Recently, we conducted series of community HCT outreach in different parts of the country, reaching nearly 1.5 million men and women. I would like to know if the sero-positivity rate from this outreach have any bearing with the actual prevalence, considering all the limitations of such outreach programs. If yes, what is the relationship and how can one make a valid case for taking the sero-positivity data seriously?
Thanks
The HIV seropositive rate in any subset of a population, regardless of how the subset is chosen and sampled, will correlate somewhat with the overall HIV seropositive rate in the entire population. But exactly how closely the rates are correlated with each other depends on dozens of factors. A few of the factors are related to how "mature" the HIV epidemic is, in the region or country that is being discussed. I a country or region where HIV has jest begun to spread in recent years, the epidemic can be very localized to a small risk group or one sub-population such as IV Drug uses who share needles in the city where the virus recently entered. In a more mature epidemic where the virus has been infecting people for more than 30 years, most of the people at highest risk of infection are already infected, most people are aware of the problem and reducing their risk behaviors, and the virus is spread more evenly in all towns and cities in the region.
Other factors are related to how the HCT outreach sampling is done and how likely it is to reflect the general population of the region being sampled. In many regions of the world injection drug use is illegal and primarily confined to people who will avoid sampling by many types of outreach workers, and thus they might be "under sampled". In other cases, the HCT outreach may specifically be sampling higher risk individuals, such as for example pregnant women who obviously have had unprotected sex. In this case there may be "over sampling" for people at higher risk.
Many viruses, such as influenza, Rhinoviruses, enteroviruses and herpesviruses are relatively uniformly distributed in human populations. HIV and quite a few other viruses are very highly concentrated in very small sub-populations of people. Everybody is at risk of being exposed to influenza virus if it come into town, but most people have zero risk of being infected with HIV even if other people living in the same village are infected.
In addition to the contributions from Brian, I have observed that sero-prevalence rate is lower in community outreaches especially places where screening has been done previously. It is possible that most people coming for screening during community outreaches have had a previous test with a negative result. However, prevalence in targeted groups at high risk usually give higher prevalence than the entire population.
HIV seropositivity rate in non-probability samples (like in your outreach programmes) is difficult to correlate to country's HIV prevalence, If you had used HIV ELISA or other tests, that test has its own sensitivity and specificity. In really speaking, true HIV prevalence should be based on a country representative probability sample and the seropositive cases should be confirmed using HIV confirmatory algorithm to say the real HIV prevalence (otherwise it is also an HIV seroprevalence). I think some kind of a modelling technique considering all the variables can give an answer to your question. Statisticians can help on that and it is out of my expertise. I just shared my thoughts. Thanks for the question. Best regards
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Invertebrate Zoology
- Entomology