The study, carried out by the Kenyan government, suggests 6.7% of people have the disease.The problem is that their survey methodology seems flawed.
Previous estimates had put the figure as high as 15% or 4.8m people.
Experts said the figures based on a sample of 8,561 households across the country are the most comprehensive to date.
This latest survey was carried out in September last year. As part of the survey, people were asked if they would be tested for HIV. Some 70% of those asked agreed.So the survey wasn't taken from a random sample. They may have asked a random sample to take HIV tests, but 30% of those asked refused. It seems very likely to me that a person's decision to agree or refuse to the test would be influenced by their own knowledge of their health and lifestyle. For instance, someone who never has sex or does drugs may refuse because they think there's no chance they have HIV and don't want to be bothered; someone who's very sick and suspects they may have HIV may refuse to take the test because they don't want to hear the bad news.
The tests were carried out by officials from the US Centers for Disease Prevention and Control. They found that 8.7% of women and 4.5% of men were HIV positive.
If either situation is very common then the results of this survey are meaningless -- and there's no way to know without testing the other 30% who refused. Therefore, although the numbers are encouraging, I'm skeptical that they really reflect the truth of the situation.