enough to reveal where the current dashboard loses meaning.

1. Choose one exposure that can create severe harm and identify its current numerator and denominator.
2. Compare the last 12 months with the actual work mix, including contractors, shutdowns, overtime, and technology changes.
3. Sample 10 records and confirm that each record uses the same inclusion rule.
4. Verify one critical control in the field and record whether the dashboard reflects its condition.
5. Present the finding to the decision owner and document the action, due date, and temporary protection.

One exposure, 12 months of history, 10 sampled records, and 30 days are enough to expose a surprising amount of dashboard weakness. The value comes from tracing the number back to work, not from adding another page to the report.

If a dashboard cannot show which high-consequence exposure changed, which control was tested, and who owns the next decision, it is reporting history while the risk keeps moving.

Safety data becomes trustworthy when the organization connects outcomes to exposure, control evidence, reporting quality, and accountable decisions. That is the shift Andreza Araujo has defended across more than 250 cultural transformation projects and in Safety Culture: From Theory to Practice. The practical standard is not a prettier dashboard. It is a dashboard that helps a leader change the work before the number becomes a consequence.