Back to browse results
Estimation of HIV incidence in Malawi from cross-sectional population-based sero-prevalence data
Authors: Humphrey E Misiri, Abdi Edriss, Odd O Aalen and Fredrik A Dahl
Source: Journal of the International AIDS Society, 2012, 15:14 doi:10.1186/1758-2652-15-14
Topic(s): HIV/AIDS
Country: Africa
Published: MAR 2012
Abstract: Background Incidence is a better measure than prevalence for monitoring AIDS, but it is not often used because longitudinal HIV data from which incidence can be computed is scarce. Our objective was to estimate the force of infection and incidence of HIV in Malawi using cross-sectional HIV sero-prevalence data from the Malawi Demographic and Health Survey conducted in 2004. Methods We formulated a recurrence relation of population prevalence as a function of a piecewise-constant force of HIV infection. The relation adjusts for natural and HIV-induced mortality. The parameters of the recurrence relation were estimated using maximum likelihood, and confidence intervals of parameter estimates were constructed by bootstrapping. We assessed the fit of the model using the Pearson Chi-square goodness of fit test. We estimated population incidence from the force of infection by accounting for the prevalence, as the force of infection applies only to the HIV-negative part of the population. Results The estimated HIV population incidence per 100,000 person-years among men is 610 for the 15-24 year age range, 2700 for the 25-34 group and 1320 for 35-49 year olds. For females, the estimates are 2030 for 15-24 year olds, 1710 for 25-34 year olds and 1730 for 35-49 year olds. Conclusions Our method provides a simple way of simultaneously estimating the incidence rate of HIV and the age-specific population prevalence for single ages using population-based cross-sectional sero-prevalence data. The estimated incidence rates depend on the HIV and natural mortalities used in the estimation process.