Susan Ellenberg started out as a high school math teacher—then became a leading biostatistician

A mathematical puzzle fascinated Susan Ellenberg as a child.

Ellenberg occasionally chipped away at the question laid out by her father, a CPA, by randomly plugging in numbers. But soon, she discovered another way to approach it.

“When I got to high school algebra, I learned that there was an actual way to solve this problem. I was so excited I knew how to do it,” Ellenberg, emerita professor of biostatistics, medical ethics, and health policy at the University of Pennsylvania Perelman School of Medicine, said on the Cancer History Project podcast.

“With my algebra homework, I would often do more problems than were actually assigned because I just thought it was so cool that there was an actual way to do this and not just do trial and error and guess,” she said.

Ellenberg never intended to become a researcher, or even end up in statistics—a STEM field that has been welcoming to women for a long time.

“It wasn’t that way in mathematics, certainly. But in statistics, it was more friendly,” Ellenberg said.

This conversation is the first in a collection of interviews with women biostatisticians that will be available on the Cancer History Project.

Ellenberg wanted to be a teacher, but she struggled to decide what subject to specialize in. That changed after seeing her college entrance exam math scores—the highest in her graduating class, despite not having been in the honors math program.

“When I saw those scores, it was like a light bulb went on over my head,” Ellenberg said. “It was like, ‘Math! I could be a math teacher!’”

In the 1960s, Ellenberg attended Radcliffe College, then the women’s college associated with Harvard University, where she took coeducational math classes and quickly achieved her dream after graduating by teaching at a nearby high school.

With my algebra homework, I would often do more problems than were actually assigned because I just thought it was so cool that there was an actual way to do this and not just do trial and error and guess.

— Susan Ellenberg

After a year at that job, Ellenberg moved to a school in Maryland because her husband, Jonas, started a job at NIH. She taught there for two years, but then took a break to start a family.

But when Ellenberg was pregnant with her first child, her friend Janet Wittes—who had been graduate school classmates with Jonas Ellenberg—asked her for help with computer programming for a project for the eminent biostatistician Jerome Cornfield, then at George Washington University.

Ellenberg agreed to take the job.

She liked it enough to carry on after her baby was born. “Then, I thought, ‘Well, I had taken a couple of statistics courses in my background, and I’m working for statisticians now. I’m a math person, so maybe I should take another statistics course,’” Ellenberg said.

Ellenberg took an evening class at GW, and then another, until Jonas said, “If you’re going to take courses, you might as well be in a degree program.”

So, she enrolled in a doctoral program in mathematical statistics and continued working with Cornfield, who helped to train her in clinical trial design and analysis. Cornfield’s friends—fellow pioneers in statistics—including Nathan Mantel, Max Halperin, and Sam Greenhouse, the latter of whom became Ellenberg’s thesis advisor, frequently visited as well.

Ellenberg remembers Mantel peeking over her shoulder as she read an analysis of a clinical trial. He pointed out why the analysis was wrong, and suggested the correct way to approach it.

After eight years, in 1980, Ellenberg earned her PhD.

As she wrapped up her degree, Ellenberg moved to the nascent Emmes Clinical Research Organization, which had been contracted to be the statistical center for the Gastrointestinal Tumor Study Group. This became Ellenberg’s first experience with cancer and first time being the primary statistician on studies, where she worked alongside leading oncologists, including Charles Moertel and Robert Mayer (The Cancer Letter, July 8, 1994).

Ellenberg then transferred to NCI in 1982 to work with Richard Simon (The Cancer Letter, May 24, 2024). For six years, she primarily took part in the Cancer Therapy Evaluation Program and expanded her knowledge of cancer research.

But then, a job opportunity arose that Ellenberg couldn’t refuse: chief of the newly founded biostatistics branch of the Division of AIDS at the National Institute of Allergy and Infectious Diseases, or NIAID.

“Honestly, I could have worked my whole career at the National Cancer Institute—I really enjoyed it,” Ellenberg said. “But I felt like I couldn’t say no to this opportunity to work in this incredibly scary and new disease, bringing my experience in doing large-scale clinical trials to an infectious disease community that really had relatively little experience in these large-scale trials.”

Beginning in 1988, Ellenberg helped NIAID coordinate and design large-scale clinical trials to investigate azidothymidine, or AZT, in HIV-infected populations as well as compare AZT to other HIV/AIDS drugs. She also managed the Data and Safety Monitoring Board for the HIV/AIDS trials, and implemented closed sessions that shared interim data only with the DSMB and lead people within the NIAID Division of AIDS, but not with FDA or pharmaceutical companies.

“That was the origin of open and closed sessions for Data and Safety Monitoring Boards, which is the standard approach nowadays, but in those days, very unusual,” Ellenberg said.

Much of her time was also dedicated to working with AIDS activists. At the 1989 International AIDS Conference, Ellenberg picked up a brochure by the AIDS Coalition to Unleash Power, or ACT UP, describing how the members believed clinical trials could be improved.

“I was very excited to see this document. I thought, ‘These people should be our partners, not our enemies,’” Ellenberg said.

She began to organize informal meetings every three months where statisticians from institutions including NIH, Harvard, and FDA—and eventually, some clinicians—would meet with AIDS activists. Later, Ellenberg found out that NIAID director Anthony Fauci was having his own back-channel conversations with AIDS activists. He told the Division of AIDS director to allow activists to attend the formal AIDS Clinical Trials Group meetings.

“They were very productive. [The activists] had good ideas about how to run these studies in a better way, and many of their ideas were adopted,” Ellenberg said. “On the other hand, we were able to explain to them why some of their ideas really weren’t workable, and they listened because we were really having good, productive interaction.”

In 1993, Ellenberg joined FDA as the inaugural director of the Office of Biostatistics and Epidemiology in the Center for Biologics Evaluation and Research.

One of the big issues Ellenberg dealt with in that role was proponents of the anti-vaccination movement. Many people feared that the diphtheria-tetanus-pertussis vaccine caused sudden infant death syndrome, and that the measles-mumps-rubella vaccine caused autism.

People who believed that they or their children had been harmed by a vaccine could report it in the Vaccine Adverse Event Reporting System—a database that Ellenberg and her colleagues reviewed and published papers on.

“It certainly didn’t stop the anti-vaxxers,” Ellenberg said. “But what you hope it will do is prevent more rational people from becoming anti-vaxxers.”

In 2004, Ellenberg returned to her first career passion by moving to Penn’s Perelman School of Medicine, where she has remained for the past two decades.

“I said I would teach a class on clinical trials, which I did all the years until I became emeritus,” she said. “I really enjoyed doing that, [going] back to my original love of teaching.”

Transcript

Susan Ellenberg: What sparked my interest in mathematics?

Well, I remember—my father was a CPA, so, he was sort of a numbers person—and I remember a problem that he told me when I was—I don’t remember how old I was—that he remembered from his school days.

It went like this: John is twice Mary’s age when John was Mary’s age. When Mary will be John’s age, the sum of their ages will be 63. How old are John and Mary?

Now, that problem fascinated me, and I had no idea… I was doing trial and error things every once in a while. And then, when I got to high school algebra, I learned that there was an actual way to solve this problem. I was so excited I knew how to do it. I would often, with my algebra homework, I would often do more problems than were actually assigned because I just thought it was so cool that there was an actual way to do this and not just sort of do trial and error and guess.

I just always really enjoyed my math classes. But I never thought of myself as somebody who was really good in math. I mean, I was a good student. I got pretty much As in all my courses, and I wasn’t even taking the honors math course, you had to decide by sophomore year in high school whether you were going to be in the honors math section, because the honors math section did plane and solid geometry in one year. And if you weren’t in that, you did just the year of plane geometry, and then a semester of solid geometry later on. And the honors math program conflicted with the modern dance group that I’d been selected for. So, I said, “Well, alright, I’ll just take the regular math.”

So, we got to senior year, and my guy friends, who were all in the honors math program, they were all taking calculus that year and I wasn’t. But when we took the standardized tests that we all take for college applications, I had the highest math scores in my class.

So, I knew I wanted to be a teacher. All the time growing up, I knew I wanted to be a teacher, but I wasn’t sure what I wanted to teach. I was really good in Spanish. I thought Spanish was my best subject. But I couldn’t see spending my life teaching Spanish. I just didn’t feel motivated enough. There wasn’t any other area. Like I said, I was a good student.

But when I saw those scores, it was like a light bulb went on over my head. It was like, “Math! I could be a math teacher! That would be so great.” So, that’s kind of my whole origin.

And my mother used to tell me… my mother was very proud of the fact that she had gone to college during the Depression, and she’d been a biology major, a science major. But she was always saying, “Why are you taking all these math courses? Math is not for girls.” She would say this to me, but I ignored her.

Then, I went to college, and I went to Radcliffe, which was the girls’ part of Harvard then, but it wasn’t separate. All the classes were together. And I had to start in beginning calculus, because I hadn’t had calculus in high school, and I was pretty scared.

I mean, I grew up in Tucson, AZ and I didn’t know what I was going to face at Harvard. It was very intimidating, and I didn’t really know if I was going to have to struggle, to flunk out or whatever. But I knew if I wanted to be a high school math teacher, I certainly had to take at least the three-year calculus sequence. Started off in math one, had the most amazing teacher. He made everything so easy. I just sailed through it. And so, I had the confidence that, yeah, I could do this, I could be a math teacher.

SE: No, not at all, because I was not taking the high-powered math courses that the math prodigies were taking. I mean, my son was a math prodigy, and he was taking advanced courses.

But I was taking the early courses. The really super math people were not in the courses that I was taking. The people in the courses that I was taking were people who they were pre-meds or maybe they were engineering majors or something else where they needed the math—economics majors, government majors, people who needed to be able to do some data analysis.

So, it wasn’t really competing with the people who would have totally intimidated me. And there were plenty of girls who were pre-meds and economics majors and others. So, I was never like the only girl in the class or feeling like somehow, I didn’t belong.

SE: Well, I really enjoyed it. When I was in college, I worked in Project Upward Bound, which was sort of the high school version of Head Start. One of the great society programs where we worked with kids, high school kids from, quote, “underprivileged backgrounds,” of which there were plenty in Cambridge, MA, and try to build them up and provide opportunities for them to be ready for going to college and things.

I always felt that I would be a good teacher. I was a confident person.

I did my student teaching. And then I taught for a year at a high school in Cambridge, a big public high school, Cambridge High, and it was called Cambridge High in Latin then. And then I taught for two years in Montgomery County, MD, after we moved to Maryland when my husband took a job at the National Institutes of Health.

And I always really enjoyed teaching. I was not bored with it. I was very motivated by it. I was one of the sponsors of the high school math team. I just really enjoyed it.

But we wanted to start our family. We figured earlier is better. You’re healthier, everything is better, and algebra and geometry aren’t going to change over a period of seven or eight years. Have two kids and go back to teaching when the youngest is in preschool.

But what happened was, the summer before our first child was born, one of my colleagues who had finished her doctorate in statistics, along with my husband—they had been graduate school classmates—asked me if I would be willing to do some computer programming for her and the very eminent biostatistician that she was doing a research associateship with until the baby came. And I thought, “Sure.”

We had just bought a house. We hadn’t counted on my having any income, but that was going to be a plus.

So, I did that, and I really enjoyed it, and it worked out well. And then they wanted me to continue after the baby was born, so they got me equipment at home—and this was back in in the early 1970s—equipment at home that I could connect to the computer at George Washington University, and I would do the work at home. And once a week, I would get a babysitter, go into the office, or my colleague would come to the house, and we would go over my output.

Then I thought, “Well, I had taken a couple of statistics courses in my background, and I’m working for statisticians, now a math person, so maybe I should take another statistics course.” And it was at George Washington, almost all the graduate statistics courses were given in the evening, because there were so many people from the federal government whose departments were sending them to get their degrees, so that the courses were at night, so the people who were working could take them.

And so, I would go down to GW. But Jonas would come home from work and take over with the baby, and I would go down for my class. And so, that that continued, and after the second course, my husband said, “If you’re going to take courses, you might as well be in a degree program.”

So, anyway, I did that, and it took me a while. I took only one course a semester, not the usual two courses that might have been standard. And it took me four years to finish a dissertation. So, it took me a long time.

I was 34 when I got my PhD. But by then, I had really a whole lot of experience working with this wonderful biostatistician, Jerry Cornfield, who I was initially working for as a computer programmer, and then that kind of morphed into being more a clinical trial statistician as my training increased. So, it was a very nice progression.

SE: Well, Cornfield, like I said, was very eminent and very well respected, and he led our group in terms of getting contracts to be the statistical group for clinical trials and other research studies. So, I had worked on a case control study as the junior statistician looking at the effect of birth control pills on myocardial infarction in young women.

I also had the opportunity to be a principal statistician on a small clinical trial just being run and conducted at George Washington University on the use of heparin in patients undergoing hip replacement surgery to see how well it prevented blood clots.

And that was interesting, an amazingly wonderful experience, because I did everything. I developed the randomization list; I prepared the data entry forms. I wrote edit programs so that I could edit the data. The data would come in on paper forms, and I would enter them into the computer, and I would run edit programs to identify inconsistencies or missing problems, would convey that back to the investigators, design the statistical plan, ultimately analyze the data from the from the study, and was part of the publication.

I felt like I couldn’t say no to this opportunity to work in this incredibly scary and new disease, bringing my experience in doing large-scale clinical trials to an infectious disease community that really had relatively little experience in these large-scale trials.

— Susan Ellenberg

It was the best possible training for clinical trials because I understood the issues of how you had to design a data entry form to minimize the chance of error, and the importance of having an edit program and how you would put that together, and just everything that, for the most part, statisticians don’t do. Other people take care of those things in large scale clinical trials, but understanding those processes was incredibly good training for me.

And the other thing about working with Cornfield is that all of his other NIH colleagues, who had recently retired from NIH and were working on research grants, would come around. So, like Nathan Mantel of the famous Mantel-Haenszel and other statistical things that he did. He was around all the time.

I remember him looking over my shoulder one time when I was writing up something, I was doing some kind of analysis of possible heterogeneity or something in clinical trials outcomes. And he looked over my shoulder and he said, “That paper is wrong.” He saw the paper that I was using, and he said, “That’s wrong.” And then he went and got me a paper that he had written with a bunch of other people, showing what was wrong with it. And I said, “Oh, well, Nathan, then what should I do?” And he said, “Well, just do a likelihood ratio test.” So, that’s what I did.

But, I mean, that experience being around in a group with people like Nathan and Max Halperin and Sam Greenhouse, who ended up being my thesis advisor—these were all the early NIH great biostatisticians—was just an amazing and wonderful experience.

SE: Yeah, that’s how I knew what I wanted to do. And I was further motivated. Cornfield would take me and Janet Wittes—who had brought me in as a programmer, initially—he would take me to take us to meetings of major clinical trials groups where he was on steering committee. So, it would be a steering committee, and there were all these eminent trialists there, and we would hear what they would be discussing about the trial and how it was going and what the issues were. And that also was a very motivating experience to kind of see how you had to manage these large, long-term clinical trials, and to see how the statistician worked with these clinicians.

And there was all this mutual respect. It really instilled in me that a statistician was a true partner in these studies, and not just somebody who kind of sat back and ran some analyzes and then let the clinicians figure out what to say about them. It was a real partnership. And that was a that was a great model for me in my life.

SE: After I left George Washington, after I got my degree, or even just before I got my degree, I went to the Emmes corporation , which was just barely starting. I mean, I was their first PhD statistician appointee.

And they had gotten a contract—so, this is my first experience with cancer—they’d gotten a contract to be the statistical center for the Gastrointestinal Tumor Study Group, which is no longer in existence.

But it was my first time being the primary statistician on studies, and that was a very formative experience, working with some of the leading oncologists of the day, people like Chuck Moertel from the Mayo Clinic and Bob Mayer from Harvard, working on important large-scale cancer studies, which was very exciting. And I really enjoyed that.

But after a few years, I saw that Richard Simon—who recently died, but was had just an amazing career, worked in cancer statistics his whole life—he was advertising for a staff fellow position in his group.

I decided to apply for that, and I was selected and that was my entry into the National Cancer Institute, where I worked for six years, mostly as part of the Cancer Therapy Evaluation Program, where I got to be connected with the Cooperative Group Program and learned you know more about cancer research—a wider variety, not just GI cancer, but all cancers—and that was that was really interesting.

And I got to work with great people. Bob Wittes was the head of CTEP when I was there. And there were other really wonderful people that I worked with, especially Rich, who was an incredible mentor to me. And after six years, I did leave then to go, and I took my first supervisory position, as the first biostatistics branch chief in the Division of AIDS at NIAID.

Honestly, I could have worked my whole career at the National Cancer Institute. I really enjoyed it, but I felt like I couldn’t say no to this opportunity to work in this incredibly scary and deadly new disease, bringing my experience in doing large-scale clinical trials to an infectious disease community that really had relatively little experience in these large-scale trials.

SE: Well, of course, it was very motivating to be working on this problem. The AIDS clinical trials group had gotten started with the plan that it would be mostly a phase I, early phase II, looking for possible treatments for this disease.

But by the time I got there, AZT had shown to be highly effective in preventing mortality, and had been approved by the FDA. So, suddenly, this group and the statistical center that had been contracted to serve it, which had been expecting to do mostly small-scale phase I studies, certainly turned into the need to do large-scale clinical trials looking at AZT in earlier HIV-infected populations, and then comparing AZT to other new drugs that were coming out from other companies.

So, I was working, sort of being in between… there was one statistical center that had been contracted, but they weren’t really suited for these large-scale clinical trials. And the Harvard group that had been working in cancer, the group that had been the statistical center for the Eastern Cooperative Oncology Group, they knew how to do these large-scale clinical trials, and so, there was a sort of a transition to that group.

And so, I had a lot of involvement in helping out with that transition, the initial presentation, the initial getting involved in the design of the large-scale trials. The clinicians in the Division of AIDS had been the experts in working with the outside clinicians, the site clinicians who were going to actually perform the trials, and they were sometimes feeling a little left out, now that there were Harvard statisticians involved, and then they were the experts in study design. So, there were things like that that I had to help mediate.

Then, I was given the task of managing the Data and Safety Monitoring Board through the AIDS trials. The group that they had was probably the most eminent DSMB that you could possibly imagine. The statisticians on it were Tom Fleming and Dave DeMets, well known to the clinical trials community and two very eminent bioethicists, both of whom with substantial experience in clinical trials in medical areas.

And then, clinicians, at that point, there were not a lot of clinicians who were really knowledgeable about HIV and AIDS. It was so new. But these were great people. And we had some work to do to tweak how the DSMB was running.

Initially, the pharmaceutical people who were providing the drug, as well as the FDA, were in the closed sessions. They were aware of all of the interim results, which surprised me, and the DSMB, it turned out, was pretty unhappy with how that was working.

So, we worked out a plan to have the interim data only available to the DSMB and to the lead people within NIAID, within the Division of AIDS, who were monitoring the trials, but the FDA and the pharmaceutical company would not have access to the interim data.

That was pretty upsetting to them, and we had meetings to try and work this out, but what we ended up with was a plan that we would have an open session where they could come, and we would talk about the progress of the study, and they could talk about other work that they were doing that was relevant.

But then, when we looked at the actual interim comparative results, they would have to leave, and that would just be looked at by the DSMB. So, that was the origin of open and closed sessions for Data and Safety Monitoring Boards, which is the standard approach nowadays, but in those days, very unusual. I mean, for the most part, the pharmaceutical companies didn’t have any access. For the clinical trials, for example, that the Heart, Lung and Blood Institute was doing, there was no involvement with pharmaceutical companies or the FDA. So, this was kind of a new thing, but that was an important part.

The other really important part from my time in the Division of AIDS was working with the AIDS activists. The clinicians were very fearful of the AIDS activists. They felt that the AIDS activists were just out to get drugs approved willy nilly with no scientific evidence, no rigorous studies, and they didn’t want to have anything to do with them.

My first experience with the ACT UP community was at the 1989 International AIDS Conference, where they were handing out a brochure or pamphlets describing what they thought needed to happen and how they could make the AIDS Clinical Trials Group clinical trials better. When I looked at what they had written, it was very clear to me that they were not out to destroy the whole clinical trial enterprise. It’s not to say that there weren’t some activists who did want to do that. If you ever saw the movie the “Dallas Buyers Club,” there was that segment.

I said I would teach a class on clinical trials, which I did all the years until I became emeritus. I really enjoyed doing that, [going] back to my original love of teaching.

— Susan Ellenberg

But ACT UP people, they wanted rigorous research. They just thought we could do it more efficiently. And there were obstacles that we were putting in the way of doing good research efficiently and humanely.

So, I was very excited to see this document. I thought, “These people should be our partners, not our enemies.” And we got a bunch of people, statisticians at the NIH—I had a meeting. I came back, and I invited the statisticians working in AIDS research, people from my group, my small group, some of the Harvard people who were working as statisticians on the trials, statisticians from the FDA. And two days before the meeting, I got a telephone call from someone from ACT UP, who had heard that we were having this meeting, and said, “We need to be there.” So, we actually had two ACT UP members. We also had some statisticians from the National Cancer Institute, as they were also doing some work in AIDS.

And it was a very exciting meeting. Everybody was very excited about seeing what the AIDS activists—or this group of AIDS activists—were really proposing, and we wanted to have another meeting right away. So, we had another meeting two weeks later, and we ended up having regular meetings every three months at the same time as the AIDS Clinical Trial Group was having their meetings. The activists weren’t allowed to be in the ACTG meetings, but they came to our meeting, and more and more people were coming to our meeting. There were maybe 20 people at the first meeting, there were 40 people at the next meeting. Pretty soon they were 100 people, 120 people. Some of the braver clinicians started coming to our meeting, some of the clinicians at NIAID and more of the FDA people.

And eventually, what I was not really aware of until later, was that Tony Fauci, who was head of the institute, he was having his own back-channel conversations with the AIDS activists. He also recognized that these people were not against rigorous research. They just wanted it to be done better. And they felt that using the tactics that they did, which sometimes involved calling all of us murderers and climbing up into the buildings of the NIH or the FDA, and making nuisances of themselves, that that was a way to get attention. So, Fauci recognized that.

Eventually, he had told our division director, who was siding with the clinicians, the ACTG clinicians, about keeping that. He told him, “You’ve got to open this up to the AIDS activists.”

And they were very productive. I mean, they had good ideas about how to run these studies in a better way. And many of their ideas were adopted. On the other hand, we were able to explain to them why some of their ideas really weren’t workable, and they listened because we were really having good, productive interaction. So, that was a very, very satisfying period in my in my career.

SE: Yes. When I look back and I think about what we did, organizing these meetings every three months, we had speakers coming in, eminent statisticians who had done some work that was related to the AIDS research experience. Sometimes somebody talked about what you can do with observational databases, because one of the AIDS groups was developing that, others talking about their research and how to account for people who weren’t adherent to their medication. So, I don’t think anybody ever said no, they wouldn’t come and talk to us.

But when I think about organizing those meetings four times a year, I say, “Well, did I really do that?”

But it was great. It was a wonderful time. I mean, it was a terrible time in many ways, but in terms of being motivated to do what we were doing, that’s what was wonderful.

SE: So, Biologics, at that time, would regulate biological therapies, which were the large molecule therapies, and which were much more complicated to manufacture. And so, the manufacturing process was something that was looked at very closely, and that’s how these drugs were differentiated from the small molecule drugs, which were regulated by the Center for Drug Evaluation and Research, but we also had responsibility for vaccines and blood products.

One of the big issues with vaccines… We are very aware of the anti-vax movement now, but there has been anti-vax for a very long time, and that was something that I had to deal with when I came into the to the FDA.

There had been pushback on the DTP vaccines. There had been a whole big thing with, does it cause sudden infant death syndrome? There had been studies that showed that it didn’t, but if you have someone whose child has had their two-month vaccines and then two weeks later, they’re found unresponsive in their crib, you can understand how parents would think—especially, they would find, there were other people who had similar experiences because, unfortunately, there were a number of SIDS deaths every year. And when you vaccinate children at birth or at one month and at two months and at four months and at six months, it’s just going to happen by chance, that there is going to be some in close timing with vaccination. So, that was difficult.

And then there was the horrendous experience with Andrew Wakefield in the UK, who wrote a paper claiming that the measles-mumps-rubella vaccine was associated with autism. And after that, the preservative thimerosal, which was in tiny amounts in some vaccines, was that going to cause autism?

So, we had to deal with all of those kinds of things. And what I did was there was this system called the Vaccine Adverse Event Reporting System, or VAERS, where anybody who felt that they or their child or somebody they cared about had been injured by a vaccine, or possibly injured by a vaccine, it could be reported. And anything that was ever reported was in this database, no matter how implausible it seemed. And so, our job was to review the data in this database to see if there was any signal that there were some serious problems with vaccines.

So, one of the things that some of the anti-vax people were saying is, “Oh, yeah, there’s this system, but nobody looks at it. Goes into a black hole. Nobody ever cares about what’s in there.” So, I had the epidemiologists in my group, as a routine exercise, after a new childhood vaccine was approved, say, maybe a year or two, by the time there was a couple of years’ worth of experience, had them review all of the data that had come in on that new vaccine. And then they would write up a report, write a paper about what they had found. And those papers ended up being published in highly reputable medical journals. Several of them appeared in the Journal of the American Medical Association, for example.

And so, that was one thing that I started to do. It certainly didn’t stop the anti-vaxxers. But what you hope it will do is prevent, more rational people from becoming anti-vaxxers, people who see this and start to understand that, yes, we are looking at these data. Yes, if there is something there, we will find it and look into it.

So, that was very challenging. And the people who were very focused on possible, or what they felt were definite reactions to the vaccine that caused great harm to their child, they were interested in research, but not in the way the AIDS activists were. They were interested in research that would prove that the vaccine harmed their child, not research to understand, the kind of observational research that they would do to see if there was any real signal that there was an increased number of adverse events.

Of course, we did find one case where there were serious adverse events associated with the first rotavirus vaccine, and the manufacturer ultimately withdrew that from the market. Anyway, so that was a big challenge.

And then, the biological therapeutics started to really come into their own with drugs like infliximab and dinutuximab, drugs that were used to treat autoimmune disease and cancer. And so, there was a lot of excitement about the new era of biological therapeutics for serious diseases.

SE: Well, I think it’s very important.

I sort of had my first experience with that in the AIDS research area, where I would get interviewed from time to time, and had to explain things to the public. I remember one time explaining—this was at a larger meeting. I can’t remember whether it was in an AIDS Clinical Trials Group meeting after the activists had been admitted, or what it was in. But I was asked to talk about the Data and Safety Monitoring Board and how it worked.

Nowadays, there are often representatives of a patient community on a DSMB. That’s true for all the cancer cooperative group DSMBs now.

In those days, I did not feel that it was a good idea to have somebody from the community—whether they were HIV infected or not—on the DSMB, because if you’re on a DSMB, you have to keep the interim data that you see that you’re monitoring, very confidential. And I just felt it was too big a burden at a time when this was a fatal disease, because I left for the FDA three years before the protease inhibitors came on the market, which turned HIV infection into a chronic disease and not an inevitably lethal disease. I just felt that it was too big a burden to ask people in that community if they saw that something was looking very promising to keep that to themselves. I just thought that was too much to ask. It was unreasonable. It was not a question of trust. It was a question of it’s just too much to ask of these people to keep that to themselves. And I explained that at this meeting, and I was not sure the reaction I was going to get. But nobody objected. Nobody complained about it.

It’s important to be able to explain these things to the public. I wrote with my CDC colleague about how to interpret the data from the VAERS system. And we published an article in the journal Public Health Reports, which I’ve ended up sending to a lot of people during the COVID time, because there’s been a lot of misuse of the VAERS database to unfairly say negative things about the COVID vaccines.

SE: Well, it was a great experience. The chair of our DSMB was Richard Whitley, who had been chair of that first AIDS Clinical Trial Group DSMB back in the days when I was working at NIAID. And he’s just a wonderful researcher and a wonderful man. And it was great that he was doing it.

And it was a very fraught time. We weren’t monitoring the Pfizer vaccine; they did it separately. But we were monitoring the clinical trials for Moderna and the J&J vaccine, the Novavax, a couple of others. And we knew, especially the Moderna, that was a new technology—nobody knew anything about it. So, we were very, very focused on the safety. But, of course, we were focused on the safety for all of them.

But the desperate need for a vaccine meant that there was not any problem in recruiting people to these studies. They were very large studies, and they recruited very quickly. And it was the glorious day when in late November—well, actually it was early November—in 2020 where it was clear that the Moderna vaccine was highly effective and very safe. And the information on the other vaccines followed somewhat after that.

But it was just a wonderful experience to be part of that, and to make sure that we were looking at everything very carefully.

SE: Well, it was certainly exciting for both my husband and me.

Neither of us had had an academic position before, but we’re both quite knowledgeable about clinical trials, and we were able to become part of important collaborative projects, which were exciting. I said I would teach a class on clinical trials, which I did all the years until I became emeritus. And I really enjoyed doing that—back to my original love of teaching. Although I did plenty of teaching all through my career, giving lectures and seminars and presenting it at national, international meetings.

But this was—actually, I’d never run a full course like this. So, it was a bit of a challenge. But I really, I really enjoyed that.

And there were a lot of advantages in being in an academic setting. I remember asking when we first came on, asking the head of our center, I said, “How many vacation days do we have? How does that work?” And he looked at me, and he said, “Vacation days? Nobody keeps track of vacation days. You do your work, and if your work doesn’t get done, somebody may look and see what you’re doing with your time.”

There were faculty that we had who were never in the office, unless they were teaching a class or meeting with project members or meeting with students. They did their work at home. And that was fine, as long as they got their work done and as long as they were there for when they needed to be there. So, there were things like that that were kind of nice parts of being an academic. And it got me back to being the principal statistician on clinical trials, which I hadn’t done since my years in the National Cancer Institute, which I really enjoyed.

So, in each position that I took, moving from one to the other, I was able to take things that I had learned in my prior position to apply in my new position. And then there were always new things to learn in the new position.

SE: I would say so.

SE: Well, I would say that statistics, the field of statistics, has been a very friendly field to women for a very long time. I mean, there was a president of the American Statistical Association who was a woman back in the 1940s, I think.

So, I never felt I never felt diminished or overlooked in any way being a woman. In fact, sometimes I felt that I had a leg up, that there was some kind of subconscious affirmative action thing that went on among statisticians that they said, “Well, here’s a competent woman, here’s a competent man, but we’re going to give an edge to the competent woman, because we want to have more women in the field.” Because it was certainly true that there were fewer women than men in the profession as a whole. I doubt that’s true now, but it certainly was when I started. And I felt I benefited from that, that I, in fact, benefited from being a woman rather than being held back by it.

There were other leaders, particularly in cancer in those early days. Judy O’Fallon was the principal statistician for the North Central Cancer Treatment Group. Carol Redmond was the principal statistician for the National Surgical Adjuvant Breast and Bowel Project. And there were other leading women statisticians—Stephanie Green in the Southwest Oncology Group, and people like Dianne Finkelstein and others in Eastern Cooperative Oncology Group.

So, there were plenty of women having leadership in my experience. I certainly wasn’t the only one.

SE: Yeah, I don’t know, and it certainly wasn’t uniform. I remember being on the visiting committee for the Harvard Department of Statistics. Now, not biostatistics, but statistics.

And I remember the dean came and talked to our committee and pointed out that the Department of Statistics had never had a tenured woman in the department, and, in fact, never even had a junior woman in the department, not since I had been an undergraduate there, however many decades ago that was. And he made it clear that if they didn’t do something about that, there wasn’t going to be any Department of Statistics anymore at Harvard.

So, you know, I wouldn’t say it was uniform that there was a welcoming to women. So, I can’t really explain why that was. I think the people that I worked with, people like Jerry Cornfield and Paul Meyer, a great biostatistician, who spent his career, first at the University of Chicago, I think, and then at Princeton, and then finally at Columbia. They were all very, very friendly to women.

So, I don’t know why it was in that career. And it wasn’t that way in mathematics, certainly. But in statistics, it was more friendly.

SE: Well, there’s been a huge amount of methodological research about the best ways to do clinical trials. Of course, I’m focused on clinical trials. There’s been lots of research of lots of other areas, too. But we’ve learned a lot about how to do clinical trials more efficiently, how to manage the adherence problem and account for it in more reliable ways. The whole era of adaptive designs and how we can adapt.

When I started out, there were no formal statistical ways to stop a trial early if the results were looking really, really positive. It was all kind of intuitive on the part of whoever was monitoring the interim data. Early on, there weren’t data monitoring committees or DSMBs. It was just study steering committees or the investigators themselves who were watching the data. And there were lots of times where they saw the data creeping up, and as soon as it crossed the 0.05 boundary, they’d say, “We have to stop. It’s unethical to go on.”

Tom Fleming and Dave DeMets and other people wrote papers to show how—well, and Peter Armitage in the UK was really the first one—to show how you inflated your type I error when you did that. And then people like Tom Fleming and Dave DeMets developed monitoring guidelines to show you where you really needed to be in order to be very confident of that p less than 0.05 when you stop the trial.

So, there have been important advances like that. And of course, you know many others. Jerry Cornfield was a Bayesian, and so, he was interested in Bayesian approaches for trials, but they weren’t very much in use. But in recent years, there’s been a much greater use, interest in Bayesian approaches. The problem 50 years ago was that you really needed a lot of computing power to do this, and we didn’t have it then, and now we have it.

And so, there’s been a lot of changes and a lot of improvements in how we do and how we manage—in design and manage and analysis of clinical trials. It’s been great.

SE: It’s been great to talk to you, and I hope you have opportunities to talk to a lot of other people, a lot of other women who have been leaders.

“Corporation” not part of the formal name, from what I understand