The follow-up window directly shapes the estimate because events are counted only during the stated period. Extending observation may capture additional new events, while a shorter window may produce a lower observed proportion. Therefore, comparisons are meaningful only when groups are evaluated over clearly defined and comparable time periods.
The denominator should represent individuals who could experience the specified event when follow-up begins. Including people who already experienced the outcome would distort the estimate because they cannot contribute a new event during observation. Clearly defining eligibility at baseline therefore supports a valid calculation and more interpretable comparisons.
Censoring or loss to follow-up means that some participants are not observed for the entire period, so their eventual outcomes may be unknown. A simple calculation can therefore misrepresent occurrence if follow-up is incomplete. Consistent outcome observation and careful interpretation are especially important when the amount of missing follow-up differs between groups.
First, define the population, specified event, and follow-up period. Next, count the new events that occur during that interval and identify the number of people at risk at the beginning. Divide the event count by that starting number, then report the result as a proportion or percentage with the observation period stated.
It is useful when groups have clearly defined populations, equivalent outcome criteria, and comparable follow-up periods. The resulting proportions can describe differences in disease occurrence or other specified events between groups. Interpretation should also consider incomplete follow-up and competing events, since these factors may affect the observed values.
An intervention study can compare the cumulative incidence of a specified outcome between treated and untreated or otherwise distinct groups during the same observation period. A difference in the resulting proportions may help assess intervention effects, provided the populations, event definitions, follow-up, and outcome observation are sufficiently consistent for comparison.