[aruanan]

Most patients are progressors because they are patients. Non-progressors don't walk in sick. It is not at all that simple.

Think of it. You're arguing from ignorance. Bad move. Though we may not know the total numbers either of those who are progressors or those who are not (for the very simple reason that one cannot have omniscient knowledge of all members of any extremely large and constantly changing category, though one can come to have pretty solid statistical knowledge), of those who are tested, found to be HIV positive, and who are tracked, the majority is found to consist of progressors. This may not be exhaustive knowledge, but it's far more solid than categorical games such as Most patients are progressors because they are patients. Non-progressors don't walk in sick. It is not all that simple.

[JasonC]

A known selective effect distorting a sample is not "ignorance". And no mythical exhaustive knowledge of a category is needed to minimize the distortion produced by this selection effect (which, in case anyone is having trouble tracking, is people with symptoms already being selected as progressors). You would use a cohort track of a random sample, e.g. from some regular institutional screening procedure, say, including negatives and a-symptomatic positives. You would get a measure of prevalence of each category in the general population as a result. Two or three such cohort studies in different places, with different reasons for the screening, would improve confidence in the results. It is not a question of mining bad data not directed at the question asked, nor merely speculating about the question asked in the absence of good data. It is a question of going out and getting good data.