Examination of only one sputum sample cannot detect all the sputum positive cases in the community. To obtain better estimates of the prevalence of bacteriologically confirmed disease in the community, a study was conducted to find out the additional yield of cases through collection and examination of eight sputum specimens and also in order to work out correction factors for estimates based on one or two sputum samples, as collecting multiple sputa is very difficult. The study was carried out in 77 villages in Nelamangala Taluk of Bangalore. In all, 5826 persons were referred for sputum examinations.

Results of all the eight culture examinations were available for 2973 (51% of the eligibles). Of these 64 persons were positive by culture of atleast one specimen. Each of the eight specimens has the chance of detecting a case and any one of them could be considered as first or second specimen etc. To overcome this difficulty 80 permutations were randomly chosen out of the total 40,320 permutations possible. Cases from first specimen and additional cases from subsequent specimens were calculated through four mathematical equations. The first equation namely Y = KXm (28.66 x-1.40) has been considered as providing the best fit to the observed data. On the basis of this equation it appears that additional positives could be obtained upto the 1Oth specimen. Out of 64 culture positive cases, only 72% of positives could be detected by first two samples. To get about 95% of the cases, it is necessary to examine at least six specimens from each individual. Multiple samples are rewarding for detecting even high grade cultures.