PIP: The Central Chest Clinic, Bangkok, Thailand, undertook a study among child contacts of newly discovered sputum-smear-positive patients with pulmonary tuberculosis to determine the effectiveness of BCG vaccination in young children. The study design resembled that of a controlled trial except that it was retrospective for vaccination, i.e., the vaccinated and control groups were not randomly selected. For this reason a number of measures were taken to allow for the comparability of groups to be verified and for adjustments to be made if necessary. The study was initiated in September 1981 and terminated in June 1984, after 971 index cases who reported contact with young children had been registered. Registration and initial examinations were completed for 1506 child contacts. The field teams could not trace 124 reported child contacts. In the case of 8 children, the initial examination was refused. Within a week of the detection of the index case, a visit was made to the household and personal data were collected. The contact children then were offered a clinical and X-ray examination at the Central Chest Clinic for the examinations. A clinical record was prepared for each child contact, the site of BCG vaccination was covered with a dressing (even if there had been no vaccination or scar), and the pediatrician administered a clinical examination and made a full-plate postero-anterior X-ray. The X-ray picture was examined by 2 readers. Suspect children were followed up for as long as there were medical indications. When indicated by the clinical or X-ray examination, a laryngeal swab was taken for culture and gland biopsies were made and examined by histopathology and culture. If tuberculosis was strongly suspected, treatment was initiated at once with rifampicin and isoniazid. For each child a final diagnosis was made at the end of the study. A scoring system proposed by WHO was used to obtain an indication of the probability of tuberculosis. As many as 1218 (81%) of the children had a BCG scar, and among those without a scar, 35 had a record of vaccination. Vaccination coverage as well as disease risk appeared to be associated with age. Stratification by age showed that this did not affect the calculated effectiveness of BCG vaccination. Apart from age, no differences between the vaccinated and the unvaccinated children were observed that call for stratification of the material. 284 tuberculosis suspects were found, 218 among the 1253 vaccinated and 66 among the 253 unvaccinated participants. The total incidence of tuberculosis was 14.5%; it was 12.6% among the vaccinated and 23.6% among the unvaccinated. Based on the data presented in Tables 1 and 8, and adjusting for the estimated 55 vaccinated children included among those without a scar or a vaccination record, the observed efficacy is 53% with 95% confidence limits of 64% and 38%; the observed number of cases among the vaccinated is 185 less than expected. Thus, although efficacy appears less, the effectiveness is far higher than with a more stringent diagnostic criteria.