Tuberculosis: a private and public health and data mix

5,6 They also suggest that HTLV-1 incidence will continue to fall in Japan as the higher risk cohorts age and die. Only a few other studies have attempted to determine the incidence of HTLV-1 and even fewer have used representative samples of general populations. Two studies in the USA and one in Brazil estimated incidence in repetitive blood donors and found results that are probably lower than, but of similar magnitude as, general population rates. Glynn and colleagues 7 and Zou and colleagues 8 in the USA reported incidence rates of 1·6 new infections per 100 000 person-years and 0·21 new infections per 100 000 person-years, respectively. The diff erence between them is probably due to diff ering geographic areas served by the sampled blood centres. In Brazil, Carneiro-Proietti and colleagues 9 reported a rate of 3·6 per 100 000 person-years, quite similar to Satake’s estimate for Japan and about three times higher than in the USA, consistent with known HTLV-1 epidemiology. Three other studies found much higher incidence, probably because they studied higher-risk populations, including a hyperendemic island in Okinawa, Japan (1·0 per 1000 person-years), 10 a sexually transmitted infection clinic