Monday, September 28, 2009

US activist group urges caution on vaccine trial results

by Keith Alcorn, via Aidsmap

The results of the RV 144 HIV vaccine trial should be treated with caution until further data are presented, US advocacy organisation Treatment Action Group said on Friday, following the announcement that the combination of vaccines used in the study resulted in a 31.2% reduction in the risk of HIV infection.

“Based on the limited amount of information that has been released, it appears that the statistical significance [of the result] hangs on very few cases of HIV infection. TAG urges caution in interpreting the findings until more detailed information is available,“ the group stated in a press release.

Seventy-four out of 8198 volunteers in the placebo group became infected, compared with 51 out of 8197 volunteers who received a combination of two vaccines, ALVAC vCP1521 and AIDSVAX B/E. Participants in both the vaccine group and the placebo group received counselling on how to reduce their risk of HIV infection.

Although the difference in the risk of infection was statistically significant, the confidence intervals for the estimate were extremely wide.

Read the rest.

1 comment:

rjreinhard said...

I think many others, including the trial sponsors themselves, the NIH, AVAC and others have urged caution about the results, justifiably so. The mechanism issues TAG and others have set out are good reasons for doing that. Still,although the statistical significance is low, the range(~1%-~50%) is entirely in the positive spectrum, there are no chances the vaccines did not have some effect. In other words, the randomization and design procedures which are there to account for confounding effects still show some beneficial effect attributable to immunization even if the effect is a very small one. The question remains whether everyone agrees the randomization, trial size and design were sufficiently robust so that even this "blip" effect is real. That is why human trials have so much interest. They can show something is happening - if they are designed well- and then leave for later how to figure out why it worked. If the numbers of infected individuals remain as they are from both arms of the study, it would be hard to ignore that something worth investigating further has occurred. We are all aware of why it should not have worked, it's not a bad effort to spend some time thinking about why it almost very likely did work

Related Posts Plugin for WordPress, Blogger...