Objective: Epidemiological surveillance of tuberculosis (TB) in Australia is dependent on the accuracy of the notification data. We have investigated the specificity of TB notification for the diagnosis of this disease. Method: We used data from notifications to the NSW Department of Health to identify members of a cohort of refugees who were screened between 1984 and 1993 and subsequently developed TB during that period. We reviewed notification data and, in most instances, case notes and x-ray films, to independently confirm or refute the diagnosis of TB. Results: Two hundred and fifty members of the cohort were identified in the notification database. After excluding refugees on treatment prior to arrival in Australia, and those who were notified as 'quiescent' and 'atypical' cases, there were 189 cases notified as active TB. There was evidence to support the diagnosis of active TB in 125 cases (66%) and evidence that subjects did not have active disease in 60 cases (32%). We could not determine the status of the remaining four notified cases. Conclusion: This study has shown that, in a population of refugees subject to screening, nearly one-third of cases notified as active tuberculosis from the study population were actually not active tuberculosis cases. Implications: The use of the TB notification database may result in overestimation of the incidence of TB in population groups who are subject to active screening.

[No Keywords available]

data base tuberculosis control tuberculosis cohort analysis normal human medical record accuracy refugee Australia Article screening X ray film human