News and Background

The big Asia public health news of the moment is the reporting of the world’s first successful HIV/AIDS vaccine trial. In the trial, started in 2006, of the 16, 395 Thai volunteers that participated half were put on a placebo, and the other half were given the vaccine, RV 144, a mix of two previously failed vaccines, Alvac-HIV and AIDSVAX.  The vaccine was designed to inoculate against HIV subtypes B/E, which are those found in Thailand.

Findings

Of the 8,198 who received a placebo, new infections occurred in 74.  Of the 8,198 who received the vaccine, new infections occurred in 51 people. The difference of 23 people is taken as the measure of effectiveness for the vaccine, or 31%.

Significance

Many public health experts, even ones with close links to the project, have been cautious about getting too excited about these results, particularly experts in countries where HIV/AIDS has become a pandemic, like in India.  Some have called it a base hit as opposed to a home run, and everyone is saying that just because this trial may have had some success, a ‘cure’ is still many years away.   In fact, experts do not even know how the vaccine mechanism works, and teams are being set up around the world to analyze data from the study in an effort to shed some light on this key question.

Controversy

There is also controversy about whether or not the data are actually statistically significant, and nothing has been released about how the trial was actually designed to work, or what the controls for the trial were.  The Lay Scientist points out that the p value of the trial data is .048, which translates into a 4.8% probability that the results of the trial are accidental.  The standard for p value set by serious scientific populations is .05, and some publications require it to be even lower.

What the Lay Scientist does not point out is that p-value is an easily customizable statistic that can be calculated from various data distributions and results can vary significantly depending on which type of relationship is analyzed.  In effect this means that the p-value can either be lower or higher or, alternatively, the p-value could be completely worthless if the right statistical tests were not used.

For me, however, the more pressing question is what steps the researchers took to assure that the make up of the two groups studied – the placebo group and the vaccine group – wasn’t responsible for the perceived effects.  The official report is that all of the people taken were volunteers from populations not thought to be at high risk of contracting AIDS/HIV; meaning no prostitutes or drug addicts.  But, people change and even assuming that the researchers did a good job of profiling all 16,000 trial participants before the start of the trials, two years in a poor country is more than enough time for someone to become a member of one of the high risk groups.  If, for example, this happened more frequently in the placebo group than the vaccine group, resulting in more chances for individuals in that group to contract HIV/AIDS than the results might not be representative of anything more than a, for lack of a better word, fluke.

My question for readers is how researchers take steps to tackle such foreseeable problems in the make up of trial participants?

I also would like to know the potential that this vaccine has to be harmful.  This is information that will come with time, but readers and observers should not forget that vaccines are not magic bullets and can sometimes create as many problems as they solve.  (Don’t take this as me being an anti-vaccine looney, I’m not, my point is simply that it is worthe remembering in the case of this trial, since it could derail the success of the vaccine.