Methodological issues in the estimation of the tuberculosis problem from tuberculin surveys.

SETTING: National tuberculin skin test surveys.
OBJECTIVES: To review the operating characteristics of the tuberculin skin test, to ascertain the validity of estimating prevalence and risk of infection from tuberculin skin test surveys under various conditions, and to review constraints in the estimation of the magnitude of the tuberculosis problem in the community from such surveys.
METHODS: This report utilizes hypothetical and selected real data obtained in regional and national surveys at various points in time to exemplify methodological issues.
RESULTS: Risk of infection, the essence to be abstracted from tuberculin skin test surveys, theoretically allows for a comparison of the extent of transmission of tubercle bacilli in various populations. However, the conduct of tuberculin skin test surveys and the analysis and interpretation of their results are not free from important technical problems. Accurate estimation of infection prevalence is particularly vulnerable to the great variability of the test's specificity under various circumstances. Furthermore, the annual risk of infection has averaging characteristics that preclude a rapid assessment of changes in transmission patterns. Finally, estimates of infection risk do not necessarily provide a standardized parameter to derive incidence of infectious cases, because of variations in the quality of intervention and varying risks of progression from latent infection to overt tuberculosis.
CONCLUSIONS: While tuberculin skin test surveys provide the currently most widely used means of assessing tuberculosis transmission patterns over prolonged periods of time in a community, results from such surveys must be interpreted with caution when accurate estimates of the tuberculosis problem are sought.