A recordable-injury rate tells you how many people already got hurt. It can never tell you who's about to.
Safety metrics fall into two categories that measure fundamentally different things in time. Lagging indicators measure outcomes that have already happened — injuries, illnesses, lost workdays, OSHA recordables — and are, by definition, historical: the harm is already done by the time the number moves. Leading indicators measure the conditions and behaviors believed to predict future incidents before they occur — near-miss reporting rates, percentage of scheduled safety audits completed, hazard corrective-action closure time, toolbox talk attendance. A site can run an entire quarter with a perfect lagging-indicator record (zero recordables) while its leading indicators — skipped inspections, ignored near-miss reports, expired PPE certifications — are quietly signaling that the next incident is overdue.
One of the most counterintuitive but well-established findings in safety management is that a sharp drop in near-miss and hazard reporting is very rarely evidence that hazards have actually decreased — it is almost always evidence that people have stopped reporting them, whether from reporting fatigue, fear of blame, or a culture that quietly discourages flagging problems. Because near-misses statistically vastly outnumber actual injuries for the same underlying hazard (this ratio is the basis of Heinrich's and later safety pyramid models), a healthy leading-indicator program treats a rising or stable near-miss report rate as a sign the safety culture is working — people feel safe raising issues before someone gets hurt — while a falling rate, especially alongside stagnant lagging indicators, should trigger investigation into whether reporting itself has broken down, not celebration that things have "gotten safer."
A zero-recordable quarter is a real and worth-celebrating outcome, but it is a laggingmeasure of luck and history combined — it says harm didn't occur in that specific window, not that the underlying risk was actually reduced. Small sample sizes make this worse: a single site with relatively few work-hours can post zero recordables for a year purely by chance, even while its leading indicators (unclosed hazard reports, unfilled safety-committee seats, expired equipment inspections) show clear deterioration. OSHA and most modern safety-management frameworks (including ANSI/ASSP Z10 and ISO 45001) explicitly recommend pairing lagging indicators with a real leading-indicator program precisely because relying on lagging metrics alone tends to produce exactly this false sense of security — reacting only after the pattern has already produced an injury, instead of catching the pattern beforehand.
Explains the OSHA/safety-management distinction between lagging indicators (TRIR, DART rate, and other post-injury metrics) and leading indicators (near-miss reporting, audit completion, corrective-action closure time), and why a falling near-miss report rate is usually a warning sign rather than good news.
Both categories show up on the same safety dashboard, both get reported to leadership in the same monthly review, and a casual read of "the numbers are good" often lumps them together as one undifferentiated safety score. The distinguishing factor is timing relative to harm: lagging indicators can only be computed after an injury or illness has already occurred, while leading indicators are computed continuously from process and behavioral data that exist whether or not any injury ever happens.
Lagging indicators are outcome-based measures derived from incidents that have already occurred — Total Recordable Incident Rate (TRIR), Days Away/Restricted/Transferred (DART) rate, lost-time injury frequency, and OSHA 300 log entries are the standard examples. They are essential for regulatory compliance (OSHA recordkeeping requirements under 29 CFR 1904), benchmarking against industry averages, and long-term trend analysis, but by construction they cannot warn anyone before the harm they measure has already happened.
Leading indicators are process- or behavior-based measures believed to correlate with future incident likelihood — near-miss/hazard report submission rates, percentage of scheduled safety inspections and audits actually completed, average time to close a corrective action, safety training completion rates, and management safety-walk frequency. Because these are measured continuously and independent of any injury occurring, they can be acted on before harm results.
ANSI/ASSP Z10 and ISO 45001 (occupational health and safety management system standards) both explicitly recommend tracking a balanced mix of leading and lagging indicators rather than managing safety performance off lagging metrics alone. OSHA itself, in its Recommended Practices for Safety and Health Programs, encourages proactive, leading-indicator-driven programs precisely because reactive management — waiting for the TRIR to rise before investigating — means the organization is always at least one injury behind. Engineering and construction firms operating under OSHA 10/30 training programs are taught to build hazard-identification and near-miss reporting systems specifically because they generate leading-indicator data no lagging metric can produce.
Total Recordable Incident Rate = (number of OSHA-recordable incidents × 200,000) ÷ total hours worked. The 200,000 figure represents 100 employees working 40 hours/week, 50 weeks/year, so TRIR is standardized to be comparable across companies of different sizes.
OSHA does not mandate specific leading indicators by regulation the way it mandates OSHA 300 log recordkeeping (a lagging-indicator requirement), but OSHA's own guidance and its Voluntary Protection Programs (VPP) strongly encourage and reward employers who track and act on leading indicators as part of a proactive safety management system.
They stay distinct by definition — a leading indicator that is never acted on doesn't "become" a lagging indicator, it simply fails to prevent whatever lagging outcome eventually results. The two categories describe when the measurement can be taken relative to harm, not a spectrum a single metric moves along.
There is no fixed number in OSHA guidance, but most mature safety-management programs pick a small, focused set (often 3–6) directly tied to the organization's specific hazard profile — tracking too many dilutes attention and makes it harder to identify which signal actually predicts incidents at that particular site.
Try our Engineering Management, Business & Professional Skills Studio
More calculators, simulators, and guides for this discipline.
Data Analysis Studio
✨ Premium ContentA structured, paid professional training program for this discipline.