Introduction

In the conduct of injury research, we often wish to measure the frequency of injury occurence, of events that might produce injuries, or of injured persons. Counts alone are inadequate for many purposes, as they include no information about the size of the population or period of time from which the counts arose. Rates, which can account for population size and time intervals, are the topic of this chapter.

Basic Concepts

Counts

Two types of counts may be defined. Incident cases are newly injured persons counted over a period of time. Prevalent cases are persons with an injury-induced condition counted at a single point in time. In injury research, incident events are sometimes counted, rather than persons; for example, traffic crashes in a year, or falls in a nursing home in one month. Note that a crash event may include several people, and a single person might fall several times.

Counts, without a denominator, are often inadequate for research purposes. Imagine we are told that there were 100 deaths due to fires in community A and 1000 deaths due to fires in community B. Without information about the size of the populations in A and B, and the time periods over which the deaths were counted in each region, we are unable to compare the hazards for death by fire in the two areas. A denominator can provide the needed information.

What is a Rate?

A rate is a count divided by a denominator. Some have expressed the view that the term rate should be reserved for incidence rates, which have incident counts in the numerator and person-time in the denominator (Rothman, 1986, pp. 23–34).