Numbers of tuberculosis cases: dreams and reality.

Abstract

Use of estimated numbers of tuberculosis (TB) cases for planning purposes in some sub-Saharan countries. To document the uncertainties of official World Health Organization estimates and problems encountered in using them for planning. Brief review of the methods used in estimation, using different sub-Saharan countries to illustrate problems. The annual risk of tuberculous infection, used for many years to calculate estimates, is no longer considered a valid method. New methods are based on an assessment of the completeness of TB notification data (the Onion Model) and prevalence surveys of bacteriologically proven pulmonary TB cases; however, these are subject to bias and are very imprecise. Examples from sub-Saharan countries reflect these difficulties and show that official estimates vary substantially, by up to a quarter of the initial values. Donors, particularly the Global Fund, rely on these estimates and push countries to arbitrarily increase planned numbers of notified cases to improve 'case detection rates'. Use of estimated numbers to monitor progress in TB control may be counterproductive, costly and risky. It would be much more realistic to accept that low-income countries plan their strategies based on TB notifications rather than on case detection rates that are more dream than reality.