A Verdecchia1, P Grossi, M Cantoni. 1. Laboratory of Epidemiology and Biostatistics, Istituto Superiore di Sanità, Rome, Italy.
Abstract
OBJECTIVES: The revision of the case definition for AIDS within a given observation period causes discontinuity points on the epidemic curve to which back-calculation procedures can be sensitive. The aim of this work is to characterize the impact of the 1993 European revision of AIDS case definition and to evaluate the degree of distortion by which back-calculation estimates obtained on Italian data are affected. METHODS: The back-calculation procedure used, is a generalization of standard methods for estimating and projecting the AIDS epidemic. Age at onset of HIV infection, incubation time distribution, susceptible population and competitive non-AIDS mortality rates for infected persons are included in the estimation process. RESULTS: Back-calculation estimates resulted to be sensitive to the discontinuity point that the 1993 revision of AIDS case definition induced on the epidemic curve. The resulting effect is a tendency to overestimate HIV incidence in recent years by 20-40%. The maximum effect was found using data with end of AIDS cases at December 1993. The overestimation is a transient effect, that is, it is expected to almost disappear in 1995-1996. CONCLUSIONS: The introduction of the 1993 European revision of AIDS case definition caused distortion of back-calculation estimates in Italy, leading to overestimate incidence and prevalence of HIV infection for recent years. It is believed that this effect has occurred in all countries that adopted the 1993 European revision of AIDS case definition, and its characterization is potentially useful for further changes of AIDS case definition in the future.
OBJECTIVES: The revision of the case definition for AIDS within a given observation period causes discontinuity points on the epidemic curve to which back-calculation procedures can be sensitive. The aim of this work is to characterize the impact of the 1993 European revision of AIDS case definition and to evaluate the degree of distortion by which back-calculation estimates obtained on Italian data are affected. METHODS: The back-calculation procedure used, is a generalization of standard methods for estimating and projecting the AIDS epidemic. Age at onset of HIV infection, incubation time distribution, susceptible population and competitive non-AIDS mortality rates for infected persons are included in the estimation process. RESULTS: Back-calculation estimates resulted to be sensitive to the discontinuity point that the 1993 revision of AIDS case definition induced on the epidemic curve. The resulting effect is a tendency to overestimate HIV incidence in recent years by 20-40%. The maximum effect was found using data with end of AIDS cases at December 1993. The overestimation is a transient effect, that is, it is expected to almost disappear in 1995-1996. CONCLUSIONS: The introduction of the 1993 European revision of AIDS case definition caused distortion of back-calculation estimates in Italy, leading to overestimate incidence and prevalence of HIV infection for recent years. It is believed that this effect has occurred in all countries that adopted the 1993 European revision of AIDS case definition, and its characterization is potentially useful for further changes of AIDS case definition in the future.