Correcting reported AIDS incidence: a statistical approach.

This study is motivated by the time lag between date of diagnosis of AIDS cases and date of reporting, which results in incomplete data about the epidemic. A maximum likelihood procedure has been developed to adjust the actual numbers of diagnosed AIDS cases for reporting delay. If a parametric function for describing past and future incidence is assumed, its parameters and the adjustment for reporting delay can be estimated simultaneously. Data from the WHO Collaborating Centre on AIDS, Paris, are used. Practical problems related to data collection are dealt with.