UNAIDS Report to Show Lower HIV/AIDS Estimates Because of Changes in Data Collection, Estimation Methods

HIV/AIDS estimates in a new UNAIDS report scheduled to be released before next month's XV International AIDS Conference in Bangkok, Thailand, are expected to be lower than those reported in 2002 primarily because of changes in estimates for sub-Saharan Africa, according to a report by officials from UNAIDS, UNICEF and the Department of Infectious Disease Epidemiology at Imperial College published in the June 26 issue of the Lancet. The UNAIDS report, which will include HIV prevalence estimates for 2003, will show "notably lower" estimates for Cameroon, Cote d'Ivoire, Ethiopia, Kenya, Rwanda, Zambia and Zimbabwe and will show "notably higher" estimates for Senegal, accoridng to Neff Walker, senior adviser on statistics, modeling and projections at UNAIDS, and colleagues. However, these changes reflect "improvements in the validity of the estimates" rather than actual fluctuations in the prevalence of the disease, the report says. For countries with generalized epidemics -- where HIV prevalence among pregnant women exceeds 1% -- overall HIV prevalence is calculated from data on women tested through prenatal care centers. The data then is adjusted and updated based on studies of HIV prevalence in specific populations. Countries with low-level or concentrated epidemics -- in which transmission is assumed to occur mainly among men who have sex with men and sex workers and their clients -- estimates are made for each group and then combined to obtain an estimate of the total number of HIV-positive people. However, many countries have started large-scale household surveys, which can offer more precise estimates than prenatal testing data. In addition, many countries have generated more precise data after attending workshops run by UNAIDS, UNICEF, the World Health Organization and CDC. Therefore, the changes in the 2003 data largely reflect "improvements in the validity of the estimates," according to the report.

Plausibility, Point Estimates
In addition, the new report will replace "point estimates," which give a precise figure, with "plausibility bounds," which give a range of figures for the best-case scenario, worst-case scenario and best estimate of HIV/AIDS prevalence. Although the point-estimate system always included information on the relative certainty of the data, the figures "encouraged a false perception of precision," the report says. Therefore, UNAIDS and WHO have decided to use the plausibility bounds to "reflect the uncertainty" around the estimates. The bounds are determined by estimating the size of major sources of error, possible assumptions in the translation of data, the prevalence levels and the scale of the epidemic. The report concludes, "Although changes in estimates do raise some concerns about the validity of previous estimates, the willingness to change and update estimates as new methods and data become available also shows that a self-correcting process is in place for making estimates of HIV/AIDS" (Walker et al., Lancet, 6/26).

Back to other news for June 30, 2004