Six years ago, San Diego obstetrician-gynecologist George Delgado set up a national hotline to help women who change their minds after taking mifepristone, the first of two drugs used in a medication abortion.

Delgado, an abortion foe, had treated a handful of such women and believed that high doses of the hormone progesterone could stop or reverse the abortion if administered before the second drug was taken.

Even though the American College of Obstetricians and Gynecologists and other medical groups have called the treatment unproved and unethical, five states — Idaho, Utah, Arizona, Arkansas, and South Dakota — have passed legislation requiring clinics to tell women about "abortion pill reversal."

This week, Delgado and colleagues published a study using data collected through the hotline, which refers women to local doctors who prescribe a wide variety of progesterone regimens because no one knows which works best. The study – in Issues in Law and Medicine, a journal with ties to the American Association for Pro-Life Obstetricians and Gynecologists – reports an overall 48 percent success rate in reversing mifepristone, with certain regimens even more effective.

Critics swiftly called the research flimsy and flawed in ways that likely inflate the purported effectiveness. They also pointed out that studies have found up to half of women who take mifepristone alone, and skip the second drug, called misoprostol, will continue their pregnancies without any further treatment.

"There's no evidence, based on this paper, that progesterone is more effective [at continuing a pregnancy] than doing nothing" after taking mifepristone, said Daniel Grossman, a University of California, San Francisco obstetrician-gynecologist who studies reproductive health issues. "It would be very concerning if more states pass laws mandating that physicians tell women about an unproven regimen."

Delgado hopes that's exactly what happens.

"Now that it is clearly shown safe and effective for women having second thoughts, they have a right to know… there is an option for reversal," said Delgado, medical director of an anti-abortion nonprofit called Culture of Life Family Services.

>> READ MORE: Study finds states regulate abortion clinics more strictly than outpatient surgicenters

No one knows how many women who begin a medication abortion change their minds.

Under the protocol, a woman takes mifepristone (brand name Mifeprex) up to 10 weeks after her last menstrual period. It blocks progesterone, a hormone needed to sustain a pregnancy. Up to 48 hours later, she takes misoprostol, which causes uterine contractions and expulsion of the grape-size fetus. The two-drug protocol evolved in the 1980s because researchers recognized that, by itself, mifepristone is not sufficiently effective at terminating pregnancy.

Medication is now used for about 31 percent of the nation's 925,000 annual abortions. In Pennsylvania, the rate is slightly higher: 36 percent in 2016.

In an interview, Delgado said he had "never heard of anyone trying to reverse mifepristone" until 2009, when he got a call from a woman in El Paso, Texas, who was desperate to do that.

"I knew mifepristone blocks progesterone receptors," he said. "And I had used progesterone in my practice. I kind of put two and two together."

Progesterone has a long history of use in infertility treatment and to prevent preterm birth. No long-term risks have been identified for women or their offspring due to its use. Delgado's theory is that flooding a woman's body with progesterone within 72 hours of taking mifepristone "would favor progesterone effects and blunt the abortifacient effects."

The new study, he said, validates this theory. Of 547 women who got treatment and were included in the final analysis, 261 – 48 percent – gave birth. All babies were healthy.

Critics say there are so many holes in the data and so many women were excluded from the final analysis – including 57 who changed their minds again and had medication or surgical abortions – that the study is meaningless.

One glaring omission that Delgado and his co-authors acknowledged in the paper: They didn't have data on ultrasounds. An unknown number of women didn't get the progesterone treatment after an ultrasound revealed their embryos had died. Without that data, critics say the treatment could seem more effective than it actually was, or at least the results were muddied.

"It is possible that those embryos who were alive at the time of the sonogram would have survived without progesterone therapy," the researchers acknowledged in the paper.

The paper comes at a time when both sides in the bitter abortion debate are seeking to use state laws to make sure women are "informed." Last month, the U.S. Supreme Court heard arguments regarding a California law that requires crisis pregnancy centers to tell clients that they are not licensed medical facilities, and that state-subsidized abortion is available. At issue is whether the requirement violates the centers' First Amendment right to free speech.

The abortion pill reversal laws are a different issue, in the view of the American College of Ob-Gyns.

"It is dangerous for outside parties like politicians to force physicians to steer their patients toward experimental treatments whose safety and efficacy is unproven," said Hal Lawrence, executive vice president of ACOG.