Top 10 Leading Safety Indicators That Predict Incidents

Lagging indicators (TRIR, LTIR, fatalities) tell you what already broke (glossary: lagging indicator). Leading indicators (glossary: leading indicator) tell you what is about to break, giving you time to act before someone gets hurt. OSHA's 2019 guide, Using Leading Indicators to Improve Safety and Health Outcomes, defines leading indicators as proactive, preventive and predictive measures of how well your safety and health activities are working, and lagging indicators as measures of events that have already happened, such as the number or rate of injuries, illnesses and fatalities.

A common problem is tracking no leading indicators at all, or tracking 30 and acting on none. OSHA's guide notes that some employers find it helpful to limit how many leading indicators they use at any one time, and suggests starting with one or two in a programme's first year. Our own suggestion, once you are past that stage, is 5-7 carefully chosen leading indicators, reviewed weekly, with clear ownership.

This guide lists 10 leading indicators, with formulas, example targets and how to track each. The order is not a ranking: we have not found published evidence that would rank them by how well they predict incidents.

Why Leading Indicators Matter

The safety triangle is usually traced to H. W. Heinrich's Industrial Accident Prevention, first published in 1931. In the 1941 second edition, Heinrich estimated that in a group of 330 similar accidents, 300 cause no injury, 29 cause minor injuries and 1 causes a major or lost-time injury. He presented the ratio as an estimate, and it should not be read as a constant for your site. The point it makes is the one that matters here: for every accident that injures someone there are many similar ones that do not, so counting only injuries misses most of the signal.

Leading indicators, in OSHA's words, measure the events leading up to injuries, illnesses and other incidents. Tracked correctly, they let you intervene before the recordable.

The Top 10 Leading Indicators

The example targets below are our own starting points, not benchmarks. OSHA's guide sets goals only inside its worked examples, and no regulator we checked publishes benchmark values for these measures. OSHA's method is to choose an indicator, set a goal for it, measure progress, and periodically reassess both the goal and the indicator. Set your targets from your own baseline.

1. Near-miss reporting rate

Formula: (Near misses reported × 200,000) ÷ Total hours worked. The 200,000-hour base is the one BLS uses for injury and illness incidence rates: 100 employees working 40 hours a week for 50 weeks. Example target: 10× your TRIR (a 10:1 near-miss-to-recordable ratio) Why it matters: A high near-miss rate can mean workers feel safe to report and supervisors are looking; a low rate may mean problems are happening but not being reported, though it can also reflect fewer hazards. Read it with the trend and the reporting breadth below. See our near-miss reporting software guide.

2. Safety observation completion rate

Formula: Observations completed / Observations scheduled × 100 Example target: 90%+ completion, with 70% identifying at least one at-risk behavior Why it matters: Completion measures how consistently the observation programme is run. A long run of zero-finding observations is worth a conversation, because it can mean observers are not looking closely. See our safety observation software guide and the behavior-based safety software page.

3. Corrective action close-out rate (and aging)

Formula: Actions closed within target ÷ Total actions opened × 100 Example target: 90% closed within 30 days; aged-action backlog under 5% Why it matters: For example, a site that opens 1,000 actions and closes 500 is left with a backlog of 500; one that opens 200 and closes 195 is left with 5. Backlog kills credibility. OSHA's guide lists a similar measure: the percentage of hazards abated on the same day, week or month in which they were identified.

4. Audit completion rate

Formula: Audits completed in period ÷ Audits planned × 100 Example target: 95%+ across self-audits; 100% for regulatory audits Why it matters: Skipped audits don't mean conditions are good; they mean you stopped looking. The safety audit checklist template is one starting point for the audits themselves. OSHA's guide lists the percentage of daily, weekly or monthly inspections completed as a leading indicator.

5. Training currency rate

Formula: Workers current on required training ÷ Total workers × 100 Example target: 98%+ for safety-critical roles Why it matters: OSHA's guide lists the percentage of workers receiving mandatory training on schedule as a leading indicator for education and training. See our training management software guide.

6. Permit compliance rate

Formula: Compliant permits ÷ Total high-risk tasks performed × 100 Example target: 99%+ for high-risk tasks (hot work, confined space, working at height) Why it matters: Permits exist for the high-hazard tasks listed above, so a task done without a compliant permit is high-hazard work without the control you designed for it. See our permit-to-work software guide.

7. Risk register currency

Formula: Risk assessments reviewed within last 12 months ÷ Total × 100 Example target: 95%+ within review window Why it matters: Stale risk registers reflect a system that was real once but has decayed. Frequent review keeps controls live.

8. Pre-task hazard analysis (PTHA / JSA) participation

Formula: Workers attesting JSA participation ÷ Workers performing tasks × 100 Example target: 95%+ for tasks requiring JSA Why it matters: OSHA's guide lists the number of workers involved in developing task-specific job safety analyses or job hazard analyses as a leading indicator of worker participation. See our job safety analysis software guide.

9. Leadership safety walks

Formula: Safety walks completed by senior leaders / Walks scheduled × 100 Example target: 90%+, with a minimum 1 walk/leader/month Why it matters: OSHA's guide suggests measures of management leadership such as the number of times each month that top management initiates discussion of a safety and health topic. Track walks like a KPI, not a nice-to-have.

10. Worker engagement (sentiment + reporting trust)

Formula: % workers agreeing "I feel safe reporting concerns without retaliation" (annual survey) Example target: 85%+ Why it matters: A workforce that doesn't trust the system stops reporting (see the safety culture assessment page). OSHA's guide lists the average score on survey questions about workers' perception of management's safety and health commitment as a leading indicator.

Two further measures for a near-miss programme

If near-miss reporting is one of your chosen indicators, two companion measures show whether the reporting system itself is healthy. Set their targets from your own baseline.

• Reporting breadth. Formula: workers who submitted at least one report or observation in the period ÷ total workers × 100. A high report count from a few people is a different programme from a moderate count spread across the workforce. OSHA's guide lists several participation counts in the same spirit, such as the number of workers participating in toolbox talks and in injury investigations.
• Time to acknowledge and act. Formula: average time from a worker's report to management's acknowledgement, and the share of reports on which corrective action started within 48 hours. Both are in OSHA's guide as measures of management leadership: the average time between a worker's report of a hazard or concern and management's acknowledgement, and the number of worker-reported hazards or concerns for which corrective action began within 48 hours, compared with the total reported. Our near-miss reporting guide sets out a response framework.

A worked example with invented numbers

The figures below are made up to show the arithmetic. A site with 52,000 worked hours in a month records the following.

Read the table the way OSHA's guide reads its own examples: choose a measure, set a goal, check progress, then reassess the goal and the measure. Here the on-time action rate is the weakest of the five against the example targets above, so a team limited to one or two indicators in its first year might pick that and observation completion. The result does not say the site is unsafe or safe; it says where the follow-through is thinnest. Pair each result with the lagging figure from your TRIR and LTIFR calculator and the TRIR and LTIR guide, so the two views are on one page, and keep the TRIR and LTIFR definitions beside the chart.

Writing an indicator definition that survives a staff change

A measure is only useful if the next person calculates it the same way. For each indicator you adopt, write down:

1. The name and the one-sentence question it answers.
2. The numerator and denominator, with the system or form each comes from.
3. What counts and what does not (for example, whether a near miss logged by a contractor is included).
4. The owner, the review interval and the date the definition was last checked.
5. The action that follows a miss, such as a review with the supervisor after two consecutive weeks below target.
6. The next review of the indicator itself, since OSHA's method includes reassessing the indicator and not only the goal.

The same rules apply to the broader list of measures on the QHSE KPIs and metrics page, and a stop work programme can supply several of the inputs (see the stop work authority guide).

How to Pick YOUR Top 5

You can't act on 10. Our suggestion is 5-7 at most. Use this prioritization:

1. Start with 1 cultural indicator (worker engagement or leadership walks)
2. Add 1-2 hazard-control indicators (permit compliance, JSA participation)
3. Add 1-2 system-health indicators (near-miss rate, action close-out)
4. Add 1 audit/training indicator (training currency or audit completion)

Our suggestion is a weekly review for the indicators you choose. OSHA's worked examples count some indicators monthly, so match the interval to how fast the hazard or the process changes, and say in the indicator definition what the interval is.

How to Track These in Software

Ask each EHS platform to demonstrate the selected indicators, including whether it can:

• Auto-calculate rates from incident, observation, and audit modules
• Push live dashboards to site leadership
• Alert when targets miss for 2 consecutive weeks
• Drill down to root cause (which crew, which task, which shift)

A spreadsheet can hold one or two indicators for a single site. As the number of indicators and sites grows, the work of collecting, reconciling and chasing owners grows with it, so test the effort in your own set-up before deciding what to buy. The capability map shows which dashboard and action-tracking capabilities the researched profiles document.

What to ask a vendor to demonstrate

The site does not rank products, and no product here is shown to predict incidents better than another. Ask each vendor to build the same three indicators from sample records and to show:

• The formula in the product, and whether you can edit the numerator and denominator.
• Drill-down from the figure to the underlying records, and who can see those records.
• How a change to the definition is dated, so a trend line does not silently change meaning.
• Whether near-miss, observation and action data sit in one reporting layer or need an export.

The demo checklist and the near-miss reporting software page give a fuller list, and the glossary entry on near misses sets out the term. For events with serious injury potential, the SIF glossary entry explains why a count of minor events may not describe the risk of the serious ones.

Common Mistakes

1. Tracking too many indicators. OSHA's guide notes that some employers find it helpful to limit how many they use at any one time.
2. Setting targets without baseline — you don't know what 90% completion means until you know what your team currently does.
3. No one owns the indicator — orphaned KPIs go nowhere.
4. Reporting only quarterly — by quarter-end the warning signs are already incidents.
5. Treating leading indicators as compliance theatre. High observation counts with zero findings show activity, not control; ask what changed because of the measure.

FAQs

How many leading indicators should I start with?

OSHA's guide suggests starting with one or two in the first year of a leading-indicator programme, and notes that some employers find it helpful to limit how many they use at any one time. Our own suggestion of five to seven is for a programme past that stage.

Should I track leading or lagging indicators?

Both. Lagging indicators are the score; leading indicators are the play-by-play. OSHA's guide puts it this way: a good programme uses leading indicators to drive change and lagging indicators to measure effectiveness.

What's the right ratio of near misses to recordables?

We know of no published benchmark we could verify. Heinrich's estimate (300 no-injury accidents for every 30 injuries) is sometimes quoted, but it comes from data collected nearly a century ago. Compare your ratio with your own history instead: a low or falling ratio can mean near misses are not being reported, not that they aren't happening.

How do I get supervisors to actually do safety walks?

Make them part of the leader's documented job (not optional), give them a simple mobile checklist, and review walk completion at the same management cadence as production metrics.

Are these indicators ISO 45001 compliant?

ISO 45001:2018 is a voluntary standard, and no indicator is compliant on its own. Clause 9.1 of the standard is headed "Monitoring, measurement, analysis and performance evaluation", as the contents of ISO's own handbook for small organizations show. The requirement text is sold rather than published, so ask your certification body how it will assess the indicators you choose; well-defined leading indicators can be part of the evidence you present. See our ISO 45001 software guide.

Can I benchmark these against industry peers?

Lagging indicators can be benchmarked: BLS publishes incidence rates by industry and establishment size, and its guide to computing your firm's rate explains how to compare yours. Leading indicators are mostly company-specific, so benchmark yourself against your own past performance month-over-month.

---

Preparing a shortlist? The Get Matched quiz returns up to five profiles using priority, industry and team-size tags. It does not verify dashboard features. Ask each vendor to build your selected indicators from sample records and demonstrate definitions, permissions and drill-down to source evidence.

Sources

• US OSHA: Using Leading Indicators to Improve Safety and Health Outcomes (OSHA 3970, June 2019) and its leading indicators page
• US Bureau of Labor Statistics: How to compute your firm's incidence rate for safety management
• H. W. Heinrich: Industrial Accident Prevention, second edition (McGraw-Hill, 1941), Internet Archive
• ISO: ISO 45001:2018, a practical guide for small organizations, preview with contents and the ISO 45001:2018 standard page

Corrected 23 September 2026: statistics without a primary source were sourced, labelled as examples or removed, and the guidance and standard cited now link to their official texts.

Revised 4 October 2026: every OSHA 3970 statement, the BLS 200,000-hour base and the ISO 45001 clause 9.1 heading were re-checked against the sources and stand. Overstated readings (what a high or low near-miss rate "means", zero-finding observations, "software is effectively required", "weekly not monthly") were softened to match the evidence, and a worked example, an indicator-definition checklist and vendor questions were added.

Updated 24 September 2026: merged the sourced sections of Near-Miss Reporting & Leading Indicators: A BBS Playbook.