The main types of prevalence used in epidemiology –– the study of diseases in populations – differ based on the window of time being assessed, and include:
- Point Prevalence: Measures the proportion of a population with a condition at a specific single point in time (e.g., a single day). It provides a “snapshot” of the disease burden.
- Lifetime Prevalence: The proportion of individuals who have ever experienced a condition or disease (i.e. the proportion of people who have had the illness of interest at any point in their lives up to the time of measurement).
- Often based on self-reported data, lifetime prevalence can be limited by recall bias and people forgetting. Additionally, researchers are unable to distinguish between active and past cases (i.e. whether every person who is counted as a ‘case’ still has the disease).
- Period Prevalence: Measures the proportion of a population with a condition at any point during a defined period, such as a month, a year, or several years (such as one year-prevalence).
- Period prevalence may not accurately capture the number of people who have a disease, especially if it is based on data which does not include individuals with mild, undiagnosed, or asymptomatic cases.
Estimates can also be grouped by data completeness i.e., what is directly visible in records versus what is statistically likely in the whole population:
- Observed Prevalence: Counts cases that can be directly identified in existing data, such as a registry or medical records.
- Observed prevalence is limited by the length of time the data source has been collecting records.
- Complete Prevalence: An estimate of the total number of people ever diagnosed with a disease who are still alive, including those diagnosed before data collection began.
- Due to gaps in data, complete prevalence is typically estimated using statistical models.
- Apparent prevalence: Based only on diagnosed cases.
- True Prevalence: Adjusts for undiagnosed or asymptomatic cases to provide a more accurate estimate of the actual occurrence.
- Pooled prevalence: Often, prevalence estimates from different studies are ‘pooled’ – combined using statistical methods – and presented alongside a range of values, known as confidence intervals, that provide an indication of the precision of an estimate. Here, narrower intervals indicate higher precision, and wider intervals indicate greater uncertainty due to sampling variation.
