Incidence (epidemiology) - Advanced Applications and Related Measures

Incidence and Prevalence: Fundamental Measures of Disease Frequency Introduction Epidemiologists use two primary measures to describe how commonly diseases occur in populations: incidence and prevalence. While these terms are sometimes confused, they answer fundamentally different questions about disease. Incidence tells us about the risk of developing a disease, whereas prevalence tells us about the burden of disease at a specific point in time. Understanding the distinction between these measures—and how they relate to one another—is essential for interpreting epidemiologic data and understanding disease patterns in populations. Conceptual Difference: Incidence vs. Prevalence Incidence measures the number of new cases of disease that develop in a population at risk during a specified time period. It focuses on the transition from health to disease and thus provides information about risk—the probability that a susceptible person will develop the disease. Prevalence, by contrast, measures the total number of cases (both new and existing) of disease in a population at a given point in time. It captures how common the disease is in the population at that moment, regardless of when people became ill. Consider a concrete example: if you're studying diabetes in a city, incidence tells you how many people are newly diagnosed with diabetes in a given year (perhaps 150 new cases), while prevalence tells you how many people in the city have diabetes at a particular moment in time (perhaps 8,500 people). The prevalence includes those newly diagnosed plus everyone who already had diabetes from previous years. Why Incidence Is Better for Understanding Disease Causes A critical insight in epidemiology is that incidence is more useful for understanding disease etiology (the causes of disease). Here's why: when incidence of a disease changes, it suggests that something in the environment or population has changed—likely the presence or absence of risk factors that trigger the onset of disease. For example, if the incidence of lung cancer increases sharply in a region over five years, this suggests that something has changed to increase people's risk of developing lung cancer (perhaps increased smoking rates or air pollution). Prevalence, on the other hand, could remain stable or even increase for other reasons that have nothing to do with new causes—such as better treatments that keep patients alive longer. Thus, changes in incidence point more directly to etiologic factors. How Disease Duration Shapes Prevalence and Incidence One of the most important concepts for understanding the relationship between these measures is recognizing that disease duration dramatically affects prevalence while having little direct effect on incidence. Consider two scenarios: Scenario 1: Long-duration disease. Suppose a disease has high incidence but people live with it for many years before recovering or dying. Over time, new cases accumulate in the population, building up a large pool of existing cases. Even if incidence begins to decline, prevalence remains high because all those existing cases persist. Scenario 2: Short-duration disease. Now suppose a disease has high incidence but resolves quickly (either through recovery or death). Cases arise rapidly but disappear rapidly. In this case, prevalence can remain relatively low even though incidence is high, because the disease doesn't persist long enough for cases to accumulate. A practical example: COVID-19 early in the pandemic had high incidence (many new cases daily) but relatively lower prevalence because most people recovered within weeks. In contrast, hypertension (high blood pressure) has lower incidence (fewer people develop it each year) but very high prevalence because people live with it for decades. The Mathematical Relationship: Prevalence and Duration When conditions are stable—specifically, when incidence remains roughly constant over time and the average duration of disease stays the same—we can express the relationship between prevalence and incidence mathematically: $$\text{Prevalence} \approx \text{Incidence} \times \text{Average Disease Duration}$$ This elegant formula captures why understanding disease duration is crucial. If you know the incidence and average duration of a disease, you can approximate how common it is in the population at any given time. Important caveat: This formula works well for the overall population, but becomes considerably more complex when examining age-specific prevalence and incidence. Age-specific relationships are influenced by additional factors such as aging, mortality patterns, and birth cohorts, which make the simple multiplication rule inadequate for these subgroups. Calculating and Interpreting Incidence Rates: A Practical Example To illustrate how incidence is calculated and interpreted, consider a hypothetical study of HIV transmission in a population. Suppose researchers follow a cohort of 2,000 individuals who are initially HIV-negative. At the start of the study, 25 individuals are found to already have HIV and are excluded from the risk set because they cannot acquire the infection again (you cannot acquire a disease you already have). This leaves 1,975 susceptible individuals. Over the follow-up period, researchers track these 1,975 individuals and observe 50 new cases of HIV. However, not all individuals are followed for the entire duration—people drop out, move away, or the study ends. The total person-time accumulated across all individuals is 1,775 person-years. The incidence rate is calculated as: $$\text{Incidence Rate} = \frac{50 \text{ new cases}}{1{,}775 \text{ person-years}} = 0.028 \text{ cases per person-year}$$ Or expressed as a more intuitive rate: 28 new cases per 1,000 person-years (or 28 per 1,000 per year). Interpreting this result: This incidence rate tells us that among susceptible individuals, approximately 28 out of every 1,000 would acquire HIV each year. This is a much more precise measure of risk than prevalence alone, because it accounts for the actual time people spent at risk during the study and excludes those who were already infected. Related Epidemiologic Measures Based on Incidence Several other important epidemiologic measures are derived from or closely related to incidence: Attack Rate Attack rate is a specific type of incidence proportion used during disease outbreaks. It measures the proportion of a defined, exposed population that becomes ill during a short, specified time period. Attack rates are commonly used in outbreak investigations—for instance, determining what percentage of people who ate at a restaurant developed food poisoning. Unlike standard incidence rates, attack rates are typically expressed as proportions (percentages) rather than person-time rates, and they apply to a clearly defined, often small population exposed to the outbreak source over a brief interval. Attributable Risk Attributable risk (also called excess risk) quantifies the difference in incidence between an exposed group and an unexposed group. Specifically: $$\text{Attributable Risk} = \text{Incidence in exposed group} - \text{Incidence in unexposed group}$$ This measure indicates how much of the disease in the exposed group can be attributed to the exposure itself. For example, if smokers have a lung cancer incidence of 150 per 100,000 per year and non-smokers have an incidence of 10 per 100,000 per year, the attributable risk is 140 per 100,000 per year—meaning smoking accounts for 140 of the 150 cases in smokers. Rate Ratio Rate ratio compares the incidence rates of two groups by expressing one as a multiple of the other: $$\text{Rate Ratio} = \frac{\text{Incidence rate in exposed group}}{\text{Incidence rate in unexposed group}}$$ Using the smoking example above, the rate ratio would be 150/10 = 15, meaning smokers have 15 times the incidence of lung cancer compared to non-smokers. Rate ratios greater than 1 indicate increased risk with exposure, while ratios less than 1 indicate decreased risk (protective effect).