Data Call Into Question Hiv Study Results

[wasabi]

Study

Unfortunately this was a likely outcome, HIV defies statistical analysis given that not everyone who contacts it will get it and not everyone who was in either group was certain to have come in contact with it. Disappointing but not surprising. It seems some therapy based on blood from Elite Suppressors is more likely to succeed. That is those with a natural immunity.

Edited October 10, 2009 by wasabi

[phetaroi]

...HIV defies statistical analysis given that not everyone who contacts it will get it...

I'm not aware of any disease where the above is not true.

[silverhawk_usa]

I also wondered how they compiled statistics on a test such as this. As part of the testing process they provided Hiv education to the recipients. In other words, how NOT to come in contact with the disease. So, if they are not coming into contact how do you know the vaccine is working? How do you have any idea what percentage of those taking the vaccine actually were exposed to the virus vs. those not taking the vaccine?

[Sheryl]

The above are not the problem. All study participants were sexually active people and assignment to vaccine and placebo group was random, so that exposure risk would be the same. Of course not everyone was exposed; one assumes majority were bnot exposed, that's a given and why such a large sample was required.

What is a problem is that the results initially reported, which were statistically significant (p<.05) failed to exclude people who did not get the full vaccine course. Which was premaure and shoddy work IMO. Assuming the numbers given in this article are true -- and they probably are as it is a very reputable publication -- the results do not in fact meet the criteria for statistical significance.

There is still an 84% chance that the vaccine reduced transmission, which might justify further work on the vaccine, but the generally accepted scientific standard of proof is 95%.

Disappointing.