Incidence and Prevalence in Epidemiology: Their similarities, differences, and relationship

Imagine that you are have recently started working as an epidemiologist in your community, and one morning, the local health authorities call upon you to give you an important task. You are to estimate the burden of diabetes in your community.

You, armed with your knowledge in medical research, epidemiological methods, and clinical acumen, promptly get on towards fulfilling the task at hand. You examine the available records of the only hospital in that community as well as that of the locally conducted household health surveys and find that 2,000 residents out of a population of 100,000 are currently living with the disease. Your colleague, another up-and-coming epidemiologist like you, meanwhile, reports that 300 residents of the community were newly diagnosed with diabetes during the preceding year.

This now creates a dilemma. Which number correctly describes the actual burden of the disease? Is it the yearly rate of new diabetes diagnoses, or is it the total number of patients actually living with it? Both numbers that you and your colleague got accurately describe diabetes in the same community. But they tell fundamentally different stories.

The first number, that of 2000 people with diabetes out of 100,000 residents describes how widespread diabetes is. The second one, which is 300 new cases of diabetes logged by the hospital in the last year tells us how many new cases have emerged over a period of time in that community.

These two perspectives form the basis of two of the most frequently encountered concepts in epidemiology: prevalence and incidence.

Although the terms are sometimes used interchangeably in everyday conversation, they answer different scientific questions. And understanding these differences is crucial when interpreting published studies, designing research projects, planning healthcare services, or evaluating public health interventions.

What Is Incidence?

The term incidence, whenever used in the context of medical science, refers to the measurement of the occurrence of new cases of a disease or health event in a defined population over a specified period of time.

Let us consider the following example:

Suppose a researcher follows 500 adults who are free from diabetes at the beginning of a study. Over the following 12 months, 30 participants develop diabetes. Assuming all 500 participants complete follow-up and remain eligible for the outcome assessment, the one-year cumulative incidence is:

This means that 6% of the initially diabetes-free participants developed diabetes during the year. Notice that the denominator consists of people who were initially at risk of developing the disease. Individuals who already had diabetes at baseline would not be included in the population at risk for a first diagnosis.

Another thing to notice here is that what we have calculated here is not simply incidence, but cumulative incidence. This is because incidence is commonly expressed in two ways, cumulative incidence, which is also known as incidence proportion, and incidence rate, which is also known as person-time incidence rate.

1. Cumulative incidence (incidence proportion)

This measures the proportion of initially disease-free individuals who develop the disease over a specified period. It is particularly useful when a clearly defined group can be followed over time. For example, a researcher may estimate the proportion of patients who develop a surgical site infection within 30 days of surgery.

Cumulative incidence can be interpreted as a risk over the stated period, provided that follow-up and competing events are appropriately handled.

2. Incidence rate (person-time incidence rate)

In many epidemiological studies, not every participant is observed for the same length of time. Some participants may enter the study late, withdraw before its completion, die, or develop the outcome during follow-up. Consequently, simply dividing the number of new cases by the original sample size may not adequately account for differences in the duration of observation. The incidence rate, also known as the person-time incidence rate or incidence density, addresses this limitation by considering the total amount of time participants remain under observation and at risk of developing the disease.

Person-time represents the combined duration of time that all participants contribute to a study while they remain at risk of developing the outcome. It is commonly expressed in person-years, person-months, or person-days, depending on the duration of follow-up.

For example, one person followed for five years contributes five person-years. Similarly, five people followed for one year each also contribute five person-years. Thus, person-time reflects the total accumulated observation time rather than simply the number of individuals enrolled.

Consider a hypothetical study involving five participants:

Although five participants were enrolled, they collectively contributed only 19 person-years of observation rather than 25 person-years. This is because some participants stopped contributing time before the study ended. An important principle is that each participant contributes person-time only while remaining at risk of the outcome. For a study examining the first occurrence of a disease, a participant who develops the disease stops contributing person-time at the time of diagnosis. Similarly, participants who withdraw, die, or reach the end of follow-up stop contributing time at those respective points.

Calculating the Incidence Rate

The incidence rate is calculated using the following formula:

Suppose a cohort contributes 900 person-years of observation, during which 18 participants develop the disease.

Expressed per 1,000 person-years:

This means that the observed frequency of new disease occurrence was 20 cases for every 1,000 person-years of accumulated time at risk. It does not mean that exactly 20 out of every 1,000 individuals developed the disease, because person-years measure accumulated observation time rather than the number of distinct individuals.

Incidence Proportion Versus Incidence Rate

The distinction between incidence proportion and incidence rate is important when interpreting longitudinal studies.

• Incidence proportion (cumulative incidence) measures the proportion of initially disease-free individuals who develop a disease during a specified period. It is interpreted as the risk of developing the disease over that period, assuming appropriate follow-up and accounting for competing events where necessary.
• Incidence rate measures how frequently new cases occur relative to the accumulated person-time at risk. It is particularly useful when participants have unequal follow-up durations or when individuals enter and leave observation at different times.

For example, an incidence proportion of 10% over five years indicates that 10% of the initially at-risk population developed the disease during that period. In contrast, an incidence rate of 20 cases per 1,000 person-years describes the frequency of new cases in relation to the total observation time contributed by participants.

What Is Prevalence?

Prevalence measures the proportion of a population that has a particular disease or condition at a specified point or during a defined period.

Unlike incidence, prevalence includes both recently developed and previously existing cases.

Suppose a community survey examines 1,000 adults and identifies 160 individuals currently living with hypertension.

The prevalence is:

In other words, 16% of the surveyed population had hypertension at the time of assessment.

We cannot determine how many of these individuals developed hypertension during the previous month or year. Some may have been diagnosed recently, while others may have been living with the condition for decades. Therefore, prevalence tells us how common the condition is, not when the affected individuals developed it.

Like incidence, prevalence can also be described in two forms:

• Point prevalence measures the proportion of people affected at a particular point in time. For example, the proportion of nursing-home residents experiencing pressure injuries on the day of a clinical assessment represents point prevalence.
• Period prevalence measures the proportion of people who had the condition at any time during a specified period, including those whose illness began before the period and continued into it. For example, a survey might estimate the proportion of adults who experienced migraine during the preceding 12 months.

Both are measures of existing disease occurrence, but they use different time frames. A study reporting prevalence should therefore make its reference period explicit.

Incidence and Prevalence: A Side-by-Side Comparison

Why High Prevalence Does Not Necessarily Mean High Incidence and the relationship between the two

Consider two hypothetical conditions.

1. Disease A develops infrequently, but affected individuals commonly live with it for several decades.
2. Disease B develops relatively frequently, but most affected individuals recover within a few days.

Which disease would have higher prevalence? Not necessarily Disease B.

Despite the greater frequency of new cases, its short duration may mean that relatively few individuals have the condition at any particular moment. Conversely, Disease A may accumulate a large number of existing cases because affected individuals remain in the population for many years. Under simplified assumptions of a stable population, a steady disease process, and relatively rare disease, the relationship is often expressed as:

This is why, when an effective treatment for a disease is found, that can render an acute severe (often lethal) disease to a chronic, but not so severe disease, sometimes prevalence of the disase is increased in the population due to the prolonging of the overall survival, without necessarily increasing incidence.

For example, improvements in long-term management of a chronic disease may allow more people to live with the condition. As a result, its prevalence may rise even if the rate of new diagnoses remains unchanged.

A rising prevalence figure should therefore never be interpreted automatically as evidence that more people are developing the disease.

How Incidence and Prevalence Guide Healthcare Decisions

Epidemiological measures are not merely calculations performed for research examinations. They provide different kinds of information for healthcare planning. These might be:

1. Understanding healthcare requirements

Prevalence helps estimate how many people may need ongoing treatment, follow-up, medication, rehabilitation, or supportive services. A district with a high prevalence of hypertension may require expanded primary care services, reliable antihypertensive drug supplies, and long-term monitoring facilities. Even if relatively few new cases emerge each year, the existing patient population may continue to place substantial demands on the healthcare system.

2. Evaluating preventive interventions

Incidence is particularly relevant when the objective is to prevent new cases. Suppose a public health programme introduces an intervention intended to reduce the occurrence of dengue. Researchers may compare dengue incidence before and after implementation, while considering seasonal variation, population changes, surveillance quality, and other possible explanations. A reduction in newly occurring cases may indicate improvement, although determining whether the intervention caused that reduction requires an appropriate study design.

3. Understanding disease patterns

Monitoring incidence over time can help identify changes in disease occurrence, detect outbreaks, and generate hypotheses about environmental, biological, or behavioural risk factors. Prevalence surveys, meanwhile, can identify populations in which existing disease burden is particularly substantial. Both types of information are valuable, but they support different interpretations.

Choosing the Correct Measure for Your Research

Imagine three postgraduate medical students planning different studies.

1. The first wants to assess how common anaemia is among pregnant women attending antenatal clinics.
2. The second wants to determine how frequently patients develop postoperative surgical site infections during a 30-day follow-up period.
3. The third wants to examine the occurrence of new tuberculosis diagnoses in a defined population over several years.

Which epidemiological measures would be appropriate?

1. For the first study, prevalence is a natural choice because the investigator wants to estimate how many participants have anaemia at the time of assessment.
2. For the second study, cumulative incidence would be appropriate if the investigator follows a clearly defined group of initially infection-free patients for the specified postoperative period, accounting properly for incomplete follow-up.
3. For the third study, an incidence rate may be useful, particularly if the population or duration of observation varies. Its denominator and case definitions must be specified carefully.

Five Common Mistakes Researchers Should Avoid

Mistake 1: Calling every percentage prevalence

A percentage does not automatically represent prevalence. If 15 out of 100 initially disease-free patients develop a complication during follow-up, the resulting 15% describes cumulative incidence, not prevalence.

The interpretation depends on what was counted and when.

Mistake 2: Including existing cases in the incidence numerator

Incidence concerns newly occurring cases. Counting participants who already had the disease at baseline would distort an estimate of first disease occurrence.

Mistake 3: Ignoring the observation period

An incidence of 5% over one month cannot be interpreted in the same way as an incidence of 5% over five years.

Time is an essential part of incidence measurement.

Mistake 4: Confusing hospital frequency with population prevalence

If 40% of patients attending a specialist clinic have diabetes, that figure does not establish a 40% prevalence in the surrounding community.

The clinic population may be highly selected.

Valid population estimates require clearly defined populations, appropriate sampling, and attention to selection bias.

Mistake 5: Comparing estimates without examining their definitions

Two prevalence estimates may differ because investigators used different diagnostic criteria, age groups, sampling strategies, or assessment periods.

Similarly, incidence comparisons require attention to follow-up duration, case ascertainment, and population characteristics.

An apparent difference in disease occurrence may reflect methodological differences rather than a genuine epidemiological change.

References and Further Reading

1. Centers for Disease Control and Prevention. Principles of Epidemiology in Public Health Practice. Third Edition. Lesson 3, Section 2: Morbidity Frequency Measures. Read here.
2. Centers for Disease Control and Prevention. Principles of Epidemiology in Public Health Practice. Lesson 3, Section 1: Frequency Measures. Read here.
3. National Cancer Institute. Surveillance, Epidemiology, and End Results Program. Cancer Incidence Rates. Read here.

Note: All numerical examples in this article are hypothetical and are provided for educational purposes.