Review safety observation trends across US worksites

A site records twenty observations in one month and forty the next. Is safety getting worse, or are workers more willing to report? The count alone cannot answer. Observation trend review looks for repeated hazards, changing exposure and whether earlier controls worked. It should lead to specific decisions at the workplace, not a decorative dashboard or a contest to reduce reporting.

This guide covers analysis of safety observations and reported hazards, separate from writing a single observation form or closing an individual corrective action. OSHA's Recommended Practices on Hazard Identification and Assessment encourage employers and workers to group similar incidents and identify trends in reported hazards, then consider severity and likelihood to prioritise action. The Recommended Practices are guidance; this article does not present a universal federal observation-frequency rule or claim that every employer must use a particular score.

Decide what belongs in the review

Use observations of conditions and work practices, hazard reports, near misses, inspection findings and relevant incident learning. Keep their source types distinct. A near miss with serious potential should not disappear inside a larger pile of minor housekeeping observations. Record site, area, task, exposure, date, reporter's description and immediate protection taken. Remove unnecessary personal detail before a broader trend review.

Agree a small set of categories based on real work: vehicle-pedestrian interface, energy isolation, falls, manual handling, chemicals or other recurring hazards. Review the categories with workers so different sites use them consistently. A classification system that only safety professionals understand will produce misleading comparisons and discourage useful reports.

Examine exposure, not only totals

Compare like with like. A busier depot with more workers and shifts may produce more reports than a small office. A reporting campaign can increase numbers without increasing underlying risk. Look at where, when and during which task observations occur, and whether the workforce had equal opportunity to report. The OSHA programme-evaluation guidance recommends both leading and lagging indicators, including worker suggestions, reported hazards and near misses, time to respond and completion of corrective actions. No single indicator proves safety performance.

Notice repeated conditions even if no injury followed: the same loading-bay conflict at several sites, repeated temporary guards left open, or recurring housekeeping after a contractor shift. Ask whether reports point to a common system issue, such as poor scheduling or unclear ownership, rather than treating each event as an isolated worker mistake.

Check severity and immediacy

Triage serious current hazards immediately. Do not wait for a monthly trend meeting before securing an unguarded machine or an active traffic conflict. For the remaining pattern, consider possible harm, exposure, existing controls and the quality of evidence. A competent person should decide priorities with worker input. An observation classification is a prompt for investigation, not an automated risk determination.

OSHA's hazard-identification practices ask employers to consider hazards in nonroutine and emergency situations as well as ordinary work. A trend review that excludes rare high-consequence tasks may be incomplete even if its average numbers look good.

Test whether controls changed the pattern

Link each significant trend to a decision: control improvement, additional inspection, maintenance review, task redesign or training need. Name an owner and a field verification method. A policy update may not help if the physical layout still forces pedestrians through a vehicle route. Return to the workplace after the control is in use, ask workers whether it works and compare later observations. Keep the original pattern visible so the effect can be assessed honestly.

If reports decline sharply, ask whether access to reporting or trust has changed. A “zero observations” target can suppress reporting. Worker participation is a core element of OSHA's safety-management guidance; avoid using individual observations as a blame score. Explain how reports led to improvements so workers see why the data matters.

Example: recurring access obstruction

Three sites report blocked emergency exits, but the observations have different labels: storage, housekeeping and loading. A reviewer groups them by the underlying obstruction and checks photographs and times. The pattern occurs after deliveries when pallets are staged temporarily. The operations manager changes the delivery staging plan, marks an alternative area and asks supervisors to check it during the next deliveries. Workers confirm whether the new route is practical. The review records the original reports, the changed control and follow-up observations, rather than declaring success when a memo is sent.

The example does not claim an OSHA-required threshold of three reports; it illustrates how seemingly different entries can reveal one operational cause.

Share a short decision record

A useful review record names the period, sites, data sources, exclusions, trend, evidence quality, risk decision, immediate controls, longer actions, owner and follow-up date. Note uncertainty: sparse reports, missing shifts or changed reporting methods. Share conclusions with affected workers and managers in language they can test against what they see. Keep claims about improvement proportionate to the evidence.

Keep the review cadence tied to the work. A high-risk new process or repeated near miss may justify a prompt focused review, while a stable low-exposure area may be reviewed on a different rhythm. Document when the next check will occur and what evidence would show the control failed, so the decision can be reopened before an injury occurs.

For current-product discussion, see Complys US OSHA compliance software. This guide does not claim Complys predicts injuries, measures risk automatically, guarantees OSHA compliance or verifies the effectiveness of a physical control.