Wednesday, May 17, 2006


AP: 240 People in SF Tested HIV Poz in '05 (New Lower Infection Rate?)


Okay, all you folks out there who know math and do your best to follow San Francisco's ever-fluctuating rate for new HIV infection, I need your help.

Back in March of this year the city health department lowered its estimate of infection, as duly reported first in the Bay Area Reporter, then in the Chronicle.

This is what the BAR said:

Now the city is estimating the number of new HIV cases will drop to 976, with gay men accounting for 772 of those cases each year.

And the Chronicle went with the same number for projected new HIV cases:

The new estimate is 976 -- 87 percent of them gay men.

So, for all intents and purposes, HIV epidemiologists here believe their best guess for new infections is near the one-thousand mark, right?

Well, an Associated Press story out tonight claims the actual number new positives is much, much lower than what the BAR and Chronicle reported just two months ago:
Last year, 240 people tested positive out of the 6,000 tested in San Francisco.

How is it the same health department can make such wildly differing claims about the HIV rate, in year 25 of the AIDS epidemic, in the one city in America that is really supposed to have its act together about EVERYTHING to do with HIV, including accurate epidemiology?

Reading the numbers, which, of course, come from the same source, SF DPH, I think we can safely say it's not reporters getting the figures wrong -- it's the health officials here doing their Keystone Kops routine in presenting HIV estimates.

By the way, looking at the stats from the AP wire, 240 out of 6,000 comes to 4 percent.

And this is my question for you: Are the 240 people tested figure and 4 percent rate a further drop in the HIV infection rate than earlier predicted by SF DPH?

1 comment:

Anonymous said...

Haven't seen the articles in context, or looked closely at the figures myself recently. But the 976 figure could be the estimated case rate including *untested* individuals, extrapolating from the 240/6000 number (assuming that the 6000 tested were high-risk, and that the 4% rate wouldn't apply to the entire population).