Richard Remington

Quotes

But Tommy [Francis] even used the phrase ‘arterial pox’ to describe coronary heart disease or arteriosclerosis. He never went so far as to say that coronary atherosclerosis was a virally-induced disease. I don’t know whether he ever believed it. He never quite said it. But he said the characteristics of this disease reminded him of a number of infectious processes. He talked about ‘kindreds’ and he wanted to build environmental exposure and familiality into a community-based study and they selected Tecumseh. (9)

Surgical trials

I came to Houston first in 1960. I got a call one evening from Dr. Debakey who was talking about a clinical trial of surgical treatment for carotid lesions. That then started a 10 to 15 year long piece of research, a very large-scale clinical trial. I came down and helped with the design of that. We did the randomization for that trial out of my office in Ann Arbor. Primitive methods in those days. This was in the earliest days of computing. My first computer program was written in about ’57 for the IBM 650. We had a weird computer set up down here for that EAO (Extracranial Arterial Occlusions) Study. Incidentally, I think that was the first U.S. study involving surgical intervention and randomization in this country. Now, you’ve got to be a little careful [about history]. There were some VA trials of surgical intervention in tuberculosis that I think had some elements of randomization in them. And I think some of the surgical adjuvant trials in the cancer field may have been. But certainly, in terms of outside the government, outside the VA, in the community in general, this was the first cooperative trial of any sort that involved randomization. [But] Debakey never randomized his patients. We had two groups of centers, ones which did and ones which didn’t. But he has always had a large referral practice down here. And its hard for me to imagine how you’re going to tell the doctor in Fargo or the patient from Fargo, who’s referred for surgery down here, “We drew the envelope and it came out… (13)

But the business of clinical trial design, Bradford Hill’s paper, “The Clinical Trial”… His 1950 [classic]… was the first statement of randomization, replication, control, double blindness, you know… and all of that business. Bradford Hill was a good friend of Felix Moore and Felix was quite an Anglophile and had been to Britain quite a bit. But then Bradford Hill visited us in Ann Arbor a number of times. I remember a seminar I gave presenting the extracranial arterial occlusion study to him. He thought it was a terrible thing that there would be centers that did not randomize. And I couldn’t in any way, I couldn’t get him to be sympathetic with the problem of referral practices or any of that. You know, in a certain sense he was right. Our conclusion was we’d go with that half of that particular loaf and ended up with two sides to the study. And it was a useful study, and I think the evidence of that, again, is that there are now two trials in that field. The field lay fallow after our results. They added an aspirin component and did some more things later on. (14)

Pickering

I was very much affected by George Pickering’s work. High Blood Pressure, that book, I considered to be the source in the field of hypertension. I was quite familiar with the literature in the late ‘50s and so on. And it was unlike anything I had read before. It was comprehensive, open, broad, not that Sir George was necessarily broad-minded. But he had a willingness to bring in genetics and statistics and demography and physiology and everything into his concern for the thing… I just thought it was the most sensible thing I had ever- read about hypertension and much better than the other things in the literature – balanced, broader. And, of course, it was the comprehensive book in the field for the next 15 years or so. (18)

Makarska Conference: Ten days of cardiovascular epidemiology began in exactly 1968 at a famous meeting in Makarska, Yugoslavia. That was a three-ring circus, a meeting of the Council of the now ISFC. There was also an international meeting on intervention trials in the cardiovascular disease area and the British were there. Igor Glasunov made a presentation about some of the Kaunus Studies. The Americans were there… And then we started the 10-days seminar there too. Geoffrey Rose was to come and teach epidemiology. He didn’t get there in time for probably very good reasons and Henry Blackburn was forced to do the basic teaching of epidemiology to begin with and he was a nervous wreck about that. We were improvising. I mean, we didn’t know what we were going to do next, any of us.

I thought it was great fun, you know, we had a few students in there – fellows – and we didn’t even know whether in the next hour we were going to be dealing with the course (I mean a little bit of exaggeration) but it might be the course, it might be the meeting on trials, might be an executive committee meeting. And all these people. Ancel, of course, presiding over the whole thing, at dinner each night and the rest of it, as the chief guru of the whole operation. They, of course, had a Seven Country Study survey stand right there in the Makarska area. That’s why that place was chosen for that meeting because their Yugoslavian unit was just down the road a little bit between Makarska south toward Dubrovnik. But Henry was doing that, too, and Geoffrey finally arrived and took over the teaching. (20)

Jumbo and MRFIT

I’ve always thought that as epidemiologists we have some kind of an obligation to try to find out what we can learn from any of these trials. I’ve never been in a trial, however single factor it was, however pure its initial design was, that didn’t look differently from the far end than it did at the beginning. With the passage of time of time alone you learn things, I wish I had included this or whatever. You’re always, always going to go back and do some sort of post-hoc analysis on the thing to find… ‘I wonder if such and such’, subgroup analyses or whatever you’d want to call them. I thought the key question was to find out whether we could begin to kick the epidemic of coronary heart disease hard. I thought we’ve got Framingham and Tecumseh and Albany now and all these things coming along. Who wants to do more of that? A lot of people like to do that and you can learn some things. But isn’t it time to take what they’ve told us and put it into a multi-factor framework?

Anyway, Stamler, Henry Taylor and I came back [from Makarska] with this resolve that the U.S. simply could not trail the world in the next step. The next step, we all agreed, was to put it all together and find out if we could reduce by anything we knew – which was throwing all information together and taking some reasonably representative group of people, could be high risk, but whatever, some group of people, and find out if we could do anything to improve their experience with cardiovascular disease and in particular, arteriosclerotic heart disease. Could we reduce mortality and morbidity from coronary heart disease by putting together everything we knew? In other words, that was a multiple-risk factor approach.

We decided that we ought to be able to it better than the trials that were presented at Yugoslavia. We thought we had the technology. We were a little critical of some of the methods that they were talking about – observational methods. They were pairing communities and things like that, comparing work units. Geoffrey [Rose] within a year was doing his factory study in Britain, remember, where he used factories as the sampling unit. I thought from the beginning, why not just go after it and use individuals as the sampling unit and randomize and do this thing “right”. We’d been randomizing people, Henry Taylor and I and Buskirk and Jerry had been involved in that too on the physical activity question. We knew we could do it. We knew we could activate these folks. (28)

I was passionate about that [the impact of treatment on total mortality] and really got to the point that I was impatient with the basic scientists. I thought they were holding us back. I thought that the mechanistic approach to this was the enemy of the public health approach and that it was high time for the community interventionist-types to have their day in the sun. Find out what the hell we can do. (30)

HDFP

And they had the finding, and they had a good finding [in the VA hypertension trials]. Few events, but a huge result. It just seemed to me that that didn’t have much to do with what I was interested in about high blood pressure as a public health problem. I didn’t think it did. They were all veterans, they were all in VA, put to bed. It looked to me like a sort of an internal medicine textbook kind of a case rather than anything I thought was going to tell us much about the epidemiology of hypertension or certainly its control. Or anything that I was interested in terms of, you know, “a half, of a half, of a half,” or anything that mattered much. Or trying to make an impact on the vital statistics. (33)

MRFIT

I think you’re best advised to take them [trials] as a blunt instrument. I never had any psychological problem with the multiple risk factor approach. An important question is, ‘can we do anything about this disease?’ Give it our best shot – can we do anything about it. And the next questions will be, ‘how did it happen?’ That seems like a logical progression to me. (34)

Middle-class and medicalization of prevention: There are those who say the whole disease prevention/health promotion business, in fact the Dean of the school at Minnesota said it last week. That whole thrust is very middle class… And we’re not dealing with middle class diseases here. We’re dealing with pandemics. Everything we’ve talked about so far. But our interventions, if you think about them, are almost uniquely middle class type interventions. We’re talking about prescription drugs, people who are under consistent medical care of a reasonable sort. We’re talking about increased physical activity and we’re talking about the ability to make substantial changes in their diets for economic and other reasons. It all has that ring about it… [based upon the belief that ‘I can control my future, that I can do something.’ Yes, that’s a middle class value. It’s a middle class attribute. But yet with hypertension, that’s not a middle class disease. That’s Blacks, people who must cure their food with salt because they can’t afford refrigeration. So if it’s hypertension, it’s stroke. And if you talk about coronary heart disease in our society, it’s the same sort of business. Fat is cheap. You can buy fatback. You buy parts of the meat that other people throw away. …What we decided to do, we called it ‘ecological’. We said we wanted to take a disease… We defined a new kind of trial and I called it an ‘ecological intervention trial’. In which we take a disease process in its community setting, we don’t exclude very many people, we take everybody who’s got this thing. They are all in our trial. And boy, we were very careful about excluding anybody because we want to know what the effect will be on mass measures of disease, vital statistics for example. So we are going to take everybody. But… we’re going to include randomization so that we get simultaneous control, so we can get internal comparison. That’s the one thing I would not want to give up. That’s the last thing to go as far as I’m concerned. I would give up a lot… You’ve been screwed repeatedly in the literature by these false comparisons and not knowing what’s happening. (36)

Tertiary Trials

I’m really think that we… have left important results on the cutting room floor. And I think maybe that’s less a reflection of statistics or epidemiology as it is a reflection of the tertiary type, if you can get the reference I’m making. The tertiary-type internist. If this person doesn’t really have this thing, in the differential diagnosis sense, if this person doesn’t have this and doesn’t have only this, should he be in the trial? That’s not the way life is out there. So MRFIT’s control group event rate was half or less than the design characteristic of the study. You go halving in the event rate in the comparison group, you’ve got a problem.

…There’s no doubt about what was going to happen. You look at the MRC data. Geez, these people are the healthiest hypertensives in the world. You’ve taken away any suggestion of any kind of a target organ involvement at all. I don’t know what that group is. I don’t even recognize that disease. That might be a disease in laboratory rats, but its not a disease in free-living people, except this one over here and that one over there, and so on. Who wants to do that?

[ed. Not epidemiologists. But again, there’s sort of an instance here it seems to me, where an Institute decision, I’m not sure how that was arrived at… an institute decision to go a particular route prevailed over investigator-instigated data saying, “Hey, look, it is crazy to go through all these exclusions because you’re going to be cutting the death rate X %”] (38)

Trials for the day

I think we make a mistake in teaching statistics and a certain brand of epidemiology. Epidemiologists are better at it than others, statistics and other basic sciences, I might say. Maybe in a certain sense we’re not Bayesian enough. Maybe we ought to take more account, not formally, but more account of the epoch of research we’re in when we do a trial.

If we are in the earliest stages of research, beyond the pilot, beyond plodding around…. It seems to me that if you’re going to develop a new hypertensive drug, for example, then you want everything, don’t you in the design. If you’re in the first stage, does this thing work, is it efficacious, and so on, you’re going to want a very tight design. You’re going to want to know to whom it’s applied, you’re going to want an even playing field for the controls, and you’ll probably want placebo and double blind and all of that stuff.

But as time goes on and you’re expanding and what you’re interested in is something about to how broad a base might this apply, or you’ve accumulated a lot of information and you’re now about to take it out for mass advice, health education advice and so on. Then it seems to me that the rules are probably kind of fundamentally different. I kind of think they are. Maybe we do a disservice by just sort of acting as if double blindness and randomization and control and exclusions and purity of the case group, and all this kind of stuff is a universal.

[ed. I wonder if the extent to which this happens is because we don’t distinguish or don’t define the question we’re trying to answer clearly enough.]

That’s another way to get at it. That’s exactly right. In HDFP… we said we wanted to do both things. I always hedged on that one. (40)

Death of “Jumbo,” birth of MRFIT, end of “The Pause”

There was even some talk of suing. I was sort of a semi-firebrand like Jerry at that stage. “They can’t do this to us.” You know, they didn’t review this. They’re not paying this, but they didn’t take it through complete review. That’s unfair. It’s an equity issue. “If you folks are going to have these rules, then you’ve got to apply them. You can’t decide as a managerial decision according to your own rules that you’re not going to do this because you don’t like what it does to your budget. Otherwise you’re going to have to change what you…”

I think we had them. I still think we had a point. The only problem was it wasn’t enough of a point to sue. I don’t know why we didn’t do that. I guess we decided, and Jerry and I talked at length on the telephone and in meetings and so on. It was always Henry Taylor, Jerry Stamler and Dick Remington. We were the three that were really fired. There were other people who came along, Milt Nichamen was involved in it. Lots of folks were involved in that in one way or another. But it was the three of us that were kind of an executive committee for putting that together and talking it through.

I guess we decided that although we had lost that battle, perhaps we won the war. Because an NHLI – I don’t know how many letters to put in the damn thing, but at any rate, the Heart Institute had moved from having no interest in intervention trials to having all kinds of interest. Within a year or two, my god, they were all over the place with these big trials in intervention.

At least I felt ‘let’s see what happens on that. We’re going to be around. We’re going to be in that business. Let’s just see what goes ahead.’ And, indeed, that’s how it worked out. We were all involved in all three or four of those trials in one way or another. Either as a reviewer, my own involvement as coordinating center director for that thing, and I was in and out of MRFIT at various points. In fact I was in on the kill. I don’t know why I get in on these things. I was part of that committee [to terminate MRFIT].

…Various people at various times were unhappy, as you know well, with the coordinating center of MRFIT. So they had enforced site visits from the coordinating center people to our coordinating center. Blackburn said to us, “You’ve got to tell these people how a coordinating center should be run.”

That should never see the light of day. How that contract ever got to be let and so forth. That was intensely interesting. You’re talking about direct institute intervention…

Public Health

We’re not very good at that as a country, in putting [public health] initiatives in place and keeping them in place. We’re very good at putting out fires and identifying fires and dealing with them, or relatively good. But somewhere along the line we seem to have lost a little bit in this country, a respect for institution building or institution maintenance. It’s not very glamorous work. (50)

All the public health problems of the past, most of them, are still with us. We eradicated small pox. But you can count the conditions that we’ve eradicated on the fingers of one finger. That means that in a peculiar sense you don’t really solve public health problems. Because almost always, and there are some exceptions, but almost always you have a maintenance problem, even if you’ve solved the acute phase. (50)

You know the thing I like about our field is its methodologic breadth and depth. Because as a methodologist-type or at least that’s what I’ve always thought of myself as being, it has really struck me that there are probably few fields in the health area in which one can use such a broad array, and appropriately use such a broad array of methodology. (57)

Lifestyle interventions: I’m not sure we have been sufficiently reductionist in our approach to lifestyle modification in general. If I were going to criticize us, it would be that what we’ve tended to do in this kind of research is to come up with an intervention, sort of a macro, sort of a packaged intervention. Then either apply that or some quasi-zero-level control to see how it worked.

Whereas, every one of those lifestyle modifications, it seems to me, or nearly every one, can be dissected into lots of little things. Lots of little methods for motivating people, which we could, perhaps, have studied much more systematically.

The drive is to get something tailored and if it works, use it, get it out there. I’m all for that, I’ve a conflict in that. (65)

Why HDFP worked: I think it was a question of readiness. I think we had everything going for us, the pharmacologic technology was ready, as we talked about. We could put together a step-care regime. There was general awareness that hypertension was no bloody good for you. We had the VA trial. So in a certain sense, HDFP was, as many of these things are, the world was ready for it. It was the trial that, perhaps, needed to be done. Something like it to broaden the result of the VA trial, to extend the blood pressure range and extend the population. In that sense, It wasn’t radical in any way. But its impact, I think, has been radical. I don’t think there’s any question about that. Its success, I’m bound to think that it was a reasonably well designed trial, reasonably well executed, and kind of at the right place and time in development.

Sometimes we push things a little more rapidly than they should be pushed. That was a case in which we didn’t. It was a manageable chunk of knowledge and information and of programming to handle. So you can get meta messages out of all that – methodologic. Going back to this notion that these cooperative trials are such blunt instruments and so massive that you want to kind of… They fit best if the situation is really ready for them, I guess. (69)