Correct and consistent condom use can prevent susceptible people from acquiring HIV infection. However, in many countries repeated cross-sectional studies reveal a trend of increasing prevalence of HIV infection alongside an increase in reported condom use. Changes in sexual behaviour that reduce the number of new HIV infections will not become apparent through changes in HIV prevalence until some time after the behaviour change takes place. Limitations in the data used to assess condom use may also explain the concurrent increases in condom use and HIV prevalence. One common indicator of condom use, the UNGASS indicator (condom use at last higher risk sex of those aged 15-24), has been chosen to illustrate how changes in the proportion of people who report using condoms do not always explain changes in the size of the group who had high risk behaviour. Indicators based on the proportion of the whole population who have sex without using a condom would be better measures of the size of the group at highest risk of HIV infection.