PAPP103 - S09: Models of populations with variable growth rates

Application 1: Expected durations from data on prevalence

One important application of this generalization of stable population relationships is to the construction of period life tables from two successive cross-sectional enumerations of a population. This method of estimation can be applied in many circumstances, the most obvious of which is to estimate mortality in national populations. More generally, it can be used to study any survival process in which experiencing the event of interest removes members of the population from that population.

Here, the variable growth method for estimating period life tables is explained with reference to estimating the expected duration of episodes of disease from two sets of data on prevalent cases of the disease. The method is then illustrated by using it to estimate the expected length of time that individuals remain members of an institutional population.

A common short-cut method for calculating the average duration of episodes of a disease (D) is the prevalence-incidence mean. This is calculated by dividing the number of prevalent cases of a disease (P) by the incidence of new cases in a defined reference period (I). It seems intuitively obvious that if, for example, there are half as many prevalent cases of a disease as people who develop the disease each year, people must be living with the disease for 6 months on average before they recover or die.