Time to abolish "gold standard"

the risk of an increase in tuberculosis related to HIV is small.2An audit of cases of tuberculosis (April 1990-March 1991) in Riverside Health Authority showed two problems with the current surveillance of HIV infection and tuberculosis based on notifi- cation data.Firstly, tuberculosis in HIV infected people is less likely to be notified than that in other people.In our study only four (40%) of 10 dual infections were notified; in contrast, 50 (74%) of 68 cases of tuberculosis in people not known to be infected with HIV were notified.Furthermore, HIV positive patients given antituberculous treat- ment without bacteriological evidence being sought were even less likely to be notified (one (18%) of six).Secondly, a rise in the number of cases of tuberculosis attributable to HIV infection may be masked by the high rates of, and changing trends in, notifications concerning people from the Indian subcontinent (India, Pakistan, Bangladesh)34 owing to the similar age ranges of the two groups of patients (mean ages in our study were 41 and 42 years).Ethnic group and country of origin are not recorded on notifications.Cooperation between the consultant in com- municable disease, microbiologists, and clinicians has corrected the problem of differential reporting from Riverside, and we recommend that other health authorities audit their cases of tuberculosis to ensure that the surveillance data are unbiased.Moreover, to measure the extent of dual infection we have begun locally to record ethnic group and country of origin for all cases of tuberculosis and recommend that other centres with a high preva- lence of HIV infection do likewise.