AIDS Vaccine Is Partially Successful 317
ifchairscouldtalk writes "A Phase III 'RV 144' study in Thailand succeeded in reducing HIV infection rate in trial with 31.2% effectiveness. The study was conducted by the Thailand Ministry of Public Health and used strains of HIV common in Thailand. It is not clear whether the vaccine, which combines AIDSVAX with Aventis Pasteur ALVAC-HIV canarypox vector, known as 'vCP1521,' would work against other strains in the United States, Africa or elsewhere. Strangely, the vaccine had no effect on levels of HIV in the blood of those who did become infected, providing 'one of the most important and intriguing findings' of the trial, according to Dr Fauci, director of the National Institute of Allergy and Infectious Diseases, which is one of the trial's sponsors."
Scary clinical trial (Score:1, Insightful)
Col. Jerome H. Kim, a physician who is manager of the armyâ(TM)s H.I.V. vaccine program, said half the 16,402 volunteers were given six doses of two vaccines in 2006 and half were given placebos. They then got regular tests for the AIDS virus for three years. Of those who got placebos, 74 became infected, while only 51 of those who got the vaccines did.
Wait, wait, wait. Did they go through the random distribution of people who may get aids? Clearly they did not just infect people with aids afterwards. The only way I can think of them getting these rates is that when someone tested positive for aids who was not in the trial noted their previous partners, then that partner list was cross-referenced with the vaccinated.
Hm, I guess the other way would be if someone who had the vaccine was told by one of their partners that they were infected and that they should get themselves checked.
Regardless, there has to be a significant margin of error on their estimates thanks to AIDS reporting and such.
Inspiring.... (Score:4, Insightful)
How in the hell could you ever do a controlled experiment like this on people if you dont control their exposure to the infection causing material? The only way you can determind improvements of real thing over placebo is if you intentionally expose the test subjects to the virus...which would be a death sentence.
Their results could mean that the group recieving the test vaccine came into contact with the virus 31.2% less.
Effectiveness (Score:4, Insightful)
Furthermore, the low effectiveness is actually a liability; the end result could be mutations in the HIV virus that make it immune to the vaccine. This is part of the reason why the influenza vaccine has limited effectiveness - influenza, like HIV, has a tendency to mutate quickly. If a new strain comes along, like H1N1 for influenza, you're defenseless.
Finally, I think there's a problem with how the vaccine will be perceived. If the vaccine is only 30% effective, I think people will see that as being too risky to even get the shot. There's already (too much IMO) FUD out there against vaccines in general. If you think that you can get influenza from the flu vaccine, there's a strong aversion to taking the HIV vaccine. For a 30% chance at being immune, that's no good. If it were 100%, that would be a totally different story.
Re:Inspiring.... (Score:5, Insightful)
Re:Lulz (Score:3, Insightful)
Why is it that on slashdot of all places that should be full of nerds we get idiots that don't grasp basic statistics and people that mod it up? As long as you got a proper control group it's simple to say "If we assume the true probability is the same, how unlikely is it that we get these results?" Of course there's something about the level of confidence - a 99% confidence means there's a 1% your observation is random fluctuations. But the whole "we reject math and logic because the numbers feel to small" sounds like the results of retarded anti-schooling.
Re:Inspiring.... (Score:4, Insightful)
here is how:
1)get the infection rate of the population
2)take a random sample from the population
3)do a double blind study of the vaccine
4)at the end of x years, compare the rate of infection of both your experimental group and your control group. If the control group is with in the statistical bounds of the population infection rate and the experimental group's infection rate is below that rate at a statisticaly significant level, then you can conclude the vaccine has a positive impact on infection rates.
Re:Vaccine Is Partially Successful (Score:4, Insightful)
Not at all. What this shows is that the vaccine likely works for some subset of the population. That doesn't mean it doesn't work at all. Viagra for example only works for about 60% of men but people don't go screaming that it doesn't work.
Bottom line here is that vaccine or no, you should still practice safe sex (afterall, HIV isn't the only bad disease lurking around out there). However, if this thing has a ~30% chance of making you immune to the disease with no other ill effects then it's certainly worth reducing your chances by that much.
Basically, to break it down, your chances of getting aids comes down to 3 factors (4 now with this in place):
a * b * c * d
Where
a = the chance that your partner is infected
b = the chance that you catch the disease during an encounter with an infected partner (having intercourse with an infected person doesn't guarantee infection)
c = the chance that your protection fails (only comes into play if you used protection - otherwise it's 100%)
d = the chance that your vaccine was ineffective (only comes into play if you actually got vaccinated - otherwise this is 100%)
Everything that is scientifically proven to reduce the final result, even if it doesn't go to 0% in the end, is a success in my opinion.
Re:Lulz (Score:3, Insightful)
It would be unethical to expose all participants to HIV. They did the next best thing.
There's nothing wrong with the basic idea of the study design. Of course, they may have fucked it up, but that's a different situation.
Stats 101 (Score:3, Insightful)
No, you don't need to control their exposure. You can study the infection rate for the general population, and provided that your study group isn't unusually different from the general population (say, by being all sexually active gay men), you can expect a similar infection rate over time.
yes, there are potentially statistical deviations that could occur, but the larger the sample group and the more test that occur, the less likely this is. Go take some stats classes if you are curious about the methodology, but if they did the trial correctly your suggested interpretation is very unlikely.
Re:I don't know how you can buy these results... (Score:1, Insightful)
I'll never understand what makes the general public think they are qualified to critique scientific studies.
Re:Inspiring.... (Score:1, Insightful)
The reason is.... randomness! Assuming that the groups were selected randomly then the results hold. You may point out, "Well, isn't it possible that all the people who were put in the vaccine group just didn't participate in risky activities?" The answer is of course it is possible, but if the groups were assigned randomly the chance of that happening is built into the result.
As long as the group assignments were truly random and they didn't treat the people in the two groups differently after they were assigned then I would expect that you would see similar results if this experiment were repeated.
Re:Effectiveness (Score:4, Insightful)
I think the big deal here is that they were able to create something that has an effectiveness greater than 0.
I'm not an expert on the subject, but I guess it's easier to go up from 30% effective than from 0% effective.
Re:New Vaccine prevents Card Accidents! (Score:1, Insightful)
Worst. Car. Analogy. Ever! You probably think a standard deviation is, "guys who dig chicks with hairy armpits."
Take 16,000 drivers at random. Send half of them through intensive accident avoidance training. Give the other half a fifteen minute lecture on traffic safety and a "Safe Driver" sticker. Let commuting take its course and tote up the body work needed, with only portion of the whole 16,000 subjects actually getting into a fender bender. Now count the number of people in both populations involved in those accidents, seeing if there's a significant statistical difference in numbers between those that got the training and those that didn't.
What you're proposing would mean is sending all 16,000 drivers into a surprise demolition derby.
Re:Thumbs up (Score:3, Insightful)
The highest-rising infected group is heterosexual women in their 20s.
Look it up.
Re:No hurry (Score:3, Insightful)
Right, because herpes and HPV didn't exist until 1980 :P
Re:Lulz (Score:4, Insightful)
that's all well and good, but if you don't understand statistics, you probably shouldn't be complaining about the statistics in a study that is undergoing peer review.
I'm not saying you're complaining about the study, I just don't think the excuse you presented holds water.
Re:HIV Vaccine (Score:2, Insightful)
yes, the main contributing factor to monogamy today is the fear of aids.
Result is NOT statistically significant (Score:2, Insightful)
According to numerous online sources, raw numbers are:
51 out of 8187 found infected in the vaccinated group;
74 out of 8198 found infected in the control group.
The most basic course of statistics tells how to proceed from here: test if the null hypothesis (vaccine has no effect) remains plausible despite this evidence. Conditions are ideal for the chi-squared test.
We get Observed values 51, 74, 8147, 8123; Expected values 62.504, 62.496, 8135.5, 8134.5; then sum((O-E)^2/E) = 4.267, with two degrees of freedom.
Conclusion: the null hypohesis is rejected with only 88% confidence level.
This is not enough to confortably say that the vaccine has any benefit. Odds of the contrary are about 1/17.
This is much less reason to trust that the vaccine reduce infection rate by 31.1%, as reported in some press articles. Odds are 1/2 that it is less efficient than this.
Francois Grieu
Re:No hurry (Score:4, Insightful)
That is no more bigotry than it is to say that heterosexual sex carries a higher risk of pregnancy.