Editing Incidence (epidemiology) (section)

===Example===
Consider the following example. Say you are looking at a sample population of 225 people, and want to determine the incidence rate of developing [[HIV]] over a 10-year period:
* At the beginning of the study (t=0) you find 25 cases of existing HIV. These people are not counted as they cannot develop HIV a second time.
* A follow-up at 5 years (t=5 years) finds 20 new cases of HIV.
* A second follow-up at the end of the study (t=10 years) finds 30 new cases.
If you were to measure prevalence you would simply take the total number of cases (25 + 20 + 30 = 75) and divide by your sample population (225). So prevalence would be 75/225 = 0.33 or 33% (by the end of the study). This tells you how widespread HIV is in your sample population, but little about the actual risk of developing HIV for any person over a coming year.

To measure incidence rate you must take into account how many years each person contributed to the study, and when they developed HIV because when a subject develops HIV he stops being at risk.  When it is not known exactly when a person develops the disease in question, epidemiologists frequently use the actuarial method, and assume it was developed at a half-way point between follow-ups.{{citation needed|date=March 2021}} In this calculation:
* At 5 yrs you found 20 new cases, so you assume they developed HIV at 2.5 years, thus contributing (20 * 2.5) = 50 person-years of disease-free life.
* At 10 years you found 30 new cases. These people did not have HIV at 5 years, but did at 10, so you assume they were infected at 7.5 years, thus contributing (30 * 7.5) = 225 person-years of disease-free life. That is a total of (225 + 50) = 275 person years so far.
* You also want to account for the 150 people who never had or developed HIV over the 10-year period, (150 * 10) contributing 1500 person-years of disease-free life.

That is a total of (1500 + 275) = 1775 person-years of life. Now take the 50 new cases of HIV, and divide by 1775 to get 0.028, or 28 cases of HIV per 1000 population, per year. In other words, if you were to follow 1000 people for one year, you would see 28 new cases of HIV. This is a much more accurate measure of risk than prevalence.