Setting: Forty-two laboratories in four countries.
Objective: To determine the number of sputum smear examinations required to identify one additional case of tuberculosis from a third serial diagnostic smear or one additional treatment failure from a second serial follow-up smear.
Material and methods: Country-specific prevalence of new cases and failures among 59 665 examinees were determined, as well as the incremental yield from serial smears. The reciprocal value of the product of the prevalence of cases or failures and the respective incremental yield from the last serial smear provided the number of slides that have to be examined to identify one additional case or failure.
Results: The expected prevalence of cases among suspects ranged from 5.4% to 32.8%; the incremental yield from a third serial smear ranged from 0.7% to 7.2%. Between 122.7 and 796.3 smears were required to identify one additional case with the third serial smear. The prevalence of failures among follow-up examinees ranged from 1.0% to 2.5%; the incremental yield from the second follow-up serial smear ranged from 4.5% to 26.9%. Between 164.8 and 2133.4 slides were required to identify one additional failure with the second serial smear.
Conclusion: The utility of serial smears can be rationally determined by careful review of program data.