Epi Corner: rates versus counts

Epidemiology Corner

January 14, 2005

Rates and counts are often used by epidemiologists to describe disease in a population. Though both rates and counts can be fairly simple to calculate, it is important to understand how they are used, how they differ and how to interpret them.

Counts, also called frequencies, are fairly straightforward. Counts are the total number of events that occur in a defined period of time. The total number of cases of Salmonellosis reported in Houston during 2000 would be an example of a count or frequency, e.g. in 2000, 227 cases of Salmonellosis were reported.

Rates are the number of events that occur in a defined period of time, divided by the average population at risk of that event. In order to estimate the rate of Salmonellosis in Houston during 2000, divide the number of reports of Salmonellosis in 2000 by the population of Houston in 2000. This is roughly equivalent to calculating the percent of people with Salmonellosis in Houston in 2000, e.g. cases in 2000 Population in 2000 = 227 1,953,631 = 0.00012.

This number is hard to interpret since the population of Houston is so large in relation to the number of reported cases of Salmonellosis. By multiplying the number by 100,000, the rate of Salmonellosis in Houston is 12 cases per 100,000 people, e.g. (Cases in 2000 Population in 2000) x 100,000 people = (227 1,953,631) x 100,000 = 12.

In other words, in 2000, for every 100,000 people in Houston, 12 developed Salmonellosis. This rate calculation is only considered an estimate since many people may have had Salmonellosis but were not sick enough to see a doctor.

Rates are often used instead of counts because they allow comparison of the level of disease or another health event in two different populations. For example, to compare the number of Salmonellosis cases reported in Houston with the number reported in another city, use rates and not counts. If two different cities both had 277 cases of Salmonellosis a year it may appear they have the same level of problem, but this may not be the case. See the figure below. This table compares two fictitious cities, one called Uoyba City and one called Notsouh. In scenario A both cities have 277 cases of Salmonellosis a year. In Scenario B, Uoyba City has over twice as many cases.

So which city has a greater problem of Salmonellosis? In scenario A, both cities have the same number of cases. But Notsouh has a higher rate of disease. That is, Notsouh has a rate of 15 cases per 10,000 people while Uoyba City has a rate of only 7 cases per 10,000 people. Notsouh may have a larger problem because it has more disease per person. In scenario B, Uoyba City has many more cases than Notsouh does. However, both cities have the same rate, 48 cases per 10,000 people. In scenario B, the cities have the same level of disease per person. In general, expect higher numbers of cases in a city with a larger population. By looking at the rates instead of just counts, one can compare which city has a greater level of disease per person.