Extra Feature: Calum MacRae Full Interview

Getting Personal: Omics of the Heart

Speaker 1: Hi everyone. As a quick introduction, this is the full length recording of Anwar Chahal's interview with Calum MacRae from August 2017. A portion of this interview was included in episode seven of the Circulation Cardiovascular Genetics podcast "Getting Personal: Omics of the Heart". As we couldn't fit everything into that regular podcast episode, we've released the unedited version as a special, feature-length podcast. Enjoy.

Dr Anwar Chahal: My name is Dr. Anwar Chahal. I'm a Cardiology Fellow in Training from London, U.K., and I'm doing my research fellowship here at the Mayo Clinic, and I'm very honored and delighted to have our guest, Dr. Calum MacRae. I searched for Dr. Calum MacRae's biography online and it came up with a Wikipedia page talking about somebody who's a rugby coach. So, Dr. MacRae, I hope that's not another one of strings to your bow, that's something else that you manage to squeeze in amongst everything else that you do in your busy and punishing schedule.

Dr Calum MacRae: I did play a little rugby in my day, but I haven't coached any, I can assure you.

Dr Anwar Chahal: So, you are the Chief of Cardiovascular Medicine, you are an MD, PhD by training, and you are Associate Professor at Harvard Medical School, and your expertise, amongst many other things, internal medicine, cardiovascular diseases, but in particular, inherited cardiovascular conditions. Is there anything else that you would add to that?

Dr Calum MacRae: No, I'm a big fan of generalism, and I am quite interested in cardiovascular involvement in systemic disease as well, but largely as a means of keeping myself abreast with the biological mechanisms in every system that seems to be relevant to cardiovascular disease.

Dr Anwar Chahal: So, that reminds me. Once I heard you talk, and you mentioned to all those people that were considering cardiovascular genetics the importance of phenotype and actually how people have become increasingly super-super-specialized, becoming the bundle branch block experts or the world's authority on the right coronary cusp of the aortic valve, and how things were now going full-circle as people actually need better and better, more general understanding so that we can accurately phenotype. And you once joked that you'd actually done residency three times, so you know the importance of having a good generalist base, so could you expand a little bit on that?

Dr Calum MacRae: Well, I have to tell you, it wasn't a joke. I did actually do residency three times. But, I think the most important element of that theme is that biological processes do not, unfortunately, obey the silos in which medical subspecialists operate. So it is increasingly important to have a broad-based vision of how phenotypes might actually impact the whole organism. That's particularly true because it helps us ratify disease, so that there are mechanistic insights that come from the different cell types and tissues and biological processes that are affected.

I think, in general, that is something that we've all appreciated, but as time goes by and people become more and more specialized, it's less regularly implemented in day to day clinical practice. And so, particularly as molecular medicine becomes more and more penetrant in clinical disease management, I think you're going to see a return toward some generalism. Obviously, procedural specialties are the exception in many ways in this setting, because you need concentrated procedural skill. But in general, particularly for translational scientists or scientists who are interested in the underlying mechanisms of disease I think, I see a general movement towards a degree of generalism.

Dr Anwar Chahal: Indeed, and in terms of, as you say, trying to understand those disease processes and trying to, let's say for example, make sense of the incredible amounts of information that can now be gathered with genomics and high throughput omics, you believe that it is actually more of a requirement to be able to understand that now that we can gather this high resolution and broad depth of data?

Dr Calum MacRae: Yes, I agree. I think one of the core elements of modern clinical medicine is that the phenotypes have, in the last 50 to 100 years, we've really focused more on improving the resolution of existing phenotypes than expanding the phenotypic space. To be completely frank, I think we've extracted a lot of the information content that we can from the phenotypic space that we've explored, and what we need to begin to do is to find ways to systematically expand that phenotypic space.

I think there are a lot of reasonable ways of doing it just by thinking about other subspecialties. So, for example, in cardiovascular disease, we've focused very heavily on anatomy and physiology, but we haven't really done much in the way of cell biology. Whereas, in immunology, partly because there's access to those cell types, it's possible to do much more detailed cellular phenotyping. In neuroscience, we're now doing functional MRI, and looking at individual subsets of cells in the brain, and their function in the context of particular challenges.

My general thesis would be that the type of strategy would serve us well and that there's also, I think, an important mismatch between the dimensionality of phenotyping that we currently undertake and the scale of the genome and epigenome, transcriptome, et cetera. So, it's not surprising that we can't be convoluted genome of 10 to the nine variants with a phenome that are present only really has about a 10 to the four phenotypes. And so, I think some systematic right-sizing of that balance will be necessary.

There are lots of things that we record that we don't even think of as phenotypes, and there are phenotypes that we record that we don't really think about how to optimize the information of content. And so that's one of the things that we have begun to invest time and energy in. And thanks to the support of the American Heart Association, Verily, and AstraZeneca, as part of the One Brave Idea, we have elected to fully focus on that area in particular in coronary disease. But I think it's a generalizable problem with much of modern medicine that we tend to have focus on phenotypes that, in many instances, date back to the turn of the last century rather than to modern molecular and cellular biology.

Dr Anwar Chahal: So, you beautifully brought us to the first question, which was to ask you about One Brave Idea. Could you just, for our listeners who aren't familiar with that, just give a little bit of a background on One Brave Idea, and you've already thanked the people who have funded that, but how did you actually reach the point where you thought that this is something that really, really needs to be done? What's the process of reaching that point of bringing this idea to fruition?

Dr Calum MacRae: I think we had recognized in many instances that the families that we were seeing in cardiovascular genetics clinics were much smaller, the diseases appeared to be less penetrant than the original families that we studied when we cloned many of the disease genes. This was work that I did as a post-doctoral fellow in John and Christine Simons lab many years ago.

One of the things that was pretty obvious was that there were subtle pre-clinically or sub-clinically affected individuals in almost every family. And that made me ... That implies that the average family is so different from the extreme family. Is it something to do with either the resolution with which we were assessing disease or are we actually just measuring the wrong elements of the underlying genetic trait? So that, for example, is a dilated cardiomyopathy family actually a family that is susceptible to dilated cardiomyopathy in the context of some unmeasured conditioning variable, maybe a viral infection or an exposure. And because we're not measuring the exposure, or we're not measuring the underlying diaphysis, we're only measuring the final state, so we only classify people as being affected if they actually have an extreme phenotype. Are we, therefore, missing the core elements of the biology?

As part of doing that, we began to look outside the heart for other phenotypes, and one of the things we recognized ... This was in cardiomyopathy ... Was that different cardiac phenotypes were really aggregates of much more granular, multi-system phenotypes. So there would be families who would have dilated cardiomyopathy, but they would also actually have abnormalities, for example, of the distal interruptus muscles, and no other muscle group in their entire body. And in fact, the distal interruptus muscle phenotype was much more obvious than any cardiomyopathic phenotype.

So you start to understand that either other extra cardiac or electrical phenotypes, or maybe even sometimes neurofunction phenotypes are more penitent features of some of these disorders, albeit rare disorders. And so that immediately leads you to think are most of the common traits that we look after really aggregates of things that really only share the relative frequency of the core phenotype, which often dates back to decades earlier when phenotyping was at a much more superficial level.

So that vicious cycle perpetuates itself if we never look more deeply or look outside the constraints of a particular subspecialty. And so we have begun many, probably almost four years ago, to build a sort of next generation phenotyping clinic where we tried to bring either cell biology or molecular biology from outside the heart into phenotyping patients in a cardiovascular clinic. That idea was in our DNA, that's probably not the right way to say it, but it's something that we had worked on in a cardiomyopathy setting.

Dr Anwar Chahal: Right.

Dr Calum MacRae: And so then when the RFP for One Brave Idea came out, it seemed like a natural expansion of that to try and think about how you could apply new phenotyping in current disease. One of the inferences from that line of thought is to move, essentially, beyond ideally much upstream of the shared final common pathway so that you can begin to identify discreet underlying mechanisms.

And then, given the success of cardiologists, and cardiology in general, in prevention, it became obvious that really what we wanted to do was to try and understand not just disease, but also wellness. And to do that in a way where we could potentially detect the transition from wellness to the very first stages of the disease or the diseases that we have labeled as atherosclerosis or coronary artery disease.

That was the genesis of the central idea of the application and something that, obviously, we were excited to get the chance to pursue as a result of the generosity of the funders, and the vision of Nancy Brown at AHA and Andy Conrad at Verily, to not only award funding in a different way, but to also really try and drive us to think differently about how we executed on a research product. How we move forward, not with a five-year plan, but with a rapid cycle early hypothesis testing, fail fast and fail early, if you are going to fail, strategy. Rethink not just the focus of the research project, but the mechanisms by which you execute on it.

I think one of the core elements of this is, obviously, we want to make sure in doing this that we build on all of the incredible work that's been done in the last 25 or 30 years in coronary disease, whether it's the pharmacologic work, or the genetics work that has emerged in the last few years. Those are all important building blocks, and what can you do that leverages all of that existing data and adds to it? Phenotype is obviously one of the most important areas where you can bring something to the table that add to existing genotypes and also layers in on top of existing pathophysiologic models.

From my standpoint, it was an efficient strategy, and one that we hoped would also help us engage the people throughout the community in different ways of using data that might already have been collected or we were going to be able to collect for the first time.

Dr Anwar Chahal: In terms of One Brave Idea, where is that right now in terms of execution, as you mentioned? What's the progress so far, and is anything that's come out already that you can share with us?

Dr Calum MacRae: Yeah, of course. So we have begun a variety of different approaches to thinking through the best way of exploring this phenotypic space. One of the obvious things is you can take a couple of strategies to move into this unknown unknown. One of them is to take an incremental approach to move slowly from the areas where we have already established knowledge, and to move into new areas from that home base. And the other is to take a more agnostic strategy, which is to say are there orthogonal ways of thinking where you could look at a particular type of biology in a very focused way in coronary disease. You can define that in lots of different ways. You can say maybe we do it at an organelle level, or maybe we do it at some orthogonal component. The microbiome might be an obvious one. Another one that has been considered would be nutritional or other common environmental exposures.

The nice thing about the flexibility of the funding is that we can afford to test multiple different hypotheses early on, see which of them has the best signal, and then invest more deeply in those that have shown early signal. At the moment, we have multiple active projects that are really testing those initial hypotheses. Is there a way of moving from the known genes that cause coronary artery disease and trying to understand are there novel phenotypes that are associated with those. And then another approach would be to take people with very early or pre-clinical disease and test areas of biology that have never been tested in atherosclerosis or in coronary disease in a systematic way.

We could imagine lots of ways of doing it, but you might think about, lets say, looking at endocytosis, a process that we know already is affected by the core genes in familial hypoglycemia, but we've never really found ways to measure that in a rigorous fashion. In large populations of individuals, are there different ... Well, we know already there are different forms endocytosis, but are there discreet port ablations that might affect those.

Another way of looking this might be to pick an organelle. Pick the peroxisome, or pick the nucleolus, pick some other element and ask how does the function of this organelle change in individuals who have early coronary disease. Where its boring each of these types of things systematically, and trying to learn not just which are the most important areas to focus on, but also trying to learn are there strategies that are useful that you could use in another disease. In other words, are there generalizable approaches to expanding phenotypic space that makes sense.

I think one of the things that perhaps we underestimate about a genome is that it is the only bounded dataset in all of biology at the moment. There are no other bounded datasets. There is an infinite number of potential exposures. There's an infinite number of potential phenotypes that we could record, or at least as far as we know, are there ways of beginning to establish the boundaries of the phenome, the boundaries of the exposure or the exposal and how do we begin to do that in a way that efficiently yields new information. That's where we, as a consortium, have focused in the last few months.

We're also, obviously, investing time and energy in thinking how do we begin to remodel the way in which research is evaluated and funded. The strategy that we've taken there is almost like a not-for-profit venture fund where we try and bring in ideas that we think might be able to leverage what's known already and move the field faster towards new pathways or new approaches to prevention, which are the core deliverables of the One Brave Idea award. As part of doing that, we obviously get the chance to interact with lots of exciting and creative scientists and that's something we're looking forward to doing in lots of different venues. We're reaching out to lots of people and lots of people are reaching out to us. We're trying to find ways to evaluate and prioritize science and then bring that science to fruition through novel approaches to funding it, either directly or as a joint venture with a foundation or some other funding source, or even as a joint venture with a commercial partner to try and move the field forward as efficiently as possible.

Dr Anwar Chahal: Thank you very much for that, and I'm sure we all eagerly look forward to the results that are going to be coming out from One Brave Idea over the next few years. I'd like to now move on to genomic medicine training and you were involved in a statement that was put out regarding this. I think training across the world has increasingly recognized the importance of genetics and genomics, but I just want to share one little anecdote.

My wife is a primary care physician, and I was visiting the GP practice where she works, and she'd mentioned that I had an interest in genetics and genomics. One of the partners came out with one of these reports that a patient had sent their sample to a private company, got this analyzed, brought it in to the clinic appointment and asked for an interpretation. The GP partner said to me, "I've absolutely no idea what any of these numbers, values, et cetera, mean, and I actually am looking forward to my retirement, because I really don't want to have to cover all this. Can you help me with it?"

I sort of remember hearing Dr. Weinshilboum talk here at Mayo Clinic, who's really pushed forward pharmacogenomics, and he's been arguing for quite some time, as I've heard you say as well, that genomics and genetics is just going to be a part of the medical record in the same way that hemoglobin or a chest x-ray is. People better catch on because it's here, it's available commercially. People can send their samples directly, without the doctor's involvement, and then it's trying to make sense of all of that.

I think, as a community, research and clinical, we have to take this very seriously. I'd be grateful for your insights on that, and then if you could then tell us what would be the best way for the up and coming generation and for programs to incorporate that into their training?

Dr Calum MacRae: So, I think you're right. There is a general tendency in the public domain to test a variety of different genotypes. And in many instances, I think, the key elements are how do we as a profession, conceive of these tests? I think one of the things that we forget, perhaps at our peril, is that many of these things are problems that we've encountered before. There's a natural cycle of different tests in medicine where they start off in the academic medical centers, they propagate into the periphery, and then eventually they're assimilated as part of internal medicine.

I think the scale of genomics is obviously somewhat broader than many individuals have seen in the types of data that they deal with on a day to day basis. But I think that's something that's happening in everybody's life. In every aspect of your life, you have many more channels to deal with. You have many more choices in the supermarket to deal with.

So, I don't see this as a sort of existential challenge to medicine. Quite the opposite. In my experience, the core things that we need to remember is that DNA is no different from any other assay except for the fact that it's relatively straightforward to do DNA diagnostics. It's technically not as sensitive a set of biochemical issues, as are many other assays that we use in day to day clinical practice.

The other thing that I think is perhaps a key element is it, as I said a few minutes ago, it's a bounded dataset, and it's stable for your whole life. You only need to have it tested once. So, to sort of invert the typical diagnostic paradigms, instead of a primary test being interpreted in the context of an ongoing clinical event, the test may have been present for decades, and the result will evolve over time, in light of the changing phenotype or some new information with respect to that genotype.

What I've actually looked on genomics as is almost an organizing principle for the way that you build care. In fact, I see quite frequently, we now probably have an average one or two new patients a month in my clinic who bring their entire whole genome with them, either an axiom or a whole genome. And so, we've begun to really get to know quite well how to manage patients. Obviously, there are a selective of patients. But one of the things that I have found is that patients are really quite astute in understanding that genotype and phenotype are not deterministic relationships. What you have to do is always interpret these things in context of a probabilistic understanding.

Most patients, I think, when they're told this, understand that we're going to learn much more about genomics going forward than we will ever imagine we could know at the present. That will involve lots of different things. It will involve new ways of displaying data, new ways of thinking about the data in the clinical context. I actually think one of the most interesting things about genomics, and to be honest, any assay is that they rarely reach any form of maturity until they are used in the clinic, until they are actually used in implementation. For example, many genetic tests at the moment, don't change therapy and they don't change outcomes. But partly, that's because they've never been studied in that context.

One of the things that I think Glen [inaudible 00:26:58] has to be really congratulated for is his focus on pharmacogenomics as being one of the early areas in which this will really move forward. I believe that by immersing ourselves in it, by actually trying it in the clinic, we're going to learn much more.

Part of that gets back to the original topic that we spoke about, which is phenotype. The only way to really begin to understand collection of phenotype is if you do it in the context of existing genotype, I think. And so, as we move into new phenotypic areas, we're not going to be able to test everything and everybody. I think there, the genome will end up being an important framework, lifelong framework for the management of a patient's diagnosis, prognostication, and then therapy, potentially in that order.

I think you need a whole different set of skills. You need a whole different set of technologies. But most importantly, you need information that you can interpret in the context of the person in front of you. Until you can make mechanistically important insights with one person, it's going to be very difficult for genomics to really change medical care. That's something that I think we should be focusing on.

I think we've tended to have an associate of strategy for genetics. We haven't driven it into the clinic. As we drive tests into the clinic, whether it's troponin T or whatever, you begin to understand much better how to use them. Although, sometimes, that can also go in quite extreme directions that you may not necessarily anticipate. Troponin originally was a stratification tool for acute coronary syndromes, and now it's virtually a diagnosis in its own right. And I think you'll see that tendency revert over time as people begin to understand the biology of troponin, of isoform switching, and peripheral tissues of the way in which troponin may represent very different disease biologies.

At the moment, it seems like it's a very simple and straightforward yes/no type of test. There's no such thing in medicine, and I think that's what we're learning about genomics. Instead of conceiving it as a series of ten to the nine yes/no tests, we're going to end up with a very different vision and view of how it can be implemented in clinical practice. And that can only come from having clinicians and geneticists work together on this. In fact, one of the things that we've been doing in the partners environment with some of our colleagues, and I have NIH funding to do this with Heidi Rehm, with Sandy Aronson, and with Sean Murphy, is to think about how we display data, but also how we collect information in light of that genomic data that helps in an iterative way and a learning fashion, informed genotype/phenotype relationships in a much more probabilistic manner than we have done to date. There are lots of efforts in that space, that just happens to be one that I'm involved in. But I think it's a generalizable approach that you're going to see moving into the clinic in the next few years.

From the standpoint of training, I think what you want to do is to get exposure to all types of genetic information so you understand common alleles, rare alleles, genomics, and individual panels. I think the best way of doing that is to have that be part of training programs. In fact, with one of my junior colleagues, Dr. Aaron Aday, we recently wrote a short piece highlighting how important it will be for all of us to come together to think about how do we start to introduce the concepts of genomics into standard clinical training programs. And that's something we're working on fairly avidly at the Brigham, and I'm sure there are ... I know there are efforts at many other institutions to do similar things.

Dr Anwar Chahal: That article was published in Circulation in July of this year, if anybody wants to download that. I think if we talk to clinical trainees and ask them what are their concerns about training, as you know, training can be very long in cardiology, which is a procedurally based specialty, whether or not you become an invasive proceduralist at the end of it, there is that component at the beginning. Do you think a standard, in the U.S. a standard three-year program with two years of clinical and one year of research, can incorporate that at a sound enough level to allow somebody to practice? Do you think we're going to look at increasingly a one-year, or a six-month, sort of add-on fellowship for those interested more on the inherited side or more on the genomic side?

I, like yourself, trained in London, and the training programs are longer in the U.K. It was probably six years when you were there, it shortened to five, and now increasingly, it's going to become six and maybe even more with a general fellowship for five years, and then a super-advanced fellowship. Inherited cardiovascular conditions, certainly there, has become a module that is encouraged for people to take and then become somewhat certified in inherited cardiovascular conditions. What do you think there, in terms of incorporating all of that as well as learning basics of echo, and device therapy, and catheterization, what are your thoughts?

Dr Calum MacRae: Again, I look at this as a spectrum. There's a trajectory for all of these types of innovation and knowledge. It starts off being super-specialized, it goes into a more general location, and then eventually, it's an integral part of everybody's clinical practice. I do think that what you're going to see is rather than, and this is already, I think, the case in many elements of medicine. Medicine has already exceeded the knowledge base, even when I was training, by probably a log order in terms of the complexity and extent of content, not that I trained that long ago.

One of the core elements that I think that we're seeing is that we need to move medicine from what I believe has become somewhat deprofessionalized state, to one where you're actually focusing not on the actual core knowledge that you bring with you to the table, but actually the way in which you integrate knowledge. So, I think the focus of training is going to change somewhat. It has had to change in other fields. Medicine, I think, for a long time favored that sort of single, comprehensive approach in one mind. And medicine is going to become more of a team sport, and it's also going to become more of a knowledge integrator profession that it has been for some time.

It's interesting, when medicine started, there was so little knowledge that you really had to have almost every physician be an experimentalist using [inaudible 00:34:48] of one experiments in front of them. I think the way that I see medicine evolving is that as the knowledge base and the rigor of that knowledge base improves, many of the things that we think of as professional activity today, will actually devolve through primary care and, to be honest, into the community. There are many things where the rigor of the underlying [inaudible 00:35:12] are as such that there's no reason for a licensed provider to be involved. We allow our patients to install their own wireless networks without a technician. I'm sure most of them could look after their own lipids pretty effectively if they were given the right information.

So, a lot of stuff will begin to move in that direction. And as that happens, I think the way in which information is displayed, the way in which data are collected, and the workflow around integrating information will change. That doesn't get past the point that you brought up, which is that that will probably take a couple of decades, and in the interim, I think people are going to end up training in modules of subspecialty, but I think one of the things that I sometimes like to ask myself is what's the end game? Where is this going to end up? And can we build systems that train directly for that end game, rather than going through these intermediate steps. I think that's something where I think we tried, in the short piece that we wrote in Circulation, to argue that everybody should have some exposure, and that that exposure can change over time. We should be equipping people, not to know genomics, but to be able to learn how genomics is impacting their patients for the next 50 years.

That model of professional training is actually the one that really was the dominant model until maybe 100 years ago. And then, for reasons that don't quite seem obvious to me at least at the moment, we sort of tended to slowly move to more of a learned knowledge base that was then applied. Physicians sort of steadily got to the point where we're now data entry clerks. The actual amount of professional and intellectual engagement has, I think, slowly diminished in many medical subspecialties and medical specialties.

The opportunity that genomics and other advancements in technology in medicine bring is the chance to, I think, reprofessionalize ourselves to move from just simply defining ourselves in terms of the knowledge base that we each bring to the table, but defining ourselves rather in terms of how we put the knowledge together around individual problems and individual patients. It's a very much more patient-centered biological approach than perhaps we've had over the last couple of decades.

I think these are ... I'm obviously stating a lot of this somewhat in extremes, but I think that these are general trends that you see in medicine. They've happened in other fields as well, and people have overcome them. It's usually a function of changing the workflow itself, of changing the way in which the information ends up in the professional's hands and how you collect the data that you use, then, to interpret the existing knowledge. That, I believe, we haven't really reworked probably since Ozler's time.

It is amazing that we still have workflow ... I mean, it's amazing in lots of ways. It's an amazing tradition, but it is quite interesting that we still have workflow that is probably largely dependent on what Ozler liked to do when he was growing up in terms of the times of day that he got up and his workflow. That's sort of instantiated in many ways in everything that we do. Nothing entirely wrong with it, but there's a lot happened since then that we haven't really changed. Medicine is not yet, in many instances, a 24/7 profession, and yet most other things that have much less in the way of impact on society, are already 24/7 professions in many settings.

So, I think you're going to see a lot of demographic changes in medicine that come from the advent of technology and other industries. And I think those will all transform the way that we imagine training in medicine, along the same sort of timeline as some of the traditional approaches that you described, building out a training module and then having a subgroup of people do a six-month or a year of extra training. I see that as a short-term solution. I think, ultimately, longer term solutions are changing the whole workflow of medicine.

Dr Anwar Chahal: What have you done in your own program at the Brigham to introduce genomic medicine training for fellows?

Dr Calum MacRae: We are building out ... Obviously we have a fairly large cardiovascular genetics clinic. I think probably the largest in the world. We have now seven, soon to be eight, providers working only and wholly in cardiovascular genetics. We therefore have the ability to have our fellows rotate through our genetics clinic. We have inpatient and outpatient genetics services. And we also, obviously, involve our fellows in a lot of the academic pursuits going on in both our genetics and genomics programs in the cardiovascular clinic.

As we do, our colleagues are no longer in training. We have regular, in our clinical conference slot, we have, several times a year, a genetics component. And then, what we have also, is an integrated training program with clinicians and pathologists that is really bringing the individuals who are understanding the technical aspects of the genetic testing with the individuals who are learning and understanding the clinical aspects of that testing. And so, we imagine over time that this will evolve into potentially the type of specialist module that you described. But also, into a fixture that goes all the way through our two-year clinical training program.

We've sort of taken the point of view that we probably need to do a bit of both. We need to, given what I've said in the last few minutes, that we need to take a thread that recognizes a short term and intermediate term need for specialization, but also recognizes that we have to equip every one of our trainees, and every one of our physicians with the ability to begin to learn the underlying sides of genomics, and the underlying approaches to using genomics in every aspect of clinical cardiology. And so, we're doing both of those things, and have active efforts in both.

Dr Anwar Chahal: You mentioned integration with pathologists, but for our colleagues who are not clinicians, what about the research angle, and the scientists, when they're in training? Is that integrated so that we are getting this meeting of minds that is essential?

Dr Calum MacRae: Absolutely. In fact we, thanks to a variety of efforts at Brigham Women's, we have now at least three separate venues in which this occurs. I mentioned cardiovascular genetics clinic. We also have a genomic medicine clinic, which I'm one of the clinical co-directors for, where we actually have cases that come through routine clinical care that seem as if they would benefit from whole genome or whole axiom sequencing. And then we have a weekly conference that's actually led by Dick Maas and Shamil Sunyaev, two of our genetics colleagues, and taped in specialists from Althrop Medicine as well as scientists from the entire Harvard Medical School environment. So we bring everybody together around mechanistically solving individual clinical cases.

And then the third venue is one that's part of a national network, the Undiagnosed Diseases Network. We are one of the sites on the national NIH-funded UDN network. And there again, one of the themes is identifying individuals or families who would benefit from both rigorous genomic analyses as well as much deeper phenotyping. That's been a program that I think has been very exciting, and one that we, again, have learned a huge amount from in terms of how do you begin to build the infrastructure that brings, not just the fresh clinician to see the patient, but somebody who ... A whole team of people, who understand and can evaluate all the biological aspects that are relevant in that patient.

It also brings to bear the scientific expertise that you might need in order to make a mechanistic connection between genotype and phenotype in that one individual. And some of that involves animal remodeling. In cancer, for example, there's a concept that has emerged over the last two to three years of what's called co-clinical modeling. Once you've identified some of the genomic features, it allows you to begin to model in an animal, in parallel with the trajectory of the patient, and individual [crosstalk 00:44:54]-

Dr Anwar Chahal: As some people call them.

Dr Calum MacRae: Exactly. Creating an avatar. And in many instances, that's an avatar that includes multiple different disease models. We have begun to do that in the cardiovascular space. I think, obviously it's early days yet, but I think there are lessons to be learned about how you build the types of infrastructure that allow people to move beyond this state where a patient's outcome is dependent on him seeing the right doctor, on the right day, at the right time.

There are actually systems that funnel the patients into the right venue based on objective criteria at every stage. I think that's the type of reorganization, re imagination of the medical system that we need. We sort of duplicate things in lots of different areas, and you're still dependent on hitting the right specialist, on the right day, at the right time. Or not seeing a specialist. Seeing a generalist on the right day, at the right time, who is able to put everything together. Or even hitting somebody who has the time to listen to your story in a way that helps you identify the exposure or the genetic basis of your condition.

If we recreate the professional environment that I talked about earlier, I think in ways that are both traditional and novel at the same time, I think we will do ourselves a great service and build a platform that lets all of the technologies, including genomics that we've talked about today, begin to impact patients in a real way on a regular basis.

Dr Anwar Chahal: Thank you for that. One question I think is important to look at from the other side, you've gone from One Brave Idea to one revolution in medicine if I can be so bold. You mentioned so many other services are 24/7. You give an example, you can book your hotel in Shanghai sat in the Midwest, and you can change your booking on an app on a phone, and yet in medicine, it's so difficult to arrange an appointment. We have resisted that 24/7 service, aside from the acutes. But for the sort of chronic workload that we have, the 24/7 model has been resisted. What do you think are some of the challenges? Because I can almost hear members of our profession saying, "Well, who wants a 24/7 service and who wants to provide that 24/7 service?", and is it always necessary to have that 24/7 service?

As you say, so many things, such as hypertension treatment, you mentioned lipid management, could actually be done reasonably well by patients who are well trained. And certainly in heart failure, you can teach patients to take their Furosemide or their Lasix by weighing themselves and adjusting it, and can do it relatively well, and relatively safely. What do you think are the challenges to get the profession to realize that this is what's going to happen, and they've got to get on board?

Dr Calum MacRae: Well, I don't think you want to make it somehow mandatory. I think there are elements. Every patient is different. I think that's something we've used as a chivalrous for many decades as a profession. The reality is that we don't do very well. It takes, from the time a medication hits the guidelines, not the trials are finished, but the time that it gets accepted into the guidelines, let's say as a Class I recommendation. The average time to reaching equilibrium in the population is 12 to 15 years in cardiovascular disease. So you'd hate to be the person who got that drug in the 11th year, if you actually end up having your event in year three or four. And yet you can upgrade software for your phone, and hundreds of millions people upgrade it in the first couple of days after a release.

So, we have to build systems that allow us to be as efficient as every other element of our lives, and yet don't, in any way, diminish the importance of the personal interaction, the healing interaction that comes from a patient provider encounter. I think we do ourselves a disservice if we just imagine everything in exactly the same way as it's always been. A lot of it just requires us to make relatively modest changes to the types of things that we do, and to cede some control over some elements of it.

People are not dependent on making cyclical appointments to have doses of drugs tritrated. But once we've identified that a drug needs to be on board as a result of a primary indication, that we allow the titration to take place in an efficient and cost-effective manner. I think a lot of what we do is driven by how we get paid. A lot of ... And that's not criticism, it's natural in every single profession on the planet. You do things the way that the system is set up to have them be done.

And so, I think with relatively little in the way of systems engineering, you can have a 24/7 system without having 24/7 physicians. There are some areas, obviously intensive care units, where you do have 24/7 coverage already, but people are so used to having asynchronous care that being able to literally come home after a night shift and make their reservation for a restaurant the following evening, on their phone, often on another continent, it is a little bit strange that we literally can't book patients into your own clinic without calling up a couple of people.

I just think that some of this is resistance for resistance's sake. Some of it is people actually simply restating the things that we all believe are important parts of medical encounters. I think we just have to be creative about how we move from here to there. I think the thing that I find perhaps most interesting is that somehow the creativity of physicians is not fully exploited. We haven't really asked doctors and patients to come up with new approaches to how care is delivered, to how patients are seen. But I think if we allowed venues where that could happen, that would be actually the way in which we would evolve a very different system.

I think some of that, as I said, just goes back to the way in which everything is structured. All of the payment models, all of the ... Even the types of places that we see patients, are very much anchored in history. They're legacy items and there are lots of reasons why that's the case. Medicine, you can't show up with a minimally viable product. You need something that works perfectly day one, because of the liability. And so, what we need are just to rethink the way in which we even move medicine forward. What we know we can't do is just keep doing what we're doing, and changing modestly, rearrange the deck chairs.

What we need to actually be able to do is find places where we can actually, or venues where we can change things and test new models of care in a relatively low risk situation. I think you already see lots of payers, the federal government, and the NIH all thinking about how you can do that. Some of the [inaudible 00:52:55] efforts, some of the ... Even the NHGRI efforts in genomics. One of the nice things about genomics is because it's a new tool, it allows you to reinvent the way in which medicine is delivered. And so, I believe things as diverse as the precision medicine initiative, and as some of the most fundamental ways in which NIH funding is being restructured, will all potentially impact the way in which creativity and innovation start to evolve within the healthcare system.

I don't want to sound revolutionary. We're all doing all of this, all of the time. It's just not structured in a way that seems to very efficiently reach reduction to practice across the entire medical ecosystem. Part of what I think we need to do is, as a profession, build better ways of identifying where the innovation is occurring, and I will tell you I think it's occurring almost evenly across the entire medical universe, it's just that it doesn't propagate. All medicine, at the moment, is quite local. I think the things that you start to see happening in the industry that will change it are the fact that medicine is becoming much more like every other area of endeavor. It's becoming linked by technology. And once information flows more efficiently, I think a lot of the things that sound as if they're revolutionary, will end up actually just seeming like a series of obvious conclusions, based on the information that we've gleaned from early outlets or success stories.

Many of the things that I've mentioned today, they're not revolutionary at all. There are entire healthcare systems that use these approaches. But they just haven't become generalized because of the way that medicine works. And so, I think that's one of the reasons that I'm a believer that technology in particular will have a transformative effect, just on the way that doctors talk to other doctors or relate to their patients, and the way in which creativity and innovation propagate through the medical system will change very rapidly as a result of that.

And that's one of the great benefits of the electronic health record. I don't think EHR's now are perfect. In fact, in many ways, they're where other industries were 15 or 20 years ago. The supply chain in many large retail organizations was much more sophisticated in the mid-80s than the average EHR is. But what they've done is begin to collect the data in the right place, and in the right way, in a structured format. But as technology begins to cut across different EHR's and across different healthcare network, you'll see things, synergies begin to emerge that will accelerate the pace of change.

It's not by chance alone that medicine has attracted different types of people over the last 50 or 100 years. I think they'll just see the types of individuals that come to medicine be more diverse and more distinctive, and that also I think will help. More distinctive in their skillset, and that will help accelerate change in ways that again, will seem far from revolutionary fairly quickly.

Dr Anwar Chahal: Thank you for that. I wanted to come to the last section of the podcast, and sort of back to where I said it was joking, and you said I wasn't joking about doing three residencies. So, could you tell us a little bit about your own training and your own path? Originally from Scotland, through to London, and then over to the U.S.

And also, if you could share some of those pearls that you've picked up that aren't obvious to us in books, or sometimes are so obvious that they're elusive and not always apparent to young, up and coming trainees, both on the research side as well as the clinical.

Dr Calum MacRae: Yeah, sure. I trained in [inaudible 00:57:15] which had I think a very healthy attitude to specialism and generalism, and the relationship between them, and instilled in all of the specialists the need to always maintain some general medical capability. To this day, I still intend on general medicine for that reason.

I then moved, I did cardiology training in London, and was fortunate to work in a couple of hospitals, one of which had a very interesting, I supposed, quaternary care clinic which had extremely complicated patients. That's where I did my second internship, at the Ross Graduate Medical School in Hammersmith. And everybody who was an intern in that setting had already basically been board certified in internal medicine, so they'd all finished their medical training, come back to do an internship in that setting.

And there, I saw some amazing cases. There was an entire service for carcinoids, there was an entire service for many rare and wonderful diseases. At that point, you began to see how super-specialist knowledge can be incredibly helpful. But it can also be restrictive if it's not applied in the right way.

And then I did cardiology training at St. George's Hospital in London with some amazing mentors. John Camm, who many people will know from his work in atrial fibrillation and sudden death. David Warr, another very well known electrophysiologist, one of the early pioneers. Bill McKenna was my primary mentor, and he was somebody who had worked on the very earliest descriptions of hypertrophic cardiomyopathy when he had originally been at the Hammersmith, and then moved to St. George's.

He taught me a lot about, well many things. First of all, focus in your career, understanding the skillsets that you needed to accumulate in order to a) build a distinctive portfolio and b) to maintain your relevance by accumulating new skillsets as you move forward. And he had actually established a collaboration with Simon's. That was one of the reasons that I ended up moving to the U.S., and had a fantastic time with John and Cricket, at one of the earliest times in genetics moving into cardiovascular disease.

I learned a huge amount from colleagues, at that stage, both at the bench. Hugh Watkins is now chair of cardiology and lecturer of medicine now in Oxford, was a bay mate who was there a couple of years ahead of me and I learned a huge amount from him. I realized ... My wife is from New York City, from Long Island rather, and I realized I had to probably stay in the U.S. for those reasons, and I retrained at that stage in internal medicine again at the Brigham where mentors such as Marshall Wolf and, actually cardiology mentors at that stage were people like Punky Mudge and Pat O'Gara, who then helped me to adapt to the U.S. system.

The only thing I will tell you is that I don't think I ever learned as much as I did in each of my internships. I think the learning curve is incredibly steep. I'd been out of clinical medicine for four or five years, focusing on the lab, before I went back to my third internship. But I still think it was one of the most amazing experiences, largely because of the fact that you learn from every colleague, and you learn from every patient. I think if you go through most of your life thinking like that, I think you can end up doing very well.

Actually, one of the other things that's really important is actually emphasizing those personal connections. The first fellow I had at Brigham and Women's when I was an intern was Joe Hill, who's now the editor of Circulation, the chair of cardiology at UT Southwestern. Almost everybody that I know in cardiovascular medicine, I've encountered in those types of settings. Either in training settings, or in research collaborations, or at research meetings. You just begin to see a whole list of people that have worked together in different ways, and have learned from each other. I think that's one of the most powerful things to take away from research or clinical training.

I then was fortunate enough to get the chance to do a second cardiology fellowship at Mass General. There, I went to Mass General actually because of the focus on zebra fish genetics. I realized at that stage to really be able to study things at the scale that I thought was going to be necessary, I needed a high [inaudible 01:02:40] system, and Mark Schwartz, before he went to Novardis, on the zebra fish and the cardiovascular system, was very inspiring and I had a great time there. And then, ended up spending some fantastic years at Mass General where I eventually became the program director. But again, there I learned an incredible amount from people like Bill Dec, from Roman Desanctis, from Dolph Hutter. All of whom had very strong clinical presence, as well as from the researchers. Mark Fishman, the late Ken Bloch, and many others.

And then also, perhaps one of the most important people in my long term training was Peter Yurchak, who had been ... He had actually defined, I think, the training programs in U.S. cardiology about 35 years earlier. He had been the program director since its inception in the 50s until he retired in 2005 I think it was. And then I became the program director and was there until I moved back to the Brigham in 2009, and became chief in 2014.

I think the trajectory is really, I outline it only to highlight the fact that it took me a long time to get where I was going, but that I spent most of my life enjoying the journey. And I think that's actually one of the most important lessons I took away from it. You can end up finding situations where you feel like you might become frustrated, but in fact, if you go into them with the right attitude, and not only that, if you do it with the right people, you can take a huge amount out of it.

I was incredibly fortunate in the fellowship class that I had at Mass General. Mark Sabatine is now the chair of TIMI, Patrick Ellinor, who is the head of EP and a pioneer in atrial fibrillation genetics. Stan Shaw, who is now the chief scientific officer with me in One Brave Idea. Danita Yoerger, who's the head of ECHO, and an outstanding ECHO researcher at Mass General. Mark Rubenstein, who's a very successful cardiologist, and a fabulous clinician. That group of people actually, I think, together helped me realize how much you could take from training no matter how old you are, and no matter how grumpy you seem when you don't get the full nights sleep.

In the research side, I think the other thing that was obvious was that so many people bring so many different things to the table in research that you should never over or underestimate any aspect of the entire profession. I think I still get remarkable insights into research questions from colleagues who are clinicians, who've never done any research, just from astute observation and declaring a problem in a way that encourages investigation. I think that's one of the most important elements of training is how do you work out what you need to do, and how do you make sure that everything that you do between the start and the finish of that journey is used to help and to improve the way in which you end up doing what you ultimately find as your sort of settling point in your career.

I think the other thing that I will say from the standpoint of research is it's always best to try and think about blending different fields together. What you don't want to do is end up being a clone of one of your supervisors or your mentors. It's really an important thing, and I encourage this in all of our trainees the importance of being a bridge between different disciplines. I think that's something that requires real emphasis.

And then, finally, never ever forget that the single most important thing in all of this, whether it's the reorganization of clinical care or the core research environments, is the biology in the patients in front of you. And so, one of the things that I'm particularly and acutely aware of almost every time I see patients is that the patients often know much more about the condition that they have than you ever will. Listening to them is actually very important piece of everything that you do.

In fact, one of the reasons that we began to move outside the heart in our heart failure research was talking to patients about their pre-clinical elements that they found in their families. So, often, when you see a family with inherited heart disease, before the gene is identified, before anybody has a phenotype that you recognize, the patients themselves can assess who's likely to develop the disease from their intrinsic knowledge of their siblings, and their cousins, and their other family members.

So, for example, one of the families that I've worked on intensely, there's a anxiety disorder that is a much more stable and much more specific part of the phenotype than any of the cardiac arrhythmias, and it's actually turned out to be quite a difficult anxiety disorder to define using even DFM criteria. But when we asked the family, they were very able to tell the people in the family who just were at the normal edge of neurotic from those who truly had the anxiety disorder that co-segregated eventually with the arrhythmia.

The lesson I've learned time and time again is that patients always are a vital and central part of the answer. And it's a pride thing to say, but particularly in genetics and genomics, I think, and particularly with the reemphasis on phenotype, that I believe is necessary, I think we do well to try and make sure our research and our clinical care, our discovery, and our disease management are very tightly aligned. And I think technology is one of the ways that will help that happen. That actually is part of what being a professional really is. If you go back to the early professional guilds, that's exactly how they were formed. It was groups of experimentalists who were interested in particular problems that formed the original professions in European cities during the Renaissance. I think that's something that we would do well to think about as we continue to remodel medicine in the 21st century.

Dr Anwar Chahal: Thank you for that. Lots of important points there, and I guess your emphasis