How DPH Cooked HIV Infection Numbers
By Terry Beswick
Get out your abacus, and follow along, if you can. There's some fancy math going on here.
"If you times the percentage of new infections by the size of the population, you learn how many people are newly infected in any year," Health Director Dr. Mitch Katz posited earlier this month. "The numbers in our presentation are very close to the numbers that were leaked earlier on, it was that we needed to do all of our homework."
Introduced by a concerned Mayor Willie Brown at an August 9 news conference, Katz was seeking to explain how his Department of Public Health, along with the University of California, San Francisco, had jointly arrived at, and leaked to the media, a new estimate that between 750 and 900 people will become infected with HIV this year in San Francisco–a minimum 50 percent increase from the 500 annual infections the two institutions had estimated in their last concerted effort three years ago.
Described by DPH epidemiologist Dr. Willi McFarland in press reports as a "surge" in HIV infections of "sub-Saharan African" levels, the shocking increase galvanized the creation of a new "HIV prevention plan," released earlier this month by Katz and UCSF AIDS Research Institute Director Dr. Tom Coates.
However, upon closer analysis, at least half of Katz's equation on which the much-ballyhooed new estimate and plan are based appears to be highly vulnerable to critical scrutiny.
"Some of it's all so slight of hand. I don't find it to be precise," commented Board of Supervisors President Tom Ammiano, who held a hearing on the matter after the mayor's press conference on August 9. "It's always going to be ambiguous."
"Follow the trend, that's what matters," pleaded Katz to reporters. "This has never been about whether San Francisco has 736 new infections or 719 new infections; that's not what matters. What matters is that infections are up.
"Condom use is down, everything else is up," Katz said.
And indeed, few experts in the DPH-UCSF-dominated world of HIV prevention in San Francisco have openly questioned the estimates, shared on August 8 with officials from the federal Centers for Disease Control and Prevention.
Yet the forecast of annual HIV infections in San Francisco, as with other population estimates, is seen as a tool to guide policy-making and budgeting, and the accuracy of the source documents used to contrive the estimates are critical to effective health planning.
The trend: X x Y = Z
As Katz noted, only the first half of the equation which produced DPH-UCSF's widely-publicized estimate is based on the "percentage of new infections." This percentage, which represents the rate of new HIV infections in San Francisco, was derived from an in-house analysis of 25 of the best available disease incidence studies of San Francisco populations, many authored by UCSF or DPH researchers.
Larger studies and studies that were seen as better designed were purportedly given more weight during the May 24 consensus deliberations, which divined that the infection incidence rate among men who have sex with men in San Francisco had gone from 1.04 percent in 1997 to 1.68 percent in 2000. Among men who have sex with men and use injection drugs -- a much smaller group -- new infections increased from 1.99 to an estimated 4.58 percent per year, while HIV is actually decreasing among heterosexuals; notably, this decrease held true whether or not they inject drugs. (Among male-to-female transgendered women, the infection rate was estimated at 2.81 percent, though figures were not available from previous years for comparison, and were therefore not included in the DPH-UCSF bar charts.)
DPH-UCSF epidemiologists then multiplied these new, higher infection rates by the estimated size of the HIV-negative "at-risk" populations living in the city at the beginning of this year, resulting in the estimated number of new annual infections.
To achieve the second figure of their equation, however, the researchers first had to use an inflated estimate of the number of people who currently live in San Francisco. As with the first number of the DPH-UCSF equation (the estimated rate of infection), the second number was also arrived at by eyeballing available data sets from various sources, and synthesizing these into a wholly new approximation.
The problem is with one of the key sources used.
According to the Health Department's Policy and Planning office, all San Francisco departments and agencies, as in other municipalities across the state, use annual population estimates from the California Department of Finance to plan all programs and to guide policy-making decisions.
The state DOF estimates that as of January 2000, there were 801,400 people living in San Francisco, an increase of 1 percent from the 793,300 living here one year earlier -- still significantly higher than the city's .1 percent increase estimated by the U.S. Census Bureau in its latest update prior to the 2000 census reports.
Presumably leaving out children and lesbians, the DPH-UCSF juggernaut estimated that there were 584,395 heterosexuals and gay men in 1997, while as of this year the city held 696,094 -- an increase of over 100,000 people, or 20 percent, at a time when the vacancy rate hovers around 1 percent.
The total gay and bisexual male population went from 43,100 to 52,000 this year, according to DPH-UCSF -- an increase of over 20 percent from 1997. In addition to the total population, this estimate is predicated in part on a couple of telephone surveys in San Francisco and other cities, estimates that were revised upward to account for closeted homosexuals, and those with no telephone.
The researchers then estimated that HIV-negative gay men who do not inject drugs increased from 27,300 in 1997 to 34,014 this year, and HIV-negative gay men who do inject drugs increased from 2,665 to 3,120.
In any case, asked about the more accurate 1 percent rate of increase, which assumes that the displacement-challenged gay community is having no better luck finding apartments than straight people, McFarland calculated that the total number of gay and bisexual men in the city would have increased from 43,100 to just 43,500 this year.
And if DPH-UCSF had used the authoritative numbers used by all other local and state government agencies -- the 1 percent rate of increase -- McFarland said that rather than a total of 37,134 HIV-negative men who have sex with men in San Francisco in 2000, there would be 31,059.
After multiplying the lower estimates by their respective estimated increases in the rate of new infections, the total number of new infections among all men who have sex with men in San Francisco in the year 2000 would be approxima tely 528, still up from the 336 estimated in 1997, but 188 fewer men than the 716 cited this month by DPH-UCSF.
If the heterosexual population also increased by just 1 percent, rather than about 20 percent, one would arrive at a total number of new infections somewhere around 600, rather than the 900 number published everywhere from the San Francisco Chronicle to the New York Times.
"This is the down side," commented Ammiano. "As [former U.S. Surgeon General C. Everett] Koop said years ago, 'this is a health issue, not a political issue.'"
Moving right along
The convoluted results are nearly impossible to evaluate without having been one of the few who participated in the private deliberations convened by DPH-UCSF. But according to McFarland, who helped formulate the figures, there is a discrepancy.
"It would be fair to say, if you went with the more 'authoritative' source and kept the gay population estimates comparable," McFarland acknowledged to the Bay Area Reporter after reworking the math, "you would go from 336 to 528 [new infections among gay males]."
"When you get into the numbers crunching you get into the trust issue," commented Ammiano. "I wouldn't be adverse to having another hearing."
Trust is indeed a key issue, not only in evaluating the new numbers announced by DPH/UCSF, but in evaluating the new "11-point action plan" published as a response to these numbers, a plan that begins with the word, "ownership."
"The appropriate question at this point I think for a reporter," said Katz, "is, 'so what is the city and county of San Francisco doing about it?'"
ACT UP San Francisco
1884 Market Street *
San Francisco, CA 94102
Phone: (415) 864-6686 * Fax: (415) 864-6687 *