Measurement of the incidence of tuberculous infection.

For certain purposes it is desirable to know the incidence of tuberculous infection in various population groups with a high degree of accuracy. The most accurate method for obtaining such information is the follow-up method, in which the population is tuberculintested on two separate occasions in order to determine the proportion who become infected with tubercle bacilli during the interval. The method is illustrated by means of data collected on children under 9 years old in Tunisia. The data indicate that the use of the same limit between "positive" and "negative" reactions in the two tests would lead to biased estimates of infection rates. A method of analysis involving the use of different limits has therefore been developed and applied to the data. The accuracy of the incidence rates obtained is indicated by a comparison of hypothetical prevalence rates computed from them with the observed prevalence rates.