Incidence and prevalence

Definitions & Key takeaways

Both incidence and prevalence are ways to measure the number of people who have a certain disease in a population. More specifically, incidence of a disease refers to the number of new cases that occur in a population over a specific period of time; whereas prevalence refers to the number of existing cases that are found in a population at any given point in time.

Watch video only
Prevalence and incidence are two ways to measure the number of people who have a certain disease in a population. Prevalence includes all individuals with a disease, or all cases, regardless of when they developed it, divided by the total population.
Whereas incidence includes the new cases of a disease - meaning anyone who develops a disease, over some period of time, divided by the population of individuals that can get that disease.
And both of these have to take into account the population at risk. Sometimes incidence and prevalence are reported as a proportion of a population like 5 per 1000, or 5 per 100, 000 if it’s a rare disease.
Let’s use a quick analogy to show the relationship between incidence and prevalence. Imagine a bucket of green marbles - with each marble representing a sick person.
New green marbles keep falling into the bucket - these are the new cases. We can count up how many green marbles enter the bucket in a year, and that would be the incidence.
Now some marbles also fall out of the bucket - perhaps some cases recover from the disease while other cases die. Either way, they don’t have the disease anymore and are no longer cases in our bucket.
We can count up the number of green marbles in the bucket and that would be our prevalence. So the prevalence really depends on the incidence of new cases, the recovery rate of the cases, and the death rate of the cases.
So for prevalence we have all cases of a disease over the population at risk. If it’s done at some specific moment, like the day that a survey is being done, then it’s called point prevalence.
And occasionally, if it’s done over a period of time, like over a month or year, then it’s called period prevalence. Now when we have new cases of a disease in a certain period of time, it’s called incidence rate.
This can’t be done at a specific moment, since we need some time window to capture the new cases, and sometimes it’s hard to follow up with everyone for the entire study period.
When this happens, we can use something called person time to find the incidence rate. That’s when you sum up the time observed for every person during the study period.
Let’s look at an example of incidence rate using person-time. Let’s say we want to figure out the incidence rate of pneumonia in a small retirement home that has 10 people, and that we do a 3 year study.
Right at the end of year 1, there’s an outbreak and 5 people get pneumonia, and sadly 2 of them died as a result. At the beginning of year 2, a new person moves into the retirement home.
Right at the end of year 2, one person gets pneumonia and this person moves back home to live with their family so we don’t have any more data on them.
At the beginning of year 3, another new person moves into the retirement home, and this person catches pneumonia, but recovers.
To calculate the incidence rate we need to add up the actual time we observed individuals, and we’ll do it in “years”. So, we have 7 people who were in the study for all 3 years, 2 people who died and were only in the study for 1 year, 1 person who was in the study for 2 years then moved away, 1 new person who was in the study for 2 years, and another new person who was in the study for 1 year.
So our person time in years is (3 + 3 + 3 + 3 + 3 + 3 + 3 + 2 + 1 + 1 + 2 + 1) = 28 person-years. If we look at the incidence, there were a total of 7 new cases during this time period.
So 7 new cases of pneumonia over 28 person years observed is 7 / 28 = 0.25 new cases per person-year, meaning in this retirement home, there’s a 25% chance every year of getting pneumonia.
And since there were 2 deaths, there’s a 2 / 28 or 7% chance every year of dying in this retirement home. Okay - let’s do another example, this time to calculate prevalence.
Let’s say there’s a town of 1000 people. The town used to have a coal-mine, but it’s been closed for many years because the coal mining company went out of business.
Now, let’s say there’s a research study on pneumoconiosis, a chronic lung disease caused by inhaling coal dust. It turns out that there aren’t any new cases of pneumoconiosis since the coal mine shut down, but that there are about 200 retired miners who have the disease.
The incidence rate is new cases/number of people at risk per year. In this case, there are no new cases, so that’s a 0 in the numerator, and the number of people at risk would be 1000 - 200 or 800 in the denominator.
That’s because the 200 individuals that already have pneumoconiosis, which is a chronic disease, can’t get it again. So the incidence is 0 / 800 which is 0% per year.
The prevalence is all cases divided by the total population, so that’s 200 divided by 1000 which is 0.2 or 20%. From this example we can see that even when the incidence is zero, the prevalence can still be 20%, which is very high.
For a chronic illness, like pneumoconiosis, that people usually have for many years, there might be a handful of new cases in a community, but because the cases stick around for a long time, the point prevalence will be high regardless of when you check.
Now on the flip side, let’s say that there’s a 24 hour stomach virus going around during the summer months, there might be a high incidence because many people might develop the illness in a community that year, but a low prevalence of individuals that have the illness if you do a point prevalence some time in December - because by then, everyone will have likely recovered.
Alright, as a quick recap. Incidence is the number of new cases of a disease in a specific period of time, divided by the susceptible population - for example, those that don’t have that disease already.
It might be calculated as the new cases in a year or it can be calculated by person years if it’s not possible to follow up on everyone.
Prevalence is all cases of a disease in a specific moment, typically at the time of survey, called point prevalence, although it can also use a specific period of time, called period prevalence.