Elizabeth Barrett-Connor

Quotes

I started Rancho Bernardo with concurrence of Joe Stokes, who several years later showed me my letter to him announcing why I thought this was a good cohort. And one of the things I said was it would be really important to study “old people over the age of 50.” I don’t know where that letter is now, but I do remember him showing it to me with some pique. I was 35, I think. (5)

So I became the epidemiologist for a Lipid Research Clinic Prevalence Study, which turned into Rancho Bernardo. When I saw the initial analysis plan, which was to measure cholesterol and triglycerides in fasting subjects and ask subjects about 15 questions, I thought this is too much work to do just that. So I asked the NIH people if I could add diabetes, because the only thing I had read about heart disease epidemiology was in a textbook. I do not remember the textbook, but it had a list of the five proven risk factors for heart disease, age, sex, cholesterol, blood pressure, smoking and diabetes. I looked in the book about anything else on diabetes and it said absolutely nothing. So I thought, this one’s for me and I’d always been interested in diabetes because that’s a problem you see a lot of when you do infectious disease. (7)

On that same little list of five or six things, it said, “age, sex.” And I figured I couldn’t do anything about either one of those, but I was really curious about why there was this huge sex difference in CVD. And being a woman and there weren’t too many women epidemiologists or anything else in those days, I just thought that was really interesting and I guess in my own selfish way I thought if I could figure out why women lived longer, men would live longer too (and there’d still be some around when I got to be 85). So that’s how I got interested in it. And I started thinking about all the data that I now show repeatedly ad nauseum about the universality of the sex differences and it just seemed like it had to be sex hormones. We measured sex hormones to test that hypothesis and were never able to show it. And no one else has been able to do that either yet. (8)

Joe Stokes took me, I think, because I had a reputation for being a good lecturer. I certainly hadn’t published much. I remember quite well that the chair of medicine at the time, a fairly well-known cardiologist, said to me that I would never get tenure because all I did was count people. But they would take me because I was a good teacher. And I remember saying in my uppity way, “I was planning to count them very well.” Although I had no idea what I was talking about. (11)

The LRC Prevalence Study

It was just lipids. They were going for the Intervention Study and I didn’t know that. So I wasn’t smart enough to know that that’s why they ( Levy and Rifkin and Co.) were doing it. And how I managed to not know that must have meant that I was really out of the loop. But, at any rate, there was this visit 1 that everybody had almost nothing. Then there was visit 2, which was going to be 15% of the population at the high part of the lipid distribution that would have more stuff. I remember at this planning meeting, I don’t remember where it was, that all of us investigators, the epidemiologists, and whoever else went to those meetings, sitting around and Stoney pointing out to them that perhaps it would be a good idea to have a random sample to compare the 10% of the top lipids with. And that was an idea that took some selling.

So you can see that the people who designed the study, who were really good lipidologists, who had contributed pipetting and laboratory methods, just did not have a clue about these kinds of studies. And that was very interesting for me to see, because although I hadn’t had much training either, it was very easy for me to see what was wrong with the whole protocol. [ed. I still remember Stoney’s complaint that “The LRC was a study designed to demonstrate that 5% of the population had cholesterols above the 95th percentile.” (Darwin Labarthe)]

No, I think it was simpler than that. I think this was a study to demonstrate that almost any fool could send blood to the laboratory at the NIH. And I don’t think they were really interested in much of anything else. I mean, I never knew, and I must admit that I had a reputation with the people from Chapel Hill, who were the coordinating center, as being extremely difficult because I kept making all these questions, like Stoney’s dumb question and asking about why I couldn’t do diabetes and wanting to measure other stuff. I just didn’t know how to present my case because I knew what I wanted to do and I knew what they were doing wasn’t enough to be worth all the time they were taking. But I really wasn’t very sophisticated so I couldn’t present it very well. I’m sure I was abrasive.. (13)

[In LRC] I added on stuff they told me that I couldn’t add on and because I followed the whole cohort to some extent, not just the 15% random sample of hyperlipidemics and the 15% of Stoney’s sample, which is what I used to call it. So we had the whole cohort. You know, good epidemiology is opportunistic and I’m good at that. (30)

Trials

As someone else, and I can’t remember, but another epidemiologist told me, that a trial was a lot of work to get a single answer. Many people who are very famous trialists just do trials and they do not innovate. They get somebody else’s ideas and they say, “OK, now it is time to prove it.” …And I thought it was much more fun to get the ideas that somebody might to a trial about. (21)

NHLBI policy

I must say that my experience with the various Institutes has been interesting. And I think it depends on the enthusiasm of the senior person for epidemiology. So my best experience, and I’ve gotten money from many different Institutes because I’m interested in all the vicissitudes of aging now, have been with NIDDK. Because they had Maureen Harris who was a senior person whose main goal in life was to make sure that all of us got to do diabetes epidemiology who wanted to. She never took any credit for it, she never put her name on the papers; she just toiled in the trenches to make it important and interesting for us. And that made it really pleasant to work with her.

My experience with other Institutes has not always been so positive. And I think the Heart Institute has not really been as supportive of observational epidemiology as I would have thought they might have been. Maybe I don’t know everything that’s happening, but I know I was thrown off at least one committee because I thought they were making very dogmatic decisions that I just refused to accept…

I think everybody has their plusses and minuses. I have finally gotten old enough to believe that. I think that there was a lot of rigidity about who decides what the priorities are for an Institute. Where they should put their money, what things should be funded. And I’ve had nasty experiences. I’ve had at least two small grants turned down, which turned out to be somebody else’s paper. I’ve had one grant application which turned out to be an extremely famous diabetologist’s whole background for his paper : my grant, including the typo in his references. (28)

Belief Systems

There’s a quote I really am very fond of that I discovered when I was preparing for the Pickering Lecture and it’s a quote from Pickering. Pickering gave this talk, God knows when, but even before I was born, in which he said that, “Clinicians have to believe that they know what’s true or they couldn’t treat anybody. Scientists have to believe that they are not sure or they couldn’t still be scientists.”

And I love that quote. It’s really the way I feel. So as a clinician I try to think that most of the time I knew what I was doing. I was doing the right thing. But I see a lot of otherwise very good epidemiologists and I think it’s been very harmful to our credibility that if they find it, they believe it’s true. That’s a difficult dilemma. Because if I find it I always figure, “Now what have I done wrong? How did this come about?”(26)

…one of the problems with a cohort study is that the interval between funding, to stay funded, is too short to see anything about change and new events. (27)

Epidemiology is #3. We Try Harder!

We had a recent dean who called me into the office very shortly after he had arrived. A new event, not something that happens regularly. People don’t just leap to talk to their epidemiologists when they become dean of a medical school. I sat down all excited because I thought he was going to be interested in epidemiology and he said to me, “I understand you have a DNA database.” I said, “Yes” and that was the end of the conversation. That was it. That was all he wanted to know.

It’s a curious thing. I think epidemiology belongs in medical schools as well as in schools of public health because that’s how we attract different kinds of people to it. And I’m very pleased that I was in a medical school and that I’ve made a very successful division out of nothing. But I feel very frustrated because we never get anything. When space is being passed out, when goodies are being passed out, when FTEs are being passed out – the school got God knows how many new cardiologists when they didn’t have any patients – did anybody say, “Do you need another epidemiologist? Do you need another FTE?” They’re finally giving us some statisticians because all the other departments want them. But that’s it. (36)

On being asked what will be important down the line for epidemiological pursuit

I think understanding endocrinology [is for the future], which I thought in my naiveté would explain the sex differences. And so far, I haven’t been able to pull out the right hormone. But I think that diabetes is part of that sex hormone difference. Because diabetes pretty much wipes it out, [sex differences in CVD]. So it’s kind of nice that all the things that I was interested in seemed to be related to each other even though I have not been able to solve them. I think other people will. At our meeting this year here, there were several people who mainly addressed it in what they call “reproductive epidemiology.” But it’s the same idea.

What is it about being a female that’s good for you and what is it about diabetes that eradicates that? I think it’s part of the same package and surely there’s got to be something related to hormones, even if it’s not the hormones per se. It’s something in that. So I think that’s the most interesting question right now. And I hope that that idea has rubbed off on some of the people that work with me.(41)

. . .what’s most fun about epidemiology is what to put in the model and what not to. And I’m not just talking about statistical models. I’m talking about a mental model. (42)