Trends in childhood tuberculosis in Hungary 1953-1983: quantitative methods for evaluation of BCG policy.

To analyse the effectiveness of the BCG policy on childhood tuberculosis in Hungary, the following three quantitative methods have been organized and systematically applied since 1959; regression analysis with multiple comparison of the incidence trends of age groups; stepwise regression analysis to select the strongest childhood incidence decrease factor; and a log-linear model to compare the risk of disease in vaccinated and non-vaccinated children. The increasing BCG coverage of children as a specific antituberculous measure (i) has decreased the childhood incidence (23-32% per year) 3-4 times more rapidly than the adult incidence (6-16% per year) as shown by the regression analysis, (ii) has acted as the strongest incidence decrease effect among other factors and (iii) has protected the vaccinated children 2.8 times better than the non-vaccinated ones. The use of valid and exact models, from statistical inferences ensure the objective interpretation of the epidemiological events in tuberculosis.