Fig. 2.

Comparison between the observed annual notification rates during the period 1967–2000 in Hong Kong among males, and model predictions of disease incidence based on the most pessimistic and optimistic assumptions about the annual risk of infection [i.e. that it declined by 2·5% or 13·5% p.a. respectively between 1950 and 1967 and by 13·5% p.a. thereafter (see Appendix)], and using values for the risks of developing disease estimated for the United Kingdom. For any given year, the predicted incidence is the same for all age groups, at least until the 1970s, since the risks of disease based on the UK notification data were identical for all adults, and the majority of adults in Hong Kong are predicted to have been infected at this time (see Fig. 1) and to be experiencing disease as a result of reactivation or re-infection. For the most pessimistic assumption about the annual risk of infection, the predicted rate of decline in the disease incidence changes in 1967 as a result of the increased rate of decline in the annual risk of infection from this year (i.e. 2·5% p.a. until 1967 and 13·5% p.a. thereafter).