Home / ESG and Compliance / Safety Incident Rate Calculator
ESG and Compliance
Safety Incident Rate Calculator (TRIR, DART, LTIFR, Severity – OSHA and ISO)
Turn hours worked and a count of injuries into the safety rates a report actually asks for. Pick the convention your standard uses, enter the hours worked and the case counts, and the tool returns the headline rate with the full metric list beside it. On the OSHA 200,000 hour base that is TRIR, the DART rate, LTIFR, and the severity rate. On the ISO and ILO 1,000,000 hour base it is the frequency index and the severity index. On the Brazil NBR 14280 base it is the taxa de frequencia and the taxa de gravidade, with the debited days from the NBR table added to the days actually lost. The metric list adapts to the convention you pick, a note warns you about the two bases and the factor of five between them, and a bar chart draws the rates with the US average line on the OSHA view. Every value stays in your browser.
This is the second flagship of the ESG and Compliance set, the safety side of a compliance report. It exists because a single injury record gets reported several ways depending on who is asking. A US site under OSHA 29 CFR 1904 reports a TRIR and a DART rate on a 200,000 hour base. A site certified to ISO 45001 or reporting under the ILO reports a frequency and severity index on a 1,000,000 hour base. A Brazilian site under NBR 14280 reports a taxa de frequencia and a taxa de gravidade, and the severity charges fixed debited days for a death or a permanent impairment. The arithmetic is the same multiply in every case; only the base and the definitions change. The tool holds those definitions so you enter the counts once and read the rate your standard wants, without mixing two bases that are not interchangeable. It is free, needs no sign-up, and runs entirely in your browser.
In short: a rate is cases times the base divided by hours worked. On the OSHA base of 200,000 hours, TRIR = (recordable cases times 200000) / hours, DART = (DART cases times 200000) / hours, LTIFR = (lost-time cases times 200000) / hours, and the severity rate = (days lost times 200000) / hours. On the ISO and ILO base of 1,000,000 hours the frequency index IF = (lost-time accidents times 1000000) / hours, and the classical severity index IG = (days lost times 1000) / hours. Under Brazil NBR 14280 the taxa de gravidade TG = (days lost plus debited days) times 1000000 / hours, where debited days are fixed charges for death and permanent impairment. The default OSHA set, 5 recordable, 3 DART, 2 lost-time, and 40 days lost in 500,000 hours, gives TRIR 2.00, DART 1.20, LTIFR 0.80, and severity 16.00. A rate on the 1,000,000 base is exactly 5 times the same rate on the 200,000 base, so never compare a TRIR against an IF or a TF directly.
TRIR total recordable incident rate
2.00TRIR total recordable incident rate
- TRIR (recordable rate)
- 2.00
- DART rate
- 1.20
- LTIFR (lost-time, 200,000 h)
- 0.80
- Severity rate (200,000 h)
- 16.00
A rate on the 1,000,000 hour base is 5 times the same rate on the 200,000 hour base, so never compare a TRIR against an IF or TF directly. TRIR 2.00: near the US private-industry average of about 2.4.
How the calculator works
The tool does one calculation, cases times a base divided by hours worked, and repeats it for each metric your convention reports. You pick the convention at the top, which sets the hour base and decides which metrics show. You enter the hours worked for the period and the case counts: recordable cases, DART cases, lost-time cases, and the days lost. On the NBR view a debited days field and a lesion selector appear so you can add the fixed NBR charges for a death or a permanent impairment. It multiplies each count by the base, divides by the hours, and lists the results with the headline rate at the top and the bar chart underneath.
The OSHA family uses a base of 200,000 hours, which is 100 full-time workers times 40 hours times 50 weeks, so a rate reads as cases per 100 workers per year. TRIR, the total recordable incident rate, is (recordable cases times 200000) / hours worked. The DART rate, for days away, restricted, or transferred, is (DART cases times 200000) / hours worked. LTIFR, the lost-time injury frequency rate, is (lost-time cases times 200000) / hours worked. The severity rate is (days lost times 200000) / hours worked. On the default set of 5 recordable, 3 DART, and 2 lost-time cases with 40 days lost in 500,000 hours, that is TRIR 2.00, DART 1.20, LTIFR 0.80, and a severity rate of 16.00.
The ISO and ILO family uses a base of 1,000,000 hours. The frequency rate IF is (lost-time accidents times 1000000) / hours worked. The classical severity index IG is (days lost times 1000) / hours worked, and the 1,000 base is not a typo: the IG is days lost per thousand hours, not per million, so it reads as a small decimal next to the frequency index. Because the frequency base is a million and the OSHA base is 200,000, the same lost-time record is exactly five times larger under ISO than under OSHA. That factor of five is the whole reason the tool keeps the conventions apart.
The Brazil NBR 14280 family also uses a base of 1,000,000 hours, but it charges severity differently. The taxa de frequencia TF is (accidents with lost time times 1000000) / hours worked. The taxa de gravidade TG is (computed time times 1000000) / hours worked, where computed time is the days actually lost plus the debited days. Debited days are fixed charges from the NBR table for outcomes that involve little calendar time but the worst harm: 6,000 days for a death or a permanent total disability, 4,500 for an arm or a leg, 3,000 for a hand, 1,800 for an eye, and smaller values down to 300 for a finger. You enter them in the debited days field or pick them from the lesion selector, and the tool adds them to the real days off before it computes the TG.
Two checks run under the arithmetic. DART cases are a subset of recordable cases and lost-time cases are a subset of DART, so DART is always at most TRIR and LTIFR is at most DART; if you enter more DART cases than recordable cases the tool flags it, because the counts are impossible rather than merely high. Near-miss reports and first-aid only injuries are not recordable, so they belong in none of these counts. The note under the result carries the base warning and, on the OSHA view, compares the TRIR against the US private-industry average of about 2.4; on the NBR view it reads the TG against the ILO severity scale instead.
What a recordable case is
Every rate on this page starts from a count of cases, so the count has to follow the definition or the rate is built on sand. Under OSHA 29 CFR 1904, a case is recordable if it is a work-related injury or illness that results in death, days away from work, restricted work or job transfer, medical treatment beyond first aid, loss of consciousness, or a significant diagnosis by a physician. First aid, by the OSHA list, is not recordable: cleaning a wound, a bandage, a non-prescription painkiller at over-the-counter dose, or a tetanus shot do not make a case recordable on their own. Near-miss events, where nobody was hurt, are worth reporting for prevention but are never recordable cases.
Within the recordable set sit the smaller subsets the DART and lost-time metrics count. A DART case is a recordable case that led to days away from work, restricted work, or a job transfer, so it is the more serious portion of the recordable log. A lost-time case is narrower still, a case that led to actual days away from work. Because each set nests inside the last, the counts have a fixed order: lost-time cases are at most DART cases, and DART cases are at most recordable cases. The tool uses that order as a validity check, and you can use it yourself to catch a miscount before it reaches a report.
Getting the definitions right matters more than the arithmetic, because the multiply is trivial and the classification is where mistakes happen. A common error is to count first-aid visits as recordable, which inflates the TRIR, or to count every recordable case as lost-time, which inflates the severity metrics. Pull the counts from your OSHA 300 log or the equivalent accident record, keep first-aid and near-miss events out of the recordable count, and confirm that your DART and lost-time counts are subsets of it. The tool cannot know whether a case was correctly classified; it can only check that the counts nest, so the judgment about each case still sits with you.
TRIR, DART, LTIFR, and the severity rate
The four OSHA metrics answer different questions from the same log. TRIR is the broadest, counting every recordable case, and it is the number most often quoted as a site’s overall safety rate. The DART rate narrows to cases serious enough to keep someone away from their normal work, so it tracks the injuries that actually cost days, and a large gap between TRIR and DART says most of your recordable cases were treated and the worker stayed on the job. LTIFR narrows again to lost-time cases only, the injuries that put someone off work entirely.
The severity rate is different in kind. TRIR, DART, and LTIFR count events, so they measure how often injuries happen. The severity rate counts days lost rather than cases, so it measures how bad the injuries were. A site can have a low TRIR and a high severity rate if it has few injuries but the ones it has are serious, or a high TRIR and a low severity rate if it has many minor recordable cases that cost few days. Reading a frequency metric and a severity metric together is what tells you whether your problem is the number of injuries or their seriousness, which point to different fixes.
On the default set the four metrics come out at TRIR 2.00, DART 1.20, LTIFR 0.80, and a severity rate of 16.00, all from the same 500,000 hours. The DART of 1.20 is at most the TRIR of 2.00, and the LTIFR of 0.80 is at most the DART, so the counts nest correctly. The severity rate of 16.00 reads on a different scale because it counts 40 days rather than a handful of cases; it is not comparable to the frequency numbers and should never be read as if it were another incidence rate. Label each metric so a reader does not put the severity rate next to the TRIR and conclude the site is eight times worse than it is.
The two hour bases and the factor of five
The single most important thing to get right on this page is that the two hour bases are not interchangeable. OSHA and the US metrics use 200,000 hours. ISO, the ILO, and Brazil NBR 14280 use 1,000,000 hours. Because the base sits in the numerator of every rate, a rate on the 1,000,000 base is exactly five times the same rate on the 200,000 base. The safety record does not change; only the number you print changes. So a frequency index of 5 under ISO and a TRIR of 1 under OSHA can describe the identical site, and reading the frequency of 5 as five times worse is simply a units mistake.
The example makes it concrete. Two lost-time cases in 500,000 hours give an LTIFR of 0.80 on the 200,000 base and a frequency of 4.00 on the 1,000,000 base. The second number is five times the first, and both come from the same two cases and the same hours. If you saw the 4.00 next to the 0.80 without knowing the bases, you might think two different sites were being compared, or that safety had deteriorated fivefold. It had not; the analyst simply switched conventions. When you have to move a rate from one base to the other, multiply or divide by five rather than recomputing from a factor you half-remember.
This is why the tool asks for the convention first and adapts the metric list to it, and why the note under the result repeats the base warning. It is easy to pull a TRIR from a US site and a frequency index from a European or Brazilian site and drop both into one benchmark table as if they were the same measure. They are not, and the resulting comparison overstates one site by a factor of five. Keep every number in one convention across a comparison, state the base next to each rate, and convert with the factor of five when you genuinely need to move between them.
NBR 14280 severity and debited days
Brazil NBR 14280 handles severity in a way the OSHA severity rate does not, and it is the reason the standard has its own view in the tool. The taxa de gravidade is not just the days a worker was actually off; it is the computed time, which adds fixed debited days for a death or a permanent impairment to the real days lost. A death may involve very few calendar days away, yet it is the worst possible outcome, so counting only the days off would score a fatal year as mild. The debited days fix that by charging a fixed number of days for the outcome itself.
The NBR table sets those charges. A death or a permanent total disability is charged 6,000 days. Losing an arm or a leg is charged 4,500, a hand 3,000, a foot 2,400, an eye 1,800, a thumb 600, and a finger 300, with the full table covering the range of permanent impairments. These are added to the days actually lost to form the computed time, and the TG is that computed time times 1,000,000 divided by the hours worked. You enter the debited days directly or pick the injury from the lesion selector, which fills the charge for you, so a fatal or disabling event lands in the severity with the weight the standard intends.
The result is a severity number that reflects harm rather than the calendar. On the NBR example, 5 accidents with lost time in 1,000,000 hours give a taxa de frequencia of 5.00, and the severity adds 150 days actually lost to 6,000 debited days for one death, for 6,150 of computed time and a taxa de gravidade of 6,150. By the ILO scale a TG over 2,000 is severe, so 6,150 reads as severe, driven almost entirely by the debited days for the death. That is the point of charging them: without the 6,000 day charge the same year would show a TG of 150, which would read as very good and hide the fatality completely.
Benchmarking against BLS or the ILO scale
A rate means little on its own; it means something next to the right benchmark. For OSHA metrics the anchor is the US private-industry average TRIR, which is about 2.4 according to the BLS. As a rough reading, a TRIR under 1.0 is strong, near 2.4 is average, and above about 2.8 is above the average and worth attention. But the all-industry average blends a warehouse and a foundry, which live at very different levels, so the honest comparison is against the average for your own industry code, not the single national figure. The tool notes the 2.4 anchor on the OSHA view as a starting point, not a target for every site.
For the NBR taxa de gravidade the benchmark is the ILO severity scale, which the tool applies on the NBR view. It reads a TG of 500 or less as very good, 500 to 1000 as good, 1000 to 2000 as poor, and over 2000 as severe. The NBR example TG of 6,150 sits well past 2,000, so it reads as severe, which matches the presence of a fatality in the year. The scale gives the severity number a plain reading rather than leaving it as a bare figure, so someone without the NBR background can still tell a very good year from a severe one.
Whichever benchmark you use, treat these as lagging indicators. Every metric on this page counts harm that already happened, so a good number this year is a record of the past, not a guarantee for next year, and a single serious event can move a small site’s rate sharply. Pair the lagging rates with leading indicators, the measures that come before an injury, so the picture is not built only on outcomes. The next section covers that pairing.
Leading versus lagging safety indicators
The rates on this page are lagging indicators. TRIR, DART, LTIFR, the severity rate, the frequency index, and the taxa de gravidade all count events that have already caused harm, so they tell you what happened, not what is about to. That makes them essential for reporting and comparison and weak for prevention, because by the time a lagging indicator moves, someone has already been hurt. A site that manages only by its TRIR is steering by the rear-view mirror.
Leading indicators come before the injury and predict it. Near-miss reports, safety inspections completed, hazards found and closed out, training hours delivered, and observations of safe or unsafe behavior all measure the conditions that produce injuries rather than the injuries themselves. They are harder to standardize than a TRIR and easier to game, but they are the numbers you can act on before an event, and a rising near-miss reporting rate usually reflects a workforce that is engaged rather than a site that is getting more dangerous.
The practical stance is to report both and read them together. Use the lagging rates from this tool for the compliance report and the year-on-year comparison, and use leading indicators to manage the risk day to day. A low TRIR with a healthy near-miss reporting rate is a genuinely strong position; a low TRIR with almost no near-miss reports may just mean people have stopped reporting. The incident record alone cannot tell those two apart, which is why the leading indicators sit alongside the rates this calculator produces rather than being replaced by them.
Preparing the OSHA 300 log or the accident record
The rate is only as good as the counts, so the record behind it deserves the care. For a US site the source is the OSHA 300 log, where each recordable case is entered with its classification as days away, restricted or transferred, or other recordable. Pull the recordable count, the DART count, and the lost-time count straight from the log for a defined period, usually the calendar year, and take the hours worked from payroll for the same period so the numerator and the denominator cover the same span. A mismatch between the injury period and the hours period is a common and quiet source of error.
For sites outside the US the record is the equivalent accident register: the CAT in Brazil, filed for each accident, with the days lost and, for the taxa de gravidade, the debited days from the NBR table for any death or permanent impairment. Under ISO 45001 the organization keeps its own incident record and reports the frequency and severity index from it. Whatever the record, the same discipline applies: count only recordable cases, keep first-aid and near-miss events out, confirm the DART and lost-time counts nest inside the recordable count, and record the hours worked for the matching period.
A note for Mexico: the IMSS indice de siniestralidad and the prima de riesgo are a different calculation, not a frequency or a severity index. They feed the employer’s workers-compensation premium through their own formula and should not be mixed with the TRIR, the frequency index, or the taxa de gravidade on this page. If you work under STPS and NOM-019 you will still recognize the frequency and severity concepts here, but keep the IMSS premium calculation separate; a future tool may cover it on its own.
Five worked examples
Example 1: the OSHA set
This one runs the full OSHA panel on one site so every metric is visible. With 5 recordable cases, 3 DART cases, 2 lost-time cases, and 40 days lost in 500,000 hours, the tool returns TRIR 2.00, DART 1.20, LTIFR 0.80, and a severity rate of 16.00. The DART of 1.20 is at most the TRIR of 2.00, which passes the sanity check that DART cannot exceed the recordable count, and the TRIR of 2.00 sits just below the US private-industry average of about 2.4. The lesson is that one input set produces every OSHA metric, and the DART-to-TRIR check catches a miscount before it reaches the report.
Example 2: the same cases on both bases
This case keeps the lost-time cases fixed and changes only the base, to show the factor of five. Two lost-time cases in 500,000 hours are an LTIFR of 0.80 on the 200,000 base and a frequency of 4.00 on the 1,000,000 base. The second number is exactly five times the first, because the base in the numerator is five times larger, not because anything about the site changed. The lesson is that the two conventions describe the same site: a frequency of 4.00 under ISO and an LTIFR of 0.80 under OSHA are not a difference in safety, only in the base, so you convert between them with the factor of five rather than reading one as safer.
Example 3: the ISO and ILO view
This case shows the frequency and severity indices together. Eight lost-time accidents in 500,000 hours give a frequency index IF of 16.00, and 150 days lost give a severity index IG of 0.30 on the classical days-per-thousand-hours base. The same data also reports a TRIR of 4.80 and a DART of 2.00 from 12 recordable and 5 DART cases. The lesson is that the IG uses the 1,000 base, so it reads as a small decimal, while the IF uses the million base and reads as 16.00; label them clearly so the 0.30 severity is not read as a frequency or mistaken for a rounding error.
Example 4: Brazil NBR 14280 with a fatality
This case shows the debited days at work. Five accidents with lost time in 1,000,000 hours give a taxa de frequencia TF of 5.00. The severity adds the 150 days actually lost to the 6,000 debited days charged for the death, for 6,150 of computed time, so the taxa de gravidade TG is 6,150. By the ILO scale that TG is severe, well over 2,000. The lesson is that the debited days from the NBR table dominate the severity when there is a death or a permanent impairment: without the 6,000 day charge the same year would show a TG of 150 and read as very good, which is exactly the outcome the debited days exist to prevent.
Example 5: a validity check
This case shows what the sanity check catches. Entering 5 DART cases against only 2 recordable cases is impossible, because DART cases are a subset of recordable cases, so DART can never exceed the recordable count, and the tool flags the entry rather than computing a rate from it. Near-miss reports and first-aid only injuries are also not recordable, so they must be left out of every count. The lesson is that the case definitions matter as much as the arithmetic: a rate built on the wrong counts is worse than no rate, because it looks authoritative while being wrong.
Three expert tips
Match the base to your standard and never compare across bases
A US site reports on the OSHA 200,000 hour base as a TRIR and a DART rate, while a site under ISO 45001 or Brazil NBR 14280 reports on the 1,000,000 hour base as a frequency and a severity index. The same accident record is exactly five times larger on the million base, so a frequency of 5 and a TRIR of 1 can be the same site. The two-base example makes the point, with two lost-time cases reading 0.80 as an LTIFR and 4.00 as a frequency. Pick the convention first, keep every number in one base across a comparison, and convert with the factor of five rather than reading the larger number as the more dangerous site.
Count severity with debited days, not just days off
A death or a permanent total disability may involve few real days away, yet it is the worst possible outcome, and a severity metric that counts only calendar days would score it as mild. The NBR 14280 severity charges a fixed 6,000 days for a death and set values for lost limbs and sight, added to the actual lost days, so the number reflects the harm and not just the calendar. The fatality example shows it: 150 real days plus 6,000 debited days give a TG of 6,150, severe on the ILO scale, where the days off alone would have shown 150. Enter the debited days from the table, or use the lesion selector, so a fatal year is not scored as a quiet one.
Grade the number against the right benchmark and treat it as lagging
A TRIR is only meaningful next to the average for your own industry, because a warehouse and a foundry live at different levels; the all-industry US average near 2.4 is a rough anchor, not a target for every site. Use your industry’s figure where you have it, and read the NBR taxa de gravidade against the ILO scale, where over 2,000 is severe. And remember these are lagging indicators that count harm already done, so pair them with leading indicators like near-miss reporting and inspections. A low TRIR with almost no near-miss reports may mean people have stopped reporting rather than that the site is safe, and the incident record alone will not show that.
Common mistakes to avoid
The first mistake is comparing rates across bases. A frequency index on the 1,000,000 base is five times the equivalent LTIFR on the 200,000 base, so dropping a US TRIR and a European frequency index into one table overstates one site fivefold; keep every number in one convention and convert with the factor of five. The second is miscounting cases: counting first-aid or near-miss events as recordable inflates the TRIR, and counting every recordable case as lost-time inflates the severity metrics. Pull the counts from the OSHA 300 log or the accident record and confirm that DART and lost-time counts nest inside the recordable count.
A third mistake is scoring severity by days off alone under NBR, which lets a fatality read as mild; add the debited days from the table so the 6,000 day charge for a death lands in the taxa de gravidade. A fourth is mismatching the injury period and the hours period, so the numerator covers the year but the denominator covers a quarter; take both from the same defined span. A fifth is reading a lagging rate as if it predicted the future, or comparing a TRIR against the all-industry average instead of your own industry. Match the base, count cases correctly, charge the debited days, align the periods, and benchmark against the right industry, and the rates will hold up.
Where this calculator fits
It suits anyone who has hours worked and a count of injuries and needs the rate a standard asks for. A safety or EHS manager can enter the year’s counts, read the TRIR, DART, LTIFR, and severity together, and produce the numbers for a compliance report or a board pack. An ISO 45001 or NBR 14280 site can switch the convention and read the frequency and severity index on the million base, with the debited days charged for any death or permanent impairment. An ESG or compliance analyst can put the safety rates next to the carbon numbers from the rest of this silo, since both belong in the same disclosure.
This is the second flagship of the ESG and Compliance set, the safety side of it. The live Scope 1 and 2 Carbon Footprint Calculator covers the emissions side, and the two together give a compliance report its environmental and safety numbers. The Electricity CO2e Calculator, the Fuel Combustion CO2e Calculator, the Carbon Intensity Calculator, and the Waste Diversion Rate Calculator are planned tools in this silo and are named here for reference; they are not yet live, so treat them as the workflow this hub will grow into. The ESG and Compliance hub gathers the tools as they go live, and the Scope 1 and 2 Carbon Footprint Calculator is the emissions flagship alongside this one.
Frequently asked questions
What does this safety incident rate calculator do?
It turns hours worked and a count of injuries into the safety rates a standard reports. You pick the convention, which sets the hour base, then enter the hours worked and the case counts: recordable cases, DART cases, lost-time cases, and days lost, plus debited days on the NBR view. On the OSHA 200,000 hour base it returns TRIR, the DART rate, LTIFR, and the severity rate. On the ISO and ILO 1,000,000 hour base it returns the frequency index and the classical severity index. On the Brazil NBR 14280 base it returns the taxa de frequencia and the taxa de gravidade, with the debited days added to the real days lost. On the default OSHA set, 5 recordable, 3 DART, 2 lost-time, and 40 days lost in 500,000 hours, it gives TRIR 2.00, DART 1.20, LTIFR 0.80, and severity 16.00. The metric list adapts to the convention, a note warns about the two bases, and a bar chart draws the rates. Every value stays in your browser.
How is TRIR calculated?
TRIR, the total recordable incident rate, is (recordable cases times 200000) divided by hours worked. The 200,000 base is 100 full-time workers times 40 hours times 50 weeks, so the rate reads as recordable cases per 100 workers per year. On the default set, 5 recordable cases in 500,000 hours give a TRIR of 2.00. TRIR counts every recordable case under OSHA 29 CFR 1904, meaning work-related injuries or illnesses that led to death, days away, restricted work or transfer, medical treatment beyond first aid, loss of consciousness, or a significant diagnosis. First-aid and near-miss events are not recordable and stay out of the count. Because DART and lost-time cases are subsets of the recordable count, the TRIR is the largest of the OSHA frequency metrics, and the DART rate and LTIFR are always at most the TRIR.
What is the difference between TRIR, DART, and LTIFR?
All three are frequency rates on the OSHA 200,000 hour base, and they differ by which cases they count. TRIR counts every recordable case, so it is the broadest. The DART rate counts only recordable cases that led to days away, restricted work, or a job transfer, so it tracks the more serious injuries that actually cost normal work. LTIFR counts only lost-time cases, the injuries that put someone off work entirely, so it is the narrowest. Because each set nests inside the last, DART is at most TRIR and LTIFR is at most DART. On the default set the three come out at TRIR 2.00, DART 1.20, and LTIFR 0.80, which nest correctly. A large gap between TRIR and DART means most recordable cases were minor enough that the worker stayed on the job.
What is the difference between the 200,000 and 1,000,000 hour bases?
They are two conventions for the same rate, and they are not interchangeable. OSHA and the US metrics use 200,000 hours, which is 100 workers times a work year. ISO, the ILO, and Brazil NBR 14280 use 1,000,000 hours. Because the base sits in the numerator, a rate on the 1,000,000 base is exactly five times the same rate on the 200,000 base. Two lost-time cases in 500,000 hours, for instance, are an LTIFR of 0.80 on the OSHA base and a frequency of 4.00 on the million base, the same two cases either way. So a frequency of 5 under ISO and a TRIR of 1 under OSHA can be the identical site. Never compare a rate on one base against a rate on the other directly; convert with the factor of five, and keep every number in one convention across a comparison.
How is the severity rate calculated?
On the OSHA base the severity rate is (days lost times 200000) divided by hours worked, so it counts days rather than cases and measures how bad the injuries were rather than how often they happened. On the default set, 40 days lost in 500,000 hours give a severity rate of 16.00. The ISO and ILO classical severity index IG uses a different base: it is (days lost times 1000) divided by hours worked, so it is days lost per thousand hours and reads as a small decimal, such as the 0.30 from 150 days in 500,000 hours. The Brazil NBR taxa de gravidade uses the million base and adds debited days: it is (days lost plus debited days) times 1000000 divided by hours worked. Because the severity metrics count days, not cases, they read on a different scale from TRIR and should never be compared with the frequency rates directly.
What are debited days under NBR 14280?
Debited days are fixed charges the Brazil NBR 14280 severity adds for a death or a permanent impairment, separate from the days a worker was actually off. A death may involve few calendar days away yet is the worst outcome, so the standard charges a fixed number of days for the outcome itself: 6,000 for a death or a permanent total disability, 4,500 for an arm or a leg, 3,000 for a hand, 2,400 for a foot, 1,800 for an eye, 600 for a thumb, and 300 for a finger. These are added to the real days lost to form the computed time, and the taxa de gravidade is that computed time times 1,000,000 divided by hours worked. On the NBR example, 150 real days plus 6,000 debited days for one death give a TG of 6,150, severe on the ILO scale. Without the charge the same year would show a TG of 150 and hide the fatality.
How do I know if a TRIR is good?
Compare it against the right benchmark. The US private-industry average TRIR is about 2.4 according to the BLS, so as a rough reading a TRIR under 1.0 is strong, near 2.4 is average, and above about 2.8 is above the average and worth attention. But the all-industry average blends very different workplaces, so the honest comparison is against the average for your own industry code, not the single national figure, because a warehouse and a foundry live at different levels. On the default set a TRIR of 2.00 sits just below the 2.4 anchor. Remember too that a TRIR is a lagging indicator that counts harm already done, so a good number is a record of the past, not a promise for next year, and a single serious event can move a small site’s rate sharply. Read it alongside leading indicators like near-miss reporting.
What counts as a recordable case?
Under OSHA 29 CFR 1904 a case is recordable if it is a work-related injury or illness that resulted in death, days away from work, restricted work or job transfer, medical treatment beyond first aid, loss of consciousness, or a significant injury or illness diagnosed by a physician. First aid, by the OSHA list, is not recordable on its own: cleaning a wound, a bandage, a non-prescription painkiller at over-the-counter dose, or a tetanus shot do not make a case recordable. Near-miss events, where nobody was hurt, are worth reporting for prevention but are never recordable. Within the recordable set, a DART case is one that led to days away, restricted work, or transfer, and a lost-time case is one that led to actual days away, so the two are subsets of the recordable count. Counting first-aid or near-miss events as recordable inflates the TRIR, which is one of the most common errors.
What is the frequency index IF under ISO and the ILO?
The frequency index IF is (lost-time accidents times 1000000) divided by hours worked, on the 1,000,000 hour base that ISO 45001 and the ILO use. Eight lost-time accidents in 500,000 hours give an IF of 16.00. It is the same idea as the OSHA LTIFR but on the million base, so it reads five times larger for the same record. The paired classical severity index IG uses a 1,000 base instead, IG equals (days lost times 1000) divided by hours worked, so it is days lost per thousand hours and reads as a small decimal, such as 0.30 from 150 days in 500,000 hours. Because the IF is on the million base and the IG on the thousand base, they look very different in size, so label them clearly. A frequency index should never be dropped into a table next to an OSHA TRIR without noting that they sit on different bases.
Why does DART cases cannot exceed recordable cases matter?
Because DART cases are a subset of recordable cases, the DART count can never be larger than the recordable count, and lost-time cases are a subset of DART, so lost-time can never exceed DART. The tool uses that order as a validity check: if you enter more DART cases than recordable cases, it flags the entry rather than computing a rate, because the counts are impossible, not just unusual. The nesting is a quick way to catch a miscount by hand too. If your DART count comes out above your recordable count, or your lost-time above your DART, one of the counts is wrong, usually because a case was classified into two buckets or a first-aid event slipped into a subset. Fixing the count before the rate is computed is the point, since a rate built on impossible counts looks authoritative while being wrong.
What is the difference between leading and lagging safety indicators?
Every rate on this page is a lagging indicator: TRIR, DART, LTIFR, the severity rate, the frequency index, and the taxa de gravidade all count harm that already happened, so they report the past rather than predict the future. They are essential for compliance and comparison but weak for prevention, because by the time one moves, someone has been hurt. Leading indicators come before the injury and predict it: near-miss reports, inspections completed, hazards found and closed, training delivered, and safe-behavior observations. They measure the conditions that produce injuries rather than the injuries themselves. The practical stance is to report both: use the lagging rates from this tool for the report and the year-on-year comparison, and use leading indicators to manage risk day to day. A low TRIR with almost no near-miss reports may just mean people stopped reporting, which the incident record alone cannot show.
Does this work for a Brazil NBR 14280 or ISO 45001 report?
Yes. Switch the convention at the top. The ISO and ILO view uses the 1,000,000 hour base and reports the frequency index IF and the classical severity index IG, which suits an ISO 45001 report. The NBR 14280 view also uses the million base and reports the taxa de frequencia and the taxa de gravidade, and it shows the debited days field and the lesion selector so you can add the fixed NBR charges for a death or a permanent impairment. On the NBR example, 5 accidents with lost time in 1,000,000 hours give a TF of 5.00, and 150 real days plus 6,000 debited days for a death give a TG of 6,150, which the tool reads as severe on the ILO scale. The English acronyms TRIR and DART stay visible on the other views because practitioners search for them, but the metric list and the base follow whichever convention you pick.
Is the IMSS prima de riesgo the same as these rates?
No. The Mexican IMSS indice de siniestralidad and the prima de riesgo are a different calculation, not a frequency or a severity index. They feed the employer’s workers-compensation premium through their own formula and should not be mixed with the TRIR, the frequency index, or the taxa de gravidade on this page. If you work under STPS and NOM-019 you will recognize the frequency and severity concepts here, and you can use this tool for those, but keep the IMSS premium calculation separate because it answers a different question and uses different inputs. Putting an IMSS siniestralidad figure into a table next to a TRIR or a frequency index would compare two things that are not the same measure. A dedicated tool for the IMSS calculation is a candidate for a future addition to this silo.
Is the tool free, and does it work in my browser?
Yes to both. The safety incident rate calculator is free with no sign-up, and every calculation runs in your browser, so the hours and case counts you enter are never sent to a server, stored, or shared. You can download a PDF of the result, export a CSV, or share a summary on WhatsApp. It computes each rate as cases times the base divided by hours worked, adapts the metric list to the convention you pick, and draws a bar chart of the rates with the US average line on the OSHA view. It is a management and reporting tool, so use it to build the safety rates for a compliance report, benchmark them against the right industry or the ILO scale, and pair them with leading indicators, and confirm the case classifications against your OSHA 300 log or accident record before you report externally.
More ESG and compliance calculators
This safety incident rate calculator is the second flagship of the silo, the safety side of a compliance report. The emissions flagship is live and linked; the other tools are on the way and are listed for reference rather than linked.
The ESG and compliance tools work as a set. Build the safety rates here for the safety side of a report, build the emissions total on the carbon footprint tool for the environmental side, and add the electricity, fuel, intensity, and waste tools as they go live. The ESG and Compliance hub gathers them, and the Scope 1 and 2 Carbon Footprint Calculator is the emissions flagship alongside this one.
Sources, disclaimer, and editorial transparency
The relationships used here are standard occupational safety rate accounting. On the OSHA 200,000 hour base, TRIR = (recordable cases times 200000) / hours worked, DART = (DART cases times 200000) / hours worked, LTIFR = (lost-time cases times 200000) / hours worked, and the severity rate = (days lost times 200000) / hours worked. On the ISO and ILO 1,000,000 hour base, the frequency index IF = (lost-time accidents times 1000000) / hours worked, and the classical severity index IG = (days lost times 1000) / hours worked. Under Brazil NBR 14280 on the 1,000,000 hour base, the taxa de frequencia TF = (accidents with lost time times 1000000) / hours worked, and the taxa de gravidade TG = (days lost plus debited days) times 1000000 / hours worked, where the debited days are fixed NBR charges of 6,000 for a death or permanent total disability, 4,500 for an arm or a leg, 3,000 for a hand, 2,400 for a foot, 1,800 for an eye, 600 for a thumb, and 300 for a finger. A rate on the 1,000,000 base is exactly five times the same rate on the 200,000 base. The worked numbers are computed from the inputs shown: the OSHA set of 5 recordable, 3 DART, 2 lost-time, and 40 days lost in 500,000 hours gives TRIR 2.00, DART 1.20, LTIFR 0.80, and severity 16.00; two lost-time cases in 500,000 hours give an LTIFR of 0.80 and a frequency of 4.00; the ISO set of 8 lost-time accidents and 150 days with 12 recordable and 5 DART in 500,000 hours gives IF 16.00, IG 0.30, TRIR 4.80, and DART 2.00; and the NBR set of 5 accidents with lost time, 150 real days, and 6,000 debited days for a death in 1,000,000 hours gives TF 5.00 and TG 6,150. The US private-industry average TRIR of about 2.4 is from the BLS; the ILO severity scale reads a TG of 500 or less as very good, 500 to 1000 as good, 1000 to 2000 as poor, and over 2000 as severe. This calculator and guide are built and reviewed by the OpsCalculators team; see our Editorial Policy for how each tool is researched, built, and tested.
Results are accurate for the formulas and the definitions above, and they are a management and reporting estimate, not a certified regulatory filing. The tool computes rates from the counts you enter and cannot know whether each case was correctly classified under OSHA 29 CFR 1904, NBR 14280, or your own record; the judgment about each case stays with you. Pull the counts from the OSHA 300 log or the accident register, keep first-aid and near-miss events out of the recordable count, confirm that DART and lost-time counts nest inside the recordable count, take the hours worked from the matching period, and add the NBR debited days for any death or permanent impairment. The Mexican IMSS indice de siniestralidad and prima de riesgo are a separate calculation and should not be mixed with these rates. See our full Disclaimer. OpsCalculators.com is operated by MAFHH INTERNATIONAL LTD. Your inputs are processed in your browser and are never stored; see our Privacy Policy.