Re: India HIV data: Joint Communique: MoHFW, WHO and UNAIDS

June 6, 2005

Dear Mr. Hota and Statistical Team,

Thank you for filling us in on the methodology. I know it’s a very sensitive and

complicated issue.

I am little confused about response number 4 when you say

" The total of new infections is more than the difference between the estimated

prevalence of 2003 and 2004 as it should also include AIDS related deaths.

Hence, the number of actual new infections is 28,000 plus the number of deaths

due to AIDS " .

As I understand incidence is number of new cases occurring in a specified time

period, usually one year in a defined population. Incidence of new cases and

deaths are two different measures and as such cannot be combined to calculate an

incidence rate. Addition of number of people died to number of new cases would

overestimate the incidence, as those who died are already counted as cases, many

might be infected in previous years and some in current year and all dying in

current year.

The HIV data and for that matter any disease data in India comes mainly from

government sector with near total lack of reporting from private sector. This is

going to lead to under reporting of HIV cases. The use of sentinel surveillance

only gives an estimate and depends upon which site are chosen as sentinel

surveillance sites. If the sites are not representative of population catered by

both public and private sectors, estimates are going to be biased.

So in my opinion data is underreported unless statistical team can give a better

explanation.

Surveillance is an extremely important task and very expensive too. Being a huge

country makes it difficult to get complete reporting. The exact number of cases

or a reasonable estimate is required for budget allocation, planning of services

and program evaluation as we are all aware.

Regards,

E. Naik, MD, PHD

Asst. Professor of Epidemiology and Biostatics

USF, FL 33612, USA

E-mail: <enaik@...>

Recommended Posts