I heard this story on NPR recently and it has me brainstorming a pharmacoepidemiology paper about exclusion criteria (types, reasons, and effect).

Report Questions Excluding Gays From Some Studies
by The Associated Press
March 17, 2010
A small but significant portion of medical studies exclude gays from participating, sometimes without an apparent scientific reason, several cancer researchers say.
In a letter in Thursday's New England Journal of Medicine, three scientists from the Fox Chase Cancer Center in Philadelphia cite several dozen studies requiring a participant to be "in a reciprocal relationship with a person of the opposite sex."
There are legitimate scientific reasons for excluding gays from certain studies. Scientists would want only heterosexuals if they were studying how HIV spreads during male-female sex, for example.
But the Fox Chase folks found cases where the reason for excluding gays is not clear: tests of a drug for attention-deficit disorder, a treatment for erection problems after prostate cancer surgery, and studies on sexual function related to diabetes, depression and benign enlargement of the prostate as men age.
Brian Egleston, a biostatistician at Fox Chase, made the observation while overseeing enrollment of patients into clinical trials at the cancer center.
"When I first saw this, I thought it was a fluke. The second time, I thought I'd dig deeper," he said.
Egleston and Roland Dunbrack Jr., a biologist, and Dr. Michael J. Hall, a medical oncologist, did a spot check of a government database of thousands of studies and turned up more examples, most of them private-industry trials.
Researchers seeking federal money for their work must explain why a study excludes a group based on gender, race or ethnicity, but no explanation is needed for exclusion based on sexual orientation, Egleston said.
Exclusion can become self-perpetuating: Researchers designing a study often "cut and paste" participation criteria from earlier trials on a similar subject.
"It becomes the way it's done," and any bias gets repeated, Egleston said.
Estimates of how much of the U.S. population is gay or bisexual vary widely; some polls have put it around 4 percent.
———
On the Net:
Medical journal: http://www.nejm.org