Last Update: October 15, 2007
This section provides updates regarding HIV epidemiology and the changing patterns and prevalence of HIV infections in different regions and countries.
All of my chapters were prepared mostly in 2005 and early 2006. I was able to update what I believe to be the most accurate estimates of HIV prevalence and HIV prevalence trends in late 2006, just before the book was sent off for printing.
In early July, NACO and UNAIDS revised the HIV prevalence in India for the 15+ population from close to 6 million to less than 2.5 million. Details of this drastic reduction and my assessment of this revised estimate are described below.
This table provides a summary of the estimated HIV prevalence and trends in major global regions since the new millennium. All of the regional estimates are based on UNAIDS reports. Estimates within the black brackets [xx] are the revised UNAIDS estimates – i.e., in the 2001 update report, UNAIDS estimated HIV prevalence in SSA at 9% and then revised its 2001 estimate downwards to 6.5% in the 2003 update report. Estimates within the red brackets [xx] are my “unofficial” estimates – i.e., I believed that HIV prevalence is grossly overestimated in SE Asia (at least a couple of million overestimated in India alone) and in Eastern Europe. UNAIDS in its December 2006 report estimated global HIV prevalence at 1% and in East Asia at 0.1%. UNAIDS was very generous in its “rounding off” of these estimates and a more accurate calculation results in lower estimates. The take home message is that even using UNAIDS’ own estimates and revisions, the AIDS pandemic is not “ever increasing and expanding.” The only region where estimated HIV prevalence seems to be increasing is Eastern Europe but my biased view is that perhaps HIV prevalence may be increasing in this region, but the estimates for this region are also grossly overestimated!
The Family Health Surveys (FHS) carried out in India in 2006 estimated HIV prevalence in the 15-49 year old population at 0.28%. However, NACO used HIV sentinel surveillance data to increase the estimate by about 25% to 0.36%. I believe that this "adjustment" is too large and needs to be justified by NACO, but no details of why this increase was warranted has been released by NACO or UNAIDS. In addition, NACO has applied the 0.36% prevalence estimate to the 15+ population and not just to the 15-49 year old population. All prior studies of the age distribution of adult HIV prevalence have shown that about 95% are within the 15-49 year old population. If the FHS finding of 0.28% in the 15-49 year old population is used, then the estimated number of HIV-infected persons in India would be about 1.6 million and if the increased estimate developed by NACO is used, the estimated number is about 2.1 million. If the new estimate is expanded to include the 50+ population, this would only add an additional 100,000 and thus the revised estimate for the 15+ population would at most be about 2.2 million.
Ethiopia was overestimated by almost 4-folds! These adjustments were based on more recent and more accurate HIV data and methods for estimating national HIV prevalence. Most of these downward revisions were forced by Demographic and Health Surveys (DHS) that are nationally-representative household surveys with large sample sizes. Most of the overestimation is due to the use of mostly urban sentinel ante-natal clinic samples where the highest HIV prevalence may be expected and then extrapolating these unrepresentative findings to the total rural and urban population. In these 15 countries the total overestimation is about 10 million. The majority of countries globally have not carried out DHS or DHS type studies. It can only be assumed that if some type of population based HIV surveys are carried out that HIV prevalence would also need to be lowered for most of these countries! It is of interest to note that no national HIV prevalence estimate has yet been found that has been underestimated!