by Alice Dreger
Looking back at this interview with Wilson, hoping that I would someday again feel capable of helping people through my work, I drafted marching orders for myself: Stop thinking the press will get it right. Stop thinking you need them to get it right. Keep working for the long term. Disengage from the immediate fight and focus on knowing more, knowing as much as you can.
But in the short term, how to work past my sense of having failed at protecting the rights of families affected by CAH and intersex? For that, I found a possible answer in the survival advice I’d also collected the year before dex from the famous psychology researcher Elizabeth Loftus. Today Loftus is generally recognized as having been right—that human memory is fallible and that you can actually implant false “memories” in a person. However, in the late 1990s, Loftus had found herself in very deep trouble over her work challenging recovery of repressed memories of alleged child sexual abuse. She and a colleague, Melvin Guyer of the University of Michigan, had decided to research the index case of recovered memory. They had ended up gathering evidence that suggested that the poster child for repressed memory of childhood sexual abuse, a woman identified in the 1997 case study as Jane Doe, almost certainly did not experience the childhood sexual abuse at the hands of her mother that had been described in the context of an ugly custody battle. Loftus and Guyer had used public documents and interviews to make their case that the girl’s “memories” of abuse likely resulted from suggestions made to Jane Doe when she was a child, not from actual events.
Before Loftus and Guyer could publish, Jane Doe—whose real name is Nicole Taus—made formal complaints, saying her privacy was being violated. As one might expect, people who believed they had recovered memories and psychiatrists engaged in memory recovery angrily rallied against Loftus and Guyer. Eventually cleared by their universities, the two went on to publish their exposé, but Taus proceeded to litigation. Given that Loftus and Guyer had employed conventional investigative journalistic methods and had not even published Taus’s real name you would think the case would have been quickly dismissed. But as the claims and counterclaims became ever more complex, the suit survived, making it all the way to the California Supreme Court—and taking over a great chunk of the researchers’ lives.
In the end, they prevailed. Because Loftus’s story had already been so well documented, what I had wanted to know most from Loftus when we had talked was how she had survived and continued to advocate for those wrongly accused through allegedly recovered memories. Now in my dex mess, I had advice from her to make my own. Some of her strategies seemed obvious: You keep working, you pay good lawyers, and you hold trustworthy friends and colleagues near. But one of her strategies had surprised me: She told me that during the worst years, she took up the habit of watching the Lifetime television network. She explained that the same basic story in different guises runs over and over again on Lifetime, the story of a woman facing tremendous adversity who somehow sticks it out and survives. Loftus’s approach made her philosophy clear: If you think you’re working for the greater good, you take the knocks and keep working, doing good research to figure out reality. You stop worrying about yourself. And so—staying firmly focused on the work that matters—you survive.
Knowing there probably wasn’t going to be a good historian who would do the work to tell me what actually happened with the whole prenatal dex scene, I realized that what Aron was telling me was right: I had to stop acting like a beaten dog and act like a professional historian. I had to treat this like a new major historical project, and I had to have the patience to work on it as long as it took to figure it out. As I made findings, I would have to present and publish them in scholarly arenas and in the mainstream press, so that people could see the evidence I’d found and help bring justice to bear. It might take several years, but this project was important enough to give years, because this really did look like DES all over again, even as to the reluctance of the government to stop the travesty, recalcitrance in the name of letting individual doctors decide what is best for their patients, no matter the science, no matter the ethics. If I did this work right—if I focused on truth and justice and not on my own misery—eventually I would help the CAH- and intersex-affected families and the ethical medical researchers, as I had originally set out to do.
So I crawled out of my hole and pulled myself forward. I started by carefully looking into the characters in the drama. I put out a series of feelers to find new leads. I tracked medical conferences and the medical literature to find any new data. I corresponded with strangers who sent me information, among them several mothers who wrote to tell me what they had been through with Maria New. (Such stories were only “anecdotal evidence,” but they gave me leads.) I learned how to use the Freedom of Information Act (FOIA) to get copies of New’s grants and information on the federal investigations we had initiated. When the OHRP FOIA office wouldn’t give me what I asked for and the FDA FOIA office wouldn’t even acknowledge my written requests, I hired a lawyer and sued the government. Meanwhile, when I had nothing new on dex on my plate to analyze, I kept up my reputation as a researcher and health and science writer by working on side projects. In private, I distracted myself with pro bono, confidential, medical-history-recovery work for individuals who had been medically traumatized, work that made me think I had some use to the universe, even if it ultimately turned out that I had been all wrong about dex. And day by slow day, week by slow week, I pressed on with investigatory historical research into dex.
After a few years of steady work, I could finally see clearly what had happened with dex. I could finally see where we’d been right, where we’d been wrong, and where we’d perhaps been played.
• • •
AS WE HAD FEARED, prenatal dex did turn out to be a story of how layers upon layers of medical research protection systems can fail—and fail even around the group recognized to be the most vulnerable: human fetuses. As we hadn’t fully understood, prenatal dex also turned out to be a story of what results when you get a perfect storm of beneficent intentions, professional loyalties, gut-level fears of sexual anomalies, unmanaged conflicts of interest, and mindless bureaucratic paper-pushing. To my horror, prenatal dex also turned out to be something of an ethics canary in the modern medical mine; whereas I had assumed at the outset of my investigations that prenatal dex would always be a story of exception, over the years I learned that in fact the same story of failed protections—and failure of truly informed consent—may well be playing out in clinics all over the United States.
That’s the short version of what I discovered in those three years of work. Here’s the slightly longer one:
Maria New’s main NIH grant, a grant to study many forms of and treatments for various adrenal disorders, had been consistently funded all the way back to the 1970s. When the idea of using prenatal dexamethasone to prevent virilization in CAH-affected females came along in the mid-1980s, New successfully rolled the intervention into her multiarmed grant renewal applications, thereafter using prenatal dex to give the NIH more reason to fund her work. She started offering prenatal dexamethasone for CAH through her Cornell clinic in 1986, and right around that time, presumably because she planned to publish on the intervention, she sought and obtained Cornell research ethics committee (IRB) approval to study what happens when you dose pregnant women and their fetuses with dexamethasone starting very early in pregnancy to try to prevent intersex in the female offspring. These pregnancy drug trials were not rigorous, in spite of the fact that the goal was to have the drug cross the placenta and change fetal development, in spite of the fact that seven out of eight of the offspring exposed in the first few weeks of embryonic development could stand no benefit and would assume all the risks. There was no placebo control, and no blinded assessments of results. In the mid-1990s, Maria New’s team published some results, comparing girls who had been exposed to prenatal dexamethasone to sisters who had not; they claimed that this showed the intervention worked
.
Contrary to what I had thought when we sent our complaints to the feds, New did seem to have consistently gotten IRB approval at Cornell for studies of prenatal dexamethasone for intersex prevention. However, New’s IRB applications at Cornell, which I obtained via FOIA, throw up multiple red flags, flags which apparently did not attract the notice of the IRB members. For example, in 1985, in New’s first Cornell IRB application to include (among many other endocrinological studies) a plan to study the use of prenatal dexamethasone for CAH starting at two to four weeks of gestation, New did not check the “special populations” boxes on the front page of the form to indicate that her subject population would include pregnant women and fetuses. Checking those important header boxes would normally have set off special layers of review. Checking those boxes would have ensured that absolutely everyone in charge of managing research subject protections knew that, deeper in the packet, there was a plan for fetal experimentation—and yet as New renewed her IRB approvals each year, no one serving on the Cornell IRB seems to have noticed these “special groups” boxes were not being checked. It took sixteen years for someone to notice.
The consent forms that the Cornell IRB approved for New’s prenatal dex work also minimized the risks of this experiment in shocking ways. Any woman who did sign them could not legitimately be said to have given informed consent to put herself and her offspring into this experiment. Instead of describing the risks as essentially unknown, the 1985 consent form approved by Cornell described potential harms of dexamethasone exposure as “transient and reversible suppression of the maternal and fetal adrenal gland.” A pregnant woman reading this description would have reasonably assumed any harm from prenatal dex would be—well, transient and reversible, especially given how the form went on to emphasize the relatively low dose and that “pregnant women and fetuses treated to date with this regimen have not experienced complications.” There is no way, given the high risks, the unknowns, and the extremely controversial status of this intervention, Cornell should ever have allowed such a reassuring description as put forth in this form. But problems with the consent forms persisted year after year, including in that they did not contain appropriately updated information about findings of possible harm to mothers and children from prenatal dex. As late as 2004, Cornell’s IRB was still approving consent forms that reassured pregnant women that prenatal dexamethasone “has been shown to be safe for the fetus” even while the supposed point of the study was to determine “long-term safety of prenatal treatment.”
New repeatedly boasted to the NIH that a strength of her research team was that her clinic had far more CAH-affected and dex-exposed patients to use as research subjects than any other researcher. By her own admission, she had this advantage of numbers specifically because offering prenatal diagnosis and dex treatment brought CAH-affected families to her door. In her 2001 grant renewal application, for example, she told the NIH, “The prenatal diagnosis and treatment program has provided a new source of patients with CAH, adding about 15 patients per month. Clearly we are in a position to expand the [clinic’s] database to provide sufficient patients to answer the [medical research] queries relevant to CAH.” That particular NIH application included a table describing human subjects under the research arm called “prenatal dx [i.e., diagnosis] and treatment in families at risk.” The total number of her research subjects for that study aim is stated as 2,144. So prenatal dex wasn’t just a boon in and of itself in terms of clinic and research income; prenatal dex allowed New a big boost in numbers of CAH-affected people who could then be used to obtain research funding for studies beyond prenatal dex.
To get and maintain her NIH funding for experimentation involving exposing fetuses to dexamethasone, New had to show that she had Cornell’s IRB’s approval for such studies. As noted above, it appears that, indeed, she consistently got Cornell’s IRB’s approval for studies that would have involved pregnant women, and that she then reported to the NIH the IRB approvals. Keep in mind she didn’t have to show the NIH signed consent forms from the mothers she was naming as subjects, or even the blank consent forms—just IRB approval to enroll subjects. The NIH would have assumed that the human subjects she was talking about in her grants had been properly enrolled—that they had signed well-vetted forms that meant they had been fully and honestly informed about potential benefits and risks before agreeing to become experimental subjects.
For many years, New’s work chugged along, approved and funded annually. By 1996, the NIH was specifically funding New to study whether prenatal dexamethasone for CAH could “succeed” in making CAH-affected females, in effect, more likely to be straight and interested in becoming mothers. No one at the NIH seems to have raised a question about whether this was a reasonable goal of clinical care or medical research.
In August 2001, the young, Greek pediatric endocrinologist Kyriakie Sarafoglou was hired to work in the pediatric research program at Cornell’s medical school. Under a new national program designed to provide on-site monitors for big NIH study locations, Sarafoglou was employed on NIH money as a “research subject advocate,” overseeing patient safety at Cornell’s Children’s Clinical Research Center. Only a few months into that job, Sarafoglou realized she was seeing all sorts of problems with the $23 million NIH grant for which New was the principal investigator, including financial and ethical irregularities. For example, Sarafoglou noted concerning irregularities on New’s IRB materials. Sarafoglou first tried raising the alarm within Cornell. This led to an internal investigation that found nothing especially concerning. The internal Cornell report, which I obtained via FOIA, praised Maria New for her storied career and suggested Sarafoglou needed replacing.
Unsatisfied with Cornell’s response, Sarafoglou contacted the OHRP with her concerns, documenting them with thousands of pages of internal paperwork. Although OHRP proceeded to investigate, the agency did so at the usual glacial pace, perhaps because Sarafoglou had sent huge quantities of evidence of systemic problems in Cornell’s pediatric research. With the investigation ultimately involving the Department of Justice, things got hot, and Sarafoglou was given notice that her position was being terminated. About a year after she had called on OHRP to investigate, Sarafoglou sent an anguished message to the agency: “It seems that my academic career is over. I just wish I had [had] the foresight to know the pain to me and damage to my career that filing with OHRP would cause.”
Based on what I found when I FOIA’ed the OHRP investigation Sarafoglou initiated in 2003, it appears a major point of concern was IRB records at Cornell that made no sense. External reviewers looking at Cornell’s pediatric research records found misleading consent forms as well as “at least one protocol in which a significant number of subjects were enrolled an [sic] on whom no documentation of informed consent could be found.” One reviewer concluded that, although the people at Cornell clearly cared about children, “The IRB members interviewed did not appear to have a thorough understanding of the additional protections for children as subjects of research.” In an angry memo that seems to have been written by Sarafoglou, someone complained that New was claiming “130 or 160 patients have been accrued each year [to New’s IRB-approved CAH studies] but [there is] not one single dropout or patient studied under another protocol,” a virtual impossibility. Yet the records from Cornell show New reporting to her IRB, in her annual applications for renewal, that not a single potential subject refused to be in, withdrew from, or had a complaint about her IRB-approved study protocol. Moreover, in a letter to her IRB, New claimed with regard to a study on prenatal dex for CAH that, “To our knowledge, there have been no adverse events in patients due to studies performed,” even while she cited in that letter a publication she had co-authored which had reported adverse events for exposed mothers and their offspring! Incidentally, all medical researchers know that IRB-reportable “adverse events” include any medical problems while in the study, not only those you can pin to the intervention. So what a
re the odds studies involving subjects who are pregnant women and developing fetuses would not involve any reportable adverse events? Again, this is virtually impossible.
From 2003 to 2005, in response to Sarafoglou’s whistleblowing, OHRP ultimately undertook a massive review of Cornell pediatric research. OHRP found so many problems, they took the extraordinary step of requiring review by Cornell’s IRB of all active pediatric research at Cornell. One of the projects about which OHRP had pointed questions at that time was New’s supposed research on prenatal dexamethasone for CAH. In a letter dated May 24, 2004, OHRP asked Cornell to “Please clarify whether the subjects described in the Journal of Clinical Endocrinology and Metabolism [1995 publication] entitled ‘Extensive Personal Experience: Prenatal Treatment and Diagnosis of Congenital Adrenal Hyperplasia[’]. . . were enrolled in a research study. If they were, please provide a copy of the complete IRB-approved protocol for these studies.” What OHRP was asking was not just whether Cornell’s IRB had given New permission to do a study, but whether she had actually had the hundreds of pregnant women research subjects enrolled in a formal study by signing the consent forms. After a lot of back-and-forth in which Cornell said they were having trouble finding the records, and in which Cornell reminded OHRP that New was now gone from Cornell, OHRP just decided to stop tracking what had happened to the pregnant women given dex. They just had too much else to sort out.
This is troubling—OHRP dropping the question of the ethics of the matter. But what is more troubling is that when, in response to OHRP’s inquiries, a Cornell administrator asked New about the pregnant women mentioned in that publication, New wrote back a curt memo informing the Cornell IRB rep that, “Prenatal diagnosis and treatment is a standard procedure at the patient’s request and has been performed in France since 1981 and in our lab since 1986.” In other words, she was stating that so far as she was concerned prenatal dex wasn’t experimental—so it didn’t require consent to research. It was “standard” practice given because patients “requested” it.