Life and Times of Leading Cardiologists: Rory Collins

E. Magnus Ohman, MD: Hello. I'm Magnus Ohman, and welcome to another edition of the Life and Times of Leading Cardiologists. I'm very fortunate today to have a leader of cardiovascular medicine, Professor Rory Collins from the University of Oxford. He has been behind the science we use every day in our practice: reperfusion therapy, thrombolysis, aspirin, blood pressure control, long-term use of aspirin, and statins—really, all the facets of medicine that we use now. Welcome.

Rory E. Collins MBBS, MSc: Thanks a lot, Magnus.

Knighthood

Dr Ohman: Rob Califf started this series some years ago, but you are the first person we have had on who has been knighted by the Queen. You are supposed to be called Sir Rory, but can I just call you Rory?

Dr Collins: Yes, I think so. It's an American program.

Dr Ohman: We recognize that, but tell me more. How did knighting come about? It does not happen every day.

Dr Collins: I received it for services to science, but really, it was for digging the funders out of a hole. They had funded a concept of establishing a very large, prospective cohort of half a million people in Britain, the UK Biobank Cohort, but they did not know how to make it happen. We had experience in Oxford of running very large-scale randomized trials in different parts of the world, and we used a lot of those methods to help with the design and the running of UK Biobank.

We were able to recruit half a million people within the budget; that involved asking them lots of questions, making lots of measurements, collecting a lot of biological samples, and then linking them into the very rich healthcare data that were available through the National Health Service. That resource is available to any researcher anywhere to use for any kind of health-related research. It's now being used by hundreds, if not thousands, of different researchers throughout the world to study a whole range of different diseases. This is not my main area of work, in terms of cardiovascular disease, but it's been an absolutely fascinating journey.

Dr Ohman: In the United States, we just get a thank you, so I think it's pretty cool. I'll call you Rory. Rory, where did you grow up?

From Hong Kong to England

A young Collins with his parents in Hong Kong. Image courtesy of Professor Collins

Dr Collins: I was born in Hong Kong. My father had been trained as an electrical engineer during the war, and the electricity did not work in Hong Kong at the end of the war. He was on the first supply ship they got in, and he turned the lights on in Hong Kong, so they offered him a job.

Dr Ohman: Problem-solving is a family trait. You solved the United Kingdom's Biobank problem on finances, and your father solved electricity for Hong Kong. That is a great analogy. Hong Kong back in those days was very different to what it is now, right?

Dr Collins: I was there until I was about 9, and then, as was typical of that time, I went back to England for boarding school and commuted back once a year to see my family.

Dr Ohman: That's so unusual. We are not used to people being so far away from their parents. How did that feel?

Dr Collins: It was quite separate. We did not have much in the way of family in the United Kingdom. I used to live on a farm near the school I was at, and when you are at school, the contact was through the blue aerogram. People won't remember, but you would write and send one to your mother and father every week, and every week you would get one back. Contact was very different from today.

Dr Ohman: That is amazing. Where in the United Kingdom was this?

Dr Collins: I went to school in Kent, which is south of London, and then I went to a school in London called Dulwich College. My father grew up as a child in south London.

Dr Ohman: Did you have brothers and sisters that had the same experience as you did?

Dr Collins: I'm in a family with only two children. My sister is 11 years older than me, and she went back to England probably when I was about 2. We did not really overlap. She trained in Oxford as an orthopedic nurse. We got to know each other later in life.

Dr Ohman: You ended up in the same place, but when you grew up you were quite far apart. Did your parents stay in Hong Kong or come back to the United Kingdom?

Dr Collins: They eventually came back to the United Kingdom and retired there.

Taking a 'Calculated' Risk and Going to Medical School

Dr Ohman: When you were in school in London, what made you go the route of medicine? Or did you have another route?

Dr Collins: I was very interested in mathematics, but I cannot say that I was a great student. Suddenly, at 15-16 years old, I had a teacher in mathematics who made it make sense to me. It was a revelation. I was very keen on doing something in mathematics, and I went to a career advisor. They said with mathematics, you could become an actuary. They explained what an actuary was, and I thought that sounded incredibly boring.

Dr Ohman: You mean like an accountant?

Dr Collins: Someone who predicts for pensions how much you should charge people on the basis of their risk.

Dr Ohman: In the end, you ended up with life-expanding therapies. You were headed in the right direction.

Dr Collins: Yes, maybe they were right. That sounded boring, and I was chatting with one of my friends, who said, "I'm going to go to medical school." I thought, "Okay, I could get into medical school." I applied, did my final exams at school, and went off hitchhiking around Europe. I came back to find that I got much better grades than I anticipated. I thought that maybe I should not do medicine, but go to Cambridge. They had a course called "natural sciences," where you could continue to do mathematics and at a later stage during that degree, you could then turn it into a medical degree.

Dr Ohman: Wow, how unusual is that? I have to get back to your teacher. What was the name of the math teacher who was influential?

Dr Collins: It was Mr Earle.

Dr Ohman: Did Mr Earle teach you any statistics that became prominent in the rest of your life?

Dr Collins: No, I had never done statistics. I failed to get into Cambridge. It raised in my mind the idea that I would like to get back to mathematics. I went to medical school at St Thomas' Hospital in London, which is right across the Thames from the House of Parliament.

Dr Ohman: You looked at Big Ben every day?

Dr Collins: Yes, I looked at Big Ben.

Weighing the Odds and Studying in America

Dr Collins: After my second year in medical school in Britain, I had the option of doing a third year and getting a bachelor's degree, or going straight into clinical medicine. It's a 5-year course or a 6-year course. I thought, "This is my opportunity; why don't I do some more mathematics?" It was at that point I thought about statistics. I'd never done any, but I thought it was relevant.

I spoke to the authorities at the medical school, and they said, "What does that have to do with medicine?" I went around to all of the colleges in London, where people could go and do what we called an intercalated bachelor's degree in the third year. You could do it in anatomy, physiology, biochemistry, psychology, or the history of medicine. No one had ever wanted to do it in statistics.

In fact, the medical school was not very happy at all and refused to support me for the funding that you could get from the government for this third year because it had nothing to do with medicine. At that point, I thought, "What am I going to do?"

At the end of my school, my last report said, "He did better this term, but he would be well-advised to get rid of his antiestablishment views." I found that that is actually not good advice. I think that was what drove me to go to the library and find the addresses of three dozen universities in America. I vaguely remembered hearing that America was more flexible. I wrote to them all, and I said, "I'd like to do a third year and get a bachelor's degree from you in statistics."

Dr Ohman: It is very different that halfway through medical school you would go to another country and do statistics.

Dr Collins: I could not do it in London.

Dr Ohman: You weren't saying no, so that's good.

Dr Collins: Several replied and said, "No, that's not possible." Harvard replied and said, "Send us a check and we'll think about it." George Washington University wrote back and said, "That's a great idea." I have to tell you, I picked these names with a pin. Some of them I'd heard of, some of them I hadn't. George Washington said I could do it. The head of community medicine or public health in St Thomas' heard about this and said, "This is great. I know the head of statistics there. Do you have any problems with this?" I said, "Yes, I've got no money. The medical school is not supporting me to get a grant." He said, "I'll deal with that."

He wrote to Jerry Cornfield, a fantastically famous biomedical statistician who set up cancer statistics within the National Institutes of Health (NIH), who was heading up the biostatistics department. They gave me a job in the biostatics center working with a guy called John Lachin on a trial of chenodeoxycholic acid for dissolving gallstones, funded by the NIH. I worked there one day a week while I was doing my bachelor's degree in statistics at George Washington University, and that was my introduction to randomized trials.

Moving to Washington, DC: From No House to the White House

Dr Ohman: This is very unique. You had to move to the United States. How was it in Washington, DC? What year is this?

Dr Collins: This was 1976. I arrived in Washington, DC, and because all of this was quite late on, no accommodation was available for me. In the United Kingdom, my mother had met the wife of a pastor from Washington, DC, at the hairdresser's, who said they had some accommodations available. It was organized that I would stay there. I arrived in my tweed suit in downtown Washington, DC, and dragged my suitcase to the Methodist ministry accommodation—which was, I think, a halfway house for people who had recently got out of prison. I stayed there a few days until I managed to find somewhere else to live.

The first week, I went along and tried to learn about the university. I went to a foreign student orientation class and met the son of the Bishop of Iran, who had just finished his degree at Oxford. He had no money, but had a lot of contacts and at that time was staying with the family of peanut farmer. They were friends of a famous peanut farmer who had just been elected president.

Dr Ohman: That would be Carter, right?

Dr Collins: That would be Jimmy Carter. I ended up going to one of Jimmy Carter's inaugural balls after I had just arrived in Washington.

Dr Ohman: In your tweed suit?

Dr Collins: Yes, in my tweed suit.

Dr Ohman: When you look back on this, what was the key thing that you learned? You had not finished medical school yet.

Rory Collins (right) and Bahram Dehqani-Tafti (left) circa 1978. Image courtesy of Professor Collins

Dr Collins: I learned that you could do things. I mention my friend Bahram Dehqani-Tafti because he was someone who gave me a belief that if you wanted to do things, you could do them. That was his approach to life. Sadly, he was assassinated because his father was the Bishop of Iran when he went back to try to and help him. He made me believe that you could do things if you wanted to do them.

Completing Medical School and Getting a Master's as Well

Dr Ohman: It's fascinating. You finished and went back to St. Thomas'. How did the next 2 years of medical school feel? Boring?

Dr Collins: I went back to St Thomas', but I had gotten the bug. While I was in Washington, I had applied and been accepted for the master's program in statistics at the London School of Hygiene and Tropical Medicine. I enrolled in that while continuing to do medicine until they found out. You were not meant to be doing two degrees simultaneously, but I did not know that. They allowed me to sit in on the lectures at the London School of Hygiene and Tropical Medicine doing my master's.

Dr Ohman: Just to give relevance, was this with Stuart Pocock, who was so behind of all the statistics in their clinical trials, or was this with someone different?

Dr Collins: This was applied statistics. It was not Stuart. One person I met who was doing the course was Garret FitzGerald from the University of Pennsylvania. He was a research registrar in London at the time and was learning some statistics.

Dr Ohman: There were two of you.

Dr Collins: I had come back to meet up with him many years later.

Dr Ohman: Of course, he got hooked on aspirin, which we'll come back to in a minute, but it's just a fascinating coincidence. Then you finished medical school?

Dr Collins: I did.

Early Medical Career

Dr Ohman: What did you do then? You had a lot of statistical knowledge when most really do not. How did you apply it?

Dr Collins: I finished medical school and did my first intern job as a surgical resident in London. I had done a lot of jobs as a student, but this was certainly the most tedious. After 3 months, I thought, "If this is medicine, I want something different." If I had done my medical residency first, I do not think I would have concluded the same. The next 6 months were absolutely fascinating, and I really got hooked on learning how to treat individual patients. But that was not the way things panned out.

I thought, "Well, what should I do?" I could go back to university and do a degree in English, which might be quite fun, or maybe I could follow up on this statistics stuff. I read a book by Peter Armitage called Statistical Methods in Medical Research. I rang up the London School of Hygiene [and Tropical Medicine] and found out that he was no longer there; he was now in Oxford. I then got his number in Oxford. I rang him up and arranged to meet him and speak with him. I explained my background, and he said that it really was not his field, but with the kinds of things I was interested in (applying it to clinical trials), that I should talk to Richard Peto, in Oxford.

He rang up Richard Peto, who said he was too busy. He said I could come around and make an appointment to speak with him. I walked around and went into this old hospital up these stairs. I was looking around for his secretary, and there was this long dusty corridor with offices and a guy in jeans and a T-shirt who was fixing a chair. He said, "What are you looking for?" I said, "I'm looking for Richard Peto's secretary." He said, "I'm Richard Peto, and my secretary is in there. Anyway, come on in." We sat and started talking, and after about an hour, Richard Doll, who of course is the famous epidemiologist who found a link between smoking and cancer, came in. He looked at me and he looked at Richard and he was incandescent with rage. Richard was meant to be finalizing their report to Congress on the cause of cancer, which was massively late, and here he was talking to someone who had literally come off the street. I was packed away with a paper to read, and Richard said, "I'll come back, and we'll go and have lunch."

We went and had lunch, and then he said, "Do you play bridge?" I said yes. Two other people sat down, and they started dealing cards. We played bridge, and I did quite well. I don't play bridge that well, but I got a small slam, which is quite unusual, and he offered me a job. I finished my clinical residency and came back to work doing research for a year to learn about clinical trials, which was my plan.

Origin and Execution of ISIS-2

Dr Ohman: You then ended up working with Salim Yusuf and Peter Sleight in designing the ISIS-2 trial^[1] and putting aspirin and streptokinase on the map. It was a very powerful study. You are now so focused in the statistics realm, what did you see next as your step beyond this? All the work you had done since that time had also been very quantitative and was dealing with very important issues.

Dr Collins: I think often when one looks at people's careers, it looks like there's a plan. I suppose what I've been trying to describe is just the randomness, not just in the evaluation of treatments, but of the way things pan out.

ISIS Trials Steering Committee

I came to Oxford, and Richard introduced me to Peter Sleight and Salim Yusuf. Salim was just completing his clinical training, but they had this idea based on his PhD work around the use of intravenous beta-blockade in the treatment of acute myocardial infarction. They had gotten some support to do a large-scale study, but there was no one to do the hands-on work. It was my job to try to get this study going.

Salim and I spent a lot of time looking at what other treatments looked like. Richard had introduced the idea of combining results from old trials to see whether there were any clear signals when you looked at all of the data in meta-analysis. We looked at all of the acute myocardial infarction trials and the different agents. The one that popped out was streptokinase. There were two dozen small trials, none of them convincing, indicating that the drug could cause bleeding without clearly having an effect on mortality. When you put it together, it looked quite compelling. It made us all think that we would like to do a trial with intravenous streptokinase.

Dr Ohman: This trial was unique in that you created suitcases with assigned treatment numbers for kits that had either streptokinase or aspirin or placebo. How did you come up with this idea of suitcases? It was a black suitcase as I recall? It sat in the cardiac care unit somewhere, and there was telephone randomization, It was quite unique. Where did this thinking come from?

Dr Collins: It was a guerilla clinical trial. When we started the ISIS-1 trial,^[2] we were learning how to do what Richard had had as a concept—the need to do much larger randomized trials to pick up moderate effects. The question was, how do you do that?

The network expanded in ISIS-1, and then we were piloting intravenous streptokinase in ISIS-2 on the basis of the meta-analysis. Our Italian colleagues in the ISIS-1 trial, Gianni Tognoni and Aldo Maggioni, said, "We can't keep people interested in the ISIS-1 trial at the moment, but what we could do is take your idea, and do a trial of streptokinase in Italy."

They came up with some very nice, colorful protocols, and I thought, "That's a lovely idea." One is a magpie—one looks and sees things that look good in another area. When we came to ISIS-2, we did the same thing by making a colorful protocol. We had these drugs and wanted to make it easy for people.

We had two different combinations: streptokinase and aspirin. We thought about using some kind of little pack that tells them exactly what to do when they opened it, and had all the labels they needed for patient notes. We tried to think about things that might put people off. Have to write "This patient is in the trial, and this is what has happened"? Just stick the label in the notes. It was all about practical simplicity.

Statin Trials

Dr Ohman: We have totally gone away from that. Somewhere, we have to get back to that simpler approach.

We'll jump forward a little bit. You had been involved in blood pressure control, and you looked at combining trials for blood pressure control and outcomes. How did your work on cholesterol-lowering therapies come about? It became something dear to your heart, I think, as time went on?

Dr Collins: Again, it emanates from Richard. He had this idea of a meta-analysis to find out whether you were missing signals, or being misled by the focus on one particular result within one particular study. Cholesterol-lowering is perhaps the example that has been the most problematic, both in terms of anxieties about hazards (the idea that lowering cholesterol might produce hazards, based on one or another trial) and the failure to be able to detect risk reductions and cardiovascular risk reductions (because the trials themselves were too small, or the agents were not lowering cholesterol enough).

He had done what we would consider now to be a crude meta-analysis of all of the trials of the old drugs and diet. It looked compelling again that the more you would lower cholesterol, the more you would lower risk, and that there was not good evidence of hazard. The problem was, one did not have the tool to test that hypothesis in terms of an agent that lowered cholesterol enough. We had created the tool that could allow us to test it and demonstrated how meta-analysis could pick up modest effects, but we did not have the agent until statins came along. It was the combination of an intervention that did what it was meant to do in terms of lowering cholesterol a lot, and the mega-trial.

Making Sense of Big Data

Dr Ohman: Once again, it's the randomness of how things came about. Where do you go from here? We have big data now. Can we use big data to make sense of medicine and go beyond meta-analysis and combining of data sets? How do you feel about going beyond this point?

Dr Collins: I think that the randomized trial is one of the greatest inventions of modern times and that people fail to understand that. They fail to understand its beauty—the idea that through a random process, you end up with two groups of individuals that differ only randomly from each other, in terms of all the things you know about them but also all the things you don't know about them that might determine outcome. The failure to understand that is, in my view, part of the reason why there is a little bit of naiveté about the way in which big data could be used. I have seen claims saying that we do not need to do randomized trials anymore because big data and real-world data will be able to work out what treatments work.

Dr Ohman: I can confess, I hate the term "real-world" data. It suggests that there is a fake world somewhere else, which is totally wrong. Everything is real-world. That is to me a big challenge.

Dr Collins: Yes, it's horses for courses. Randomized trials provide some of the information that you need to determine whether a treatment is efficacious and safe. Population data can help you apply that to different settings, in terms of the risks for particular outcomes. Then, results of the randomized trials are applied to the risks to those outcomes in the population, to determine what the balance of benefit and risk would be of that treatment.

The idea that population studies, observational studies, can avoid the biases inherent in those methods, irrespective of how big the data are, is not correct. In fact, the bigger the data are, the more likely you are to get statistically precise wrong answers. Despite how clever the analysis or how advanced the computers, I think you cannot be sure that you have dealt with all the biases.

Look at Google. Google randomizes everything they do. Every time they make a change in their systems, they randomize it to evaluate whether it works.

Dr Ohman: When you sign your agreement to log on to Google, you are actually giving permission for that to happen. It's an interesting dynamic.

Advice for the Young, Numbers-Minded Doctor

Dr Ohman: Let's finish up. What would you recommend for a young doctor who had your interest in numbers? How does someone go forward? Go with your feeling and let the randomness take care of things?

Dr Collins: Medicine is now increasingly mathematical. Bioinformatics, the ability to understand numbers, is central to medicine in a way it really was not when I qualified in 1981. I think the future is yours.

Increasingly, we need to make sure that doctors have more mathematical understanding, irrespective of what kind of doctor they are. If you are a family practitioner, how do you determine truth from nontruth? Are you going to depend on people writing stuff on the Internet? Surely not. You need to give yourselves the skills to be able to judge things for yourself.

I think the same is true of a lab scientist. Lab scientists need to be able to use the data they get in the most efficient way.

Dr Ohman: You are the perfect example of someone who said that numbers are important, and then put a whole career together around these issues. I want to thank you for a great conversation about the importance of randomized trials and potential challenges with big data, which I'm glad you brought up. It's been great. Thank you, Rory.

Dr Collins: Thanks very much, Magnus.