Incidence

Incidence is the occurrence of new cases of a disease, injury, or other health event in a defined population during a specified period. In epidemiology, the term may refer to a count of new cases or to a measure relating those cases to the population at risk. Its two principal forms are incidence proportion, which measures risk over an interval, and incidence rate, which measures how rapidly new cases occur. (archive.cdc.gov)

Incidence proportion

Cumulative incidence, also called incidence proportion, is the fraction of an initially disease-free population that develops the condition during a stated interval:

It is a probability, ranging from 0 to 1, and is commonly expressed as a percentage. The observation period is essential: a one-year risk and a ten-year risk describe different quantities. During an outbreak of an infectious disease, an incidence proportion is often called an attack rate, although it is a proportion rather than a person-time rate. (archive.cdc.gov)

For a hypothetical group of 1,000 initially disease-free people, all followed for one year, 20 new cases give a one-year incidence proportion of $20/1{,}000=2\%$.

Incidence rate and person-time

Incidence rate, also called incidence density or a person-time rate, relates new cases to the total time during which people were observed and at risk:

Person-time combines the number of people observed with their individual observation durations. It can be expressed in person-days, person-months, or person-years. Unlike incidence proportion, incidence rate has a time dimension and is not itself a probability. (archive.cdc.gov)

For example, 20 new cases observed during 950 person-years give an incidence rate of approximately 21.1 cases per 1,000 person-years. This describes the frequency of events relative to accumulated observation time, not a 21.1% annual probability.

Population rates and standardization

Population reporting often expresses incidence as new cases per 100,000 people per year. In cancer surveillance, the numerator may count newly diagnosed primary cancers rather than unique patients: one person can contribute more than one primary cancer. Recurrences are generally excluded. The denominator must match the population eligible for the condition being counted. (seer.cancer.gov)

A crude rate describes occurrence in the population as observed. Because age distributions differ between populations, comparisons often use age-standardization. An age-standardized incidence rate is a weighted average of age-specific rates, using the age distribution of a standard population:

where $I_a$ is the incidence rate in age group $a$, and $w_a$ is that group's standard-population weight. Comparisons require the same standard population; adjustment reduces the potential confounding effect of age. (seer.cancer.gov)

Difference from prevalence

Prevalence measures existing cases at a specified time or during a specified period, including cases that began earlier. Incidence concerns newly occurring cases. Thus, prevalence describes how widespread a condition is, whereas incidence describes its occurrence over time. (archive.cdc.gov)

Prevalence depends partly on both incidence and the length of time people remain affected. A high prevalence can therefore reflect frequent new disease, prolonged survival without cure, or both; it does not necessarily imply high incidence. A mortality rate, by contrast, measures deaths rather than new diagnoses. (archive.cdc.gov)

Measurement and interpretation

Incidence is used in public health surveillance to describe disease occurrence and changes over time. Reliable measurement requires a defined population, observation interval, and procedure for identifying new cases. Births, deaths, and migration can change the population at risk, while uncertain dates of onset can make it difficult to distinguish new cases from pre-existing disease. (cdc.gov)

Recorded incidence also depends on detection. The date of diagnosis may not coincide with biological onset. In cancer statistics, a rise in diagnosed incidence can reflect a genuine increase in disease occurrence, but it can also follow expanded screening or overdiagnosis—the detection of cancers that would not have caused problems during a person's lifetime. Incidence trends therefore cannot, by themselves, establish whether underlying disease risk or health outcomes have changed. (cancer.gov)

References

1. Principles of Epidemiology | Lesson 3 - Section 2archive.cdc.gov
2. Principles of Epidemiology | Lesson 3 - Section 1archive.cdc.gov
3. Principles of Epidemiology: Glossaryarchive.cdc.gov
4. Cancer Incidence Rates - SEER Cancer Statisticsseer.cancer.gov
5. Principles of Epidemiology in Public Health Practice, Third Edition: An Introductionarchive.cdc.gov
6. Incidence - Health, United Statescdc.gov
7. Cancer Statistics - NCIcancer.gov