Calculate incidence rate as new cases divided by person-time at risk, per 1,000 or 100,000, or solve for expected cases, person-time or cumulative risk.

Optional settings

Disclaimer: This calculator is for educational/statistical purposes and does not provide medical or public-health advice. For official definitions and guidance, consult resources from organizations such as the CDC or WHO.

Incidence Rate Formula

The incidence rate, also called incidence density, measures how quickly new cases occur across the total amount of time that people remain at risk. It is especially useful when participants are followed for different lengths of time.

IR = (NC) / (PT)

When you want the result on a standard reporting scale, multiply the basic rate by a constant such as 100, 1,000, or 100,000.

IRₖ = (NC) / (PT) × k
  • IR = incidence rate
  • NC = number of new cases during the time period
  • PT = total person-time at risk
  • k = reporting multiplier such as 100, 1,000, or 100,000

If person-time is entered in person-years, the result can be reported as cases per person-year, per 1,000 person-years, or per 100,000 person-years.

Rearranged Formulas

Choose the value to solve for. These rearrangements use an unscaled rate IR per one person-time unit; divide a displayed scaled rate by its reporting multiplier k before using it.

NC = IR × PT
PT = (NC) / (IR)

Comparing Two Groups: Incidence Rate Ratio

Select Incidence rate ratio to compare new cases per person-time in two groups. Enter each group’s whole-number case count and positive person-time. Group 1 is the numerator and Group 2 is the reference. The calculator converts both denominators to a common time unit and displays each rate on the selected reporting scale.

Rate ratio: IRR = (a / PT1) / (b / PT0)

Standard error: SE[ln(IRR)] = √(1/a + 1/b)

Approximate 95% CI: exp[ln(IRR) ± 1.96 × SE]

Here, a and b are the new case counts, and PT1 and PT0 are the matching person-time totals in the same units. An IRR of 1 means equal observed rates; 2 means Group 1’s rate is twice Group 2’s. This is a rate ratio, not a cumulative-risk ratio, and it does not establish causation.

For example, 40 cases over 2,000 person-years versus 20 cases over 2,000 person-years gives rates of 20 and 10 per 1,000 person-years. The IRR is 2, with an approximate 95% interval of 1.169269 to 3.42094.

The interval uses an approximate Poisson log-Wald method for independent counts and fixed person-time denominators. Sparse counts can make it unreliable; if either group has fewer than five cases, an exact method is preferable. No interval is calculated when either count is zero, and no arbitrary continuity correction is added. A zero reference rate has no finite ratio; two zero rates give an undefined ratio. Confounding, clustering and overdispersion need other analysis methods.

Sources: CDC: rate ratios and NIOSH evaluation guide, Appendix B, pages 105–106.

What Person-Time at Risk Means

Person-time is the sum of all observation time contributed by people while they are still at risk for the outcome being studied. This denominator is what makes incidence rate more flexible than a simple proportion.

  • 12 people followed for 1 year each = 12 person-years
  • 40 people followed for 6 months each = 20 person-years
  • 1 person followed for 3 months contributes 0.25 person-years

Only count time while a participant is actually at risk. Follow-up typically stops when the event occurs, the person leaves the study, the person is no longer at risk, or the study period ends.

How to Use the Incidence Rate Calculator

  1. Enter the number of new cases observed during the study period.
  2. Enter the total person-time at risk for the same period.
  3. Select the reporting scale you want, such as per 1 person-year, per 1,000 person-years, or per 100,000 person-years.
  4. If you already know the rate, enter the rate and one other variable to solve for the missing case count or missing person-time.

Keep units consistent. If your denominator is person-months, do not label the output as person-years unless you convert the denominator first.

Example

Suppose a study records 18 new cases over 9,000 person-years of follow-up.

IR = (18) / (9000) = 0.002

This equals 0.002 cases per person-year. To express the same result on more common scales:

0.002 × 1000 = 2
0.002 × 100000 = 200

The same incidence rate can therefore be reported as 2 cases per 1,000 person-years or 200 cases per 100,000 person-years.

How to Interpret the Result

  • A higher incidence rate means new cases are occurring more quickly in the observed population.
  • A lower incidence rate means new cases are occurring more slowly.
  • The rate describes the intensity of occurrence over time, not just the share of people affected by the end of the study.
  • Two studies can report the same number of cases but very different incidence rates if total person-time is different.

Incidence Rate vs. Other Common Measures

Measure Primary denominator Best use What it tells you
Incidence rate Person-time at risk When follow-up time differs across people How fast new cases occur
Cumulative incidence People initially at risk When you want risk over a fixed period The proportion that develops the outcome during that period
Prevalence Total population When measuring how common a condition is How many existing cases are present at a point or during a period

Common Input Mistakes

  • Using existing cases instead of new cases.
  • Using total population size as the denominator instead of total person-time.
  • Mixing time units without converting them first.
  • Continuing to count follow-up time after a participant is no longer at risk.
  • Interpreting a rate as if it were exactly the same as a probability or percentage.

FAQ

What does an incidence rate of 2 per 1,000 person-years mean?
It means that, on average, the study observed 2 new cases for every 1,000 person-years of time at risk.

Why is person-time better than just counting people?
Person-time accounts for unequal follow-up. If some participants are observed for a full year and others for only a few months, person-time keeps the denominator accurate.

Should case counts be whole numbers?
Usually yes. Case counts represent real events and are typically entered as whole numbers. Fractional values most often appear only when cases are reverse-calculated from a rounded rate.