Incidence Density

Incidence density, also called the incidence rate, is the number of new cases of a condition divided by the total person-time at risk accumulated by the population under observation. By using person-time rather than a count of people as its denominator, it expresses how rapidly new cases arise per unit of time and accommodates individuals who are observed for differing lengths of time.

Definition

Incidence density is the incidence rate computed as the number of new cases divided by the total person-time at risk accumulated by the population, expressed in units of cases per person-time (for example, per 1,000 person-years).

Scope

This entry covers incidence density as the rate form of incidence: the concept of person-time, the units of inverse time it carries, how it handles varying and incomplete follow-up, and how it differs from the cumulative incidence proportion. It is methodological and does not provide clinical guidance.

Key concepts

• Person-time at risk
• New cases (numerator)
• Rate with units of inverse time
• Varying follow-up duration
• Handling of loss to follow-up
• Distinction from cumulative incidence

Mechanisms

Each person at risk contributes time to the denominator only while they remain under observation and free of the condition; their contribution ends when they develop the outcome, are lost to follow-up, or reach the end of the study. Summing these contributions gives the total person-time at risk, and dividing the count of new cases by it yields a rate that measures the instantaneous speed at which new cases occur per unit of time. Because the denominator is time, not a fixed group of people, incidence density naturally accommodates subjects observed for unequal periods and those who enter or leave the at-risk pool, which is what makes it the appropriate incidence measure for open or dynamic populations and for studies with substantial loss to follow-up. Unlike a proportion, the rate is not bounded by 1 and carries units of inverse time.

Clinical relevance

Incidence density is the frequency measure behind rate ratios and rate differences and is the form of incidence used when follow-up time varies across individuals, so it is central to appraising time-to-event evidence. It describes population-level speed of occurrence and is not a basis for individual diagnostic or treatment decisions.

Epidemiology

Incidence density is the standard incidence measure in cohort studies and registries where people are followed for different durations, and in dynamic populations whose membership changes over time. Its reliance on accurately accruing person-time, including correctly stopping each person's contribution at their outcome or censoring time, links it to survival-analytic methods for time-to-event data.

History

The use of person-time denominators to express incidence as a true rate was formalised as twentieth-century epidemiology distinguished the rate form of incidence from the proportion form. Standard textbooks and dictionaries codified incidence density and its person-time denominator, and its connection to the hazard concept of survival analysis was developed in the methodological literature.

Debates

When is a person-time rate preferable to a risk proportion?

Person-time rates handle unequal and incomplete follow-up that a simple proportion cannot, but they assume the rate is reasonably constant over the contributed time; when that assumption is doubtful, summarising heterogeneous follow-up into a single rate can obscure how risk changes over time.

Seminal works

• grimes-cohort-2002
• rothman-2008
• porta-2014

Frequently asked questions

What is person-time?

Person-time is the sum of the time each person in a study is observed while at risk of the outcome. One person followed for two years and two people followed for one year each contribute two person-years; it lets a study combine individuals observed for different lengths of time.

Why can an incidence density exceed 1?

Because it is a rate per unit of person-time rather than a proportion of people, incidence density is not bounded by 1. A value such as 1,200 cases per 1,000 person-years simply reflects the chosen time unit and the speed at which cases occur.