The Truth About Evidence-Based Medicine with Dr. Gordon Guyatt

Gordon: 
In other words, a hundred fold difference between the increase with risk of smoking and lung cancer and the apparent increase in risk with red meat and cardiovascular events. And that very small apparent increase from the observational studies is easily explained by bias.

Bret: 
Welcome to the Metabolic Mind Podcast. I’m your host, Dr. Bret Scher. Metabolic Mind is a nonprofit initiative of Baszucki Group where we’re providing information about the intersection of metabolic health and mental health and metabolic therapies such as nutritional ketosis as therapies for mental illness.

Thank you for joining us. Although our podcast is for informational purposes only, and we aren’t giving medical advice, we hope you will learn from our content and it will help facilitate discussions with your healthcare providers to see if you could benefit from exploring the connection between metabolic and mental health.

What do you think of when you hear the term evidence-based medicine? And how are you either as an individual or a clinician supposed to see? Just the whole breadth of evidence that is out there and understand what it means for you or what it means for a patient? Today, I am joined by Dr. Gordon Guyatt, who’s known as the Father of Evidence-Based Medicine, and he’ll talk about that.

But we really go through this concept of evidence-based medicine, the strength of different evidence to maybe help you walk away with a better understanding of how do I assess scientific medical evidence? And how do I know what it means for me or as a clinician? How do I know what it means for the patient in front of me?

So I hope you find this helpful. And here’s my interview with Dr. Gordon .

Dr. Gordon Guyatt, thank you so much for joining me today on Metabolic Mind. 

Gordon: 
Pleasure to be here. 

Bret: 
Now we all, I think it’s an overused term to say, this person does not need an introduction. And I would say that’s the case for you with all your publications and accolades within the world of evidence-based medicine.

But, take half a minute here just to tell our listeners who you are, if they haven’t heard of you before. 

Gordon: 
So I’m a distinguished professor of medicine and health research methods at McMaster University. I’ve been in medical research for 45 years or so in terms of the most relevant thing, I suppose by background, that in 1990, to describe what we were doing in running a residency program at Master University.

I was then running, I coined the term evidence-based medicine, and since that time have been writing a lot about evidence-based medicine and advocating the use of evidence-based approaches in clinical decision making worldwide. 

Bret: 
Yeah, and I think that really hits what I want to talk about today, this concept of evidence-based medicine.

You’ve been called the father of the Godfather of Evidence-Based Medicine because you coined that term in that article back in 1990. But evidence-based medicine can mean different things and I wonder if what you meant by it in 1990, maybe it’s used a little bit differently now? So I’m curious how you see the term evidence-based medicine.

What does it mean to you and how do you think maybe it’s evolved over time? 

Gordon: 
We, the way I see it, there are three key principles of evidence-based medicine. One is that some evidence is more trustworthy than others, and we actually have a way of distinguishing the more versus the less trustworthy evidence.

Secondly, when making clinical decisions, we need systematic reviews of all of the highest quality evidence or we get misled, and so we need those reviews to be understand the best evidence available to apply it to our patients. And the third principle, perhaps a little ironic, is that evidence itself never tells you what to do.

And it’s always evidence in the context of patient values and preferences. And if you were talking about the evolution of evidence-based medicine in the publication that sort of sprung evidence-based medicine on the work to the world, the JAMA publication in 1992, we didn’t mention values and preferences in that publication.

By the year 2000, the values and preferences were very prominently part of evidence-based medicine in another publication in JAMA. And since that time, in terms of it being not understood, people still are 25 years behind the times. Some of them, and they do not grasp that. The attention to values and preferences is central and crucial to evidence-based medicine.

Bret: 
Yeah. I think that is so important and something that’s not discussed enough. So give us an example of where evidence-based medicine needs to rely on that, the preferences and how it’s so central. 

Gordon: 
So the, I think you are going to want to go to the area of nutrition.

So I’ll pick a nutritional example. For particular dietary habits, whether it is red meat or trans fats or whatever, the evidence is still definitely not high quality so that we can’t tell people, if you want to be healthy, do eat this way or that way. And so the people have to say, okay, there’s some evidence to suggest your health might be advanced by one way of eating or another way of eating.

And the choice is a value and preference-sensitive choice. You could even, so we have evidence, for instance, so that’s just focusing into health in the nutrition area. We have other consequences. So some people might say we shouldn’t be eating red meat because of the environmental cons, the issues that are related to that and how much you value the environmental issues will influence your dietary habits.

So there is an example from nutrition, if you wanted an example from more traditional, some people might consider more traditional medical areas, there are a number of conditions that increase the risk of stroke. And one of the things we do to prevent the strokes is anticoagulation so that decreases the risk of stroke.

However, it increases the risk of serious bleeding, and some people may say anything that decreases stroke, I’m not too worried about the bleeding. I will use the anticoagulate. And other people might say that bleeding scares me. There are a lot, I’ll put up with the risk of stroke to avoid the, bleeds that are caused by the anticoagulants.

So there’s another value and preference-sensitive choice. 

Bret: 
Yeah, and that one really hits home for me because for years and years, we use this CHAâ‚‚DSâ‚‚-VASc score, which is an and abbreviation for how you evaluate someone’s risk of a stroke with atrial fibrillation. And you just plug in the numbers, and you get a score.

And if you’re above a score, you anticoagulate. If you don’t, if you’re below it, you don’t. But that completely ignores exactly what you’re saying, the patient’s desire. So I really like that you make.. 

Gordon: 
So when I have a patient with atrial fibrillation, I will calculate their CHAâ‚‚DSâ‚‚-VASc score and say, here’s your risk if you don’t of stroke, if you don’t take anticoagulation, here’s your risk.

If you do a stroke and here’s the increase in the bleeding risk, a serious gastrointestinal bleeding particularly that comes with the anticoag. What do you think? 

Bret: 
Yeah, it’s our job to guide, not to dictate. I’m glad you brought up the example of nutrition because we often hear the evidence shows, the evidence says, the studies say and without further clarification of what evidence, what quality of study.

It really doesn’t mean much. And I guess that’s where I think this term of evidence-based medicine has been really misused. That if there’s any kind of a study supporting a claim, it is now evidence-based medicine. So how do you feel about that use of the term evidence-based medicine in that setting?

Gordon: 
A few minutes ago, I said the first, principle of evidence-based medicine is that some evidence is more trustworthy than others. And the second principle is to make decisions, we need systematic reviews of the best evidence, which will include at the end some assessment of the certainty of that evidence.

And we have an approach that we have developed over the last 20 years that’s been used over the last 20 years now widely used called GRADE, which is about assessing the certainty of the evidence. 

Bret: 
Sorry for the interruption, but just want to clarify real quick. When he talks about the GRADE approach, it’s an abbreviation for grading of recommendations, assessment, development, and evaluation.

And it’s a whole different approach to how we evaluate the quality of evidence, and we’ll link to some descriptions about it in the description below. 

Gordon: 
And GRADE has four categories, high, moderate, low, and very low-certainty evidence. And it may be, it is often the case, unfortunately, that we only have low or very low-certainty evidence, and make decisions.

And that needs to be considered. One example that we, give of how things went very wrong in the days before evidence-based medicine was fully developed was, the use of hormone replacement therapy in women who were post-menopausal. There was, as it turns out, low-quality evidence of a decreasing cardiovascular risk.

Hormone replacement therapy, and on the basis of that low-quality evidence, big campaign to get all the women to use hormone replacement therapy. When randomized trials were done, sorry, did not lower the risk of of cardiovascular events and people had to say, whoops, we’ve been giving women the wrong advice for the last while, which was pretty awkward and pretty unfortunate for the women who were giving to bad advice.

It would’ve been reasonable earlier to say we only have low or very low-quality evidence that hormone replacement therapy decreases the risk of cardiovascular events, and that we also have some evidence that maybe it increases the risk of breast cancer. What do you want to do? And that would’ve been, that would then have been a valiant preference-sensitive choice.

And ignoring the fact that it was only low-quality evidence for the reduction in cardiovascular risk was very problematic. 

Bret: 
And that opens up another whole can of worms from an evidence standpoint to say, how do I relate this to the patient in front of me? Because how old are they? When did they go through menopause?

What version of hormone replacement are you using? These are all these factors that can’t be studied in one study that may make a difference in that patient’s risk. So again, perfect example of taking the evidence we have and the imperfect world of trying to apply it to the patient in front of you.

So to bring it back to nutrition. Then, we often hear that observational studies are what told us smoking causes cancer. So if an observational study tells us red meat increases the risk of heart disease, aren’t those the same? And I frequently point out what the difference is

the odds ratio. The hazardous ratio are dramatically different. So, tell us the impact that a hazard ratio or an odds ratio of a study has on being able to believe the conclusions and extrapolate it to the patient in front of you. 

Gordon: 
Okay. Great question. Excuse me, as it turns out, the odds ratio, or you could do a risk ratio or odd ratio or hazard ratio, whatever, which one you pick. Heavy smoking increases the risk of lung cancer by, in observational studies, by a factor of 10. Your risk of lung cancer goes up 10 times on the basis of the observational studies. The observational studies are inevitably. Open to bias, but bias cannot explain a tenfold increase.

There’s no way that bias can explain a tenfold increase that’s much, much too big to be explained by the bias. For instance, red meat and the risk of cardiovascular events, it’s about a 10%. Where it’s tenfold risk, it’s about 1.1.

Okay, in other words, a hundred fold difference between the increase with risk of smoking and lung cancer and the apparent increase in risk with red meat and cardiovascular events. And that very small apparent increase from the observational studies is easily explained by bias. So huge effects from observational studies, which we seldom see, but we did see with smoking and lung cancer cannot be explained by bias.

Very small increases in risk from observational studies. And of the order of the hormone replacement therapy and reduction in cardiovascular risk, which is actually greater than from the observational studies, was greater than what is seen if you stopped eating red meat. That can easily be explained by bias.

So that then is untrustworthy. When a finding can easily be explained by the limitations of the study and the possible bias, then it’s only low-quality evidence. 

Bret: 
Yeah. And then you had previously mentioned meta-analysis, and one of the arguments we frequently hear is that every study says the same, which is not true.

Not every study does, but a number of different observational studies reach the same conclusion. But if they’re all prone to the same bias and the same low hazard ratio, does it matter how many you have that say the same? That conclude the same? 

Gordon: 
You’ve got it, and very well stated. Let me give you another example.

There was study after study, observational study, that said antioxidant vitamins lowered the risk of cancer and cardiovascular disease. People who took anti antioxidant vitamins had an appreciable reduction in cancer and cardiovascular risk, and it was true. People who took antioxidant vitamins did have a lower risk of cancer and cardiovascular disease.

Unfortunately, it had nothing to do with antioxidant vitamins. When the randomized trials were done, people who were randomized to receive or not receive the antioxidant vitamins, no reduction in cardiovascular risk, no reduction in cancer. And what it is, is that people who, the nature of the people who take antioxidant vitamins, they are people who, for a variety of reasons, have lower risk of cancer and cardiovascular risk.

And that at just as you said, perfectly as you said, those biases existed in every one of those observational studies. And so they kept being nicely replicated. But the reason it was replicated was the bias was similar in each of those studies. 

Bret: 
Yeah, that’s a great example of the need to follow up a study like that with a higher quality of evidence with the randomized controlled trial.

So, one of the main biases in that study, or in the saturated fat studies or red meat studies, is healthy user bias. So when you look at the baseline characteristic, the people who eat less saturated fat or less red meat, exercise more, they smoke less. They have higher socioeconomic status and higher education.

And it’s not that eating saturated fat causes you to smoke or exercise less, but that’s just the baseline characteristics. But now the authors say they adjust for that and still reach similar conclusions, although with low hazard ratio. So give us an idea of how do you adjust for that and how accurate is that at trying to counteract healthy user bias?

Gordon: 
Okay, so great points. And that’s exactly what the people who do the observational studies and when they do them well, they do this adjustment. However, so how do you adjust for socioeconomic status? Perhaps you, in your studies, all you may have is your information from the people about where they live.

Rich people tend to live in some areas and poor peoples tend to live in other areas. And so you adjust for the postal code. Unfortunately, that is a not a great adjustment. And even if you ask people about their incomes, people with similar incomes may, their socioeconomic status still differs because of all sorts of other factors related to socioeconomic status.

Ask them how much they smoke. That may, that helps, but you may not get accurate. That may not be completely accurate. So all these factors that we adjust for, typically, have limitations of their specific accuracy for the underlying feature. And there’s always things that we don’t understand or can’t measure or can’t adjust for.

And so the limitations of the data in those things we do adjust for. And the fact that there are things out there that we don’t know about, that we don’t adjust for, all of that limits the adjustment. So in the antioxidant vitamin, the people who did the antioxidant vitamin observational studies were good epidemiologists and they adjusted for everything, but it didn’t work.

It still suggest, it still gave us the false conclusion that it was the antioxidant vitamins when it wasn’t. It was the nature of the people who took or did not take antioxidant vitamins. 

Bret: 
Yeah. Yeah. Very good example. Very good example.

Before we continue, I want to take a brief moment to let our practitioners know about a couple of fantastic free CME courses developed in partnership with Baszucki Group by Dr. Georgia Ede and Dr. Chris Palmer. Both of these free CME sessions provide excellent insight on incorporating metabolic therapies for mental illness into your practice. They’re approved for a MA category one credits, CNE nursing credit hours, and continuing education credits for psychologists, and they’re completely free of charge on mycme.com. There’s a link in the description. I highly recommend you check them both out. Now, back to the video.

Now you had mentioned the GRADE criteria before, and so there, people I guess mentioned different criteria when looking at studies. So I want to get into the GRADE criteria in a minute, but there’s also the Bradford Hill criteria, which has been cited many times.

And I want to at least read what it is. it was back in like the sixties and they said you have to factor in strength of evidence, consistency, specificity, temporality, biologic, gradient, plausibility, coherence. And then experiment and analogy. So a lot of different terms, which by themselves probably, may not mean much, but in the concept of trying to decide how relevant a study is, so did those factors play into the GRADE criteria and give us a, more of a definition of the grade criteria as well.

Gordon: 
Okay, so we’ve actually written a paper that says, GRADE is consistent with the Bradford Hill criteria. Okay, so Bradford Hill had it basically, right? But you read off these things and the issue is how in fact do you apply them? Which ones are more important? Which ones are less important? What grade has done is build in a considerably more, I would say, sophisticated and that certainly specific way, how those factors play out.

And so we’ve reformulated those. And so, first of all, study design. Study design is crucial. So in the high, moderate, low, and very low-certainty or quality of evidence. Randomized in grade, randomized studies start as high and observational studies start as low, but randomized trials may be limited. They may be limited by risk of bias.

They may give inconsistent results. They may be indirect or done in one population, you’re interested in another. They may have small sample sizes and be emphasized. They may have all these problems. So although randomized studies start as high-certainty evidence, the sophisticated get GRADE approach considers these things, and maybe it’ll only end up moderate or maybe even low or on occasion.

If it’s got all those problems, it may end up as very low. Observational studies start as low, but in particular, as we’ve said earlier in this conversation, if the effects are large or very large, it may end up as high-certainty evidence. So in fact, we have high-certainty evidence that smoking causes lung cancer.

We have high-certainty evidence that replacing your hip does, for most people, cures your hip osteoarthritis. We have high-certainty evidence that if you’re dying of renal failure, dialysis will keep you alive longer. We have high-certainty evidence that if you’re presenting with diabetic ketoacidosis in the emergency room, insulin will get you through your episode of diabetic ketoacidosis and so on.

There aren’t a huge number of these. I’ve listed a few, but there are instances in which the effects are very large, and particularly if they act quickly when we have only, all we have is observational studies. But because of those very large effects and quick action, we have high-certainty evidence from observational studies.

So basically, I summarize the fact that GRADE has a sophisticated, everything in grade is consistent with Bradford Hill, but it’s a very structured, sophisticated approach, that allows us for any particular intervention to, when we look at the evidence, to classify it from high to very low. 

Bret: 
That makes a lot of sense.

And when I see a paper that says this is a GRADE analysis of the evidence, I pay a lot more attention to it, but I’m curious what you think. Has it been, has this GRADE process specifically, or even just the concept in general, do you think it’s been adopted among clinicians, guideline writers, researchers?

Do you think it’s pretty widespread that people recognize this, or do you think there’s still work to do to, even within the world of scientific evidence to get people to understand these concepts more? 

Gordon: 
We’ve done, we think we’ve done, very well. Over 120 organizations worldwide have adopted GRADE.

They include the World Health Organization, the Cochrane Collaboration, very prominent in the world of evidence-based medicine, leading American organizations like the American College of Physicians, and the American Thoracic Society. And the most popular worldwide source of guidance for PR practice up to date has over 10,000 GRADED recommendations.

So we’ve, we think we’ve done very well. Where there is resistance is perhaps understandable in areas where there are only observational studies or largely observational studies. The people working in those areas would like people to pay attention to their work. They would like their findings to be implemented, but sadly, very often there are going to be stuck with only low-quality evidence and it’s makes it very hard to say, oh, I’ve shown that red meat causes cancer.

Oh, unfortunately it’s only low-quality evidence. So, perhaps, maybe we should still not push this for this idea that red meat causes cancer too hard. I can understand that’s not, that’s a little sad situation. However, we of course, say the solution isn’t to pretend that you’ve got high-certainty evidence when you don’t.

However, in areas rare disease, nutrition, in some areas of public health where you’re only going to have, or largely have, only low-certainty evidence. People in those areas are a little reluctant to adopt GRADE. I can understand it, and it’s a little hard to say, oh, even if we do our best, we’re only going to get low-certainty evidence.

However, obviously, we think it’s a mistake to pretend you have higher certainty evidence than in fact you have. 

Bret: 
Yeah, I think that was a fantastic explanation. And it’s funny, it’s easy to fault the researchers. Oh, the researchers are doing this low-quality evidence, like terrible researchers.

No, they’re working within the confines of what they have, but it’s really how everybody else takes that research and what they do with it. And so I, one of the most glaring examples I think is with the United States dietary guidelines, how we take low-quality evidence and say, okay, we know what is healthy.

We know what you should eat to be healthy, and so we’re going to, we’re going to dictate for the whole country what you should eat because we have the evidence. But to me that seems like a gross misuse of the quality of evidence. Would you, it sounds like you would agree, but I don’t want to put words in your mouth.

Gordon: 
No. Yes, it would be reasonable to say, we have some indication, but perhaps it is better to eat this way than that way. But please don’t take this too seriously. There’s some suggestions, but that’s really all we can say, and it’s the, it’s the presentation. 

The most recent, alcohol causes cancer. No, maybe low odds ratios. And the evidence isn’t completely consistent. It’s possible that there’s a small increase in risk. But that’s not how it’s, that’s how it should be presented, but unfortunately, not. 

Bret: 
So that’s another good example of a specific aspect of the studies that would be helpful to talk about, like the gradation of risks.

Because with alcohol, if you are an alcoholic, chronically abusing alcohol, right? The risk of cancer is substantial. If you’re more of a moderate a user of alcohol, not such a high risk, much smaller risk. And you can use the same for when you look at the red meat studies, even though they’re so flawed, it’s only really the highest quintile compared to the lowest quintile with no gradation in between.

Whereas for tobacco and smoking, every step along the way, there was an increased risk for cancer. So I guess I’m trying to formulate this question as I’m speaking, like how do you see that in terms of the difference between a clear gradation versus the highest risk, there are the highest users to the lowest users?

Does that affect the quality of evidence in your mind? 

Gordon: 
There was some evidence of dose response, which is what you’re talking about, with the red meat. People say, oh, that supports it. However, if it, if the gradient risk goes from 1.01 to 1.15, that’s a very different thing from going from 2 to 10 with the smoking.

Yes, dose response can help, but only again, if you’ve got big gradients, a small dose response across a very narrow range of increasing risk is not very convincing. 

Bret: 
Yeah. Now I want to take another quick transition here and talk about observational and epidemiologic studies and if they’re helpful in the negative.

So, I’ll give an example. In the low carb and ketogenic world, we often hear people say, vegetables and plants have antinutrients and phytates, and they can be harmful to your health. But then you could say, look at all these observational studies and the people eating the most vegetables and the most plant foods live the longest and are the healthiest in these observational studies.

Now it’s low-quality evidence. It doesn’t prove that the vegetables by themselves are healthy, but it it makes a good argument for how bad can they be if they’re doing well. So do you see the the observational study in the negative as being maybe more powerful than the positive?

Gordon: 
It depends on the circumstances. Let me give you a example where that argument holds very nicely, and that is in the association between vaccines and autism. And the, when the studies were done properly, no association with vaccines and autism. That is, despite the fact that the word got out, vaccines cause autism.

And so what biases would you expect if your kid was autistic? Your life would say, I remember the kid got the vaccine, the kid doesn’t have autism. Maybe you forget, and that would create a spurious association between the vaccine and the autism. So all the biases that you would expect would create a spurious or false association, and yet they still didn’t find the association.

So there’s one where a negative study is really very powerful for the nutrition, I am not so sure. And the reason is that the people who quote eat healthy, tend to be higher socioeconomic status exercise, don’t smoke, et cetera, et cetera. And so the bias is going to be in favor of the way they eat, apparently doing better.

When you have all these other things that make them do better, that may well be responsible. And so as a result, the ruling out the harm is less compelling. So you’re right, if all of it goes in the same positive direction, it becomes less likely to be harmful. I would quite agree. But if the biases are still, to make it look better than it really is, you don’t completely rule out the harm.

Bret: 
Okay. Yeah. Yeah. that makes sense. Now with this discussion, I could see somebody watching or listening and being like, oh my goodness, this is so complicated. Like, how am I, as an individual supposed to know what to trust? Some people are saying, and, I guess, the vaccine example is a great example, because it’s such a hotly divided discussion.

But let’s focus on nutrition, too, more because someone’s going to say, look, the experts say this, or the, how am I supposed to be able to interpret a study and know whether to believe it or not? So what kind of advice can you give to the average person, right? Not even the clinician or the scientist, maybe we can talk about them separately, but the average person, what kind of advice would you give them on how to decipher all this?

Gordon: 
What you, they would have to know that there are evidence that is more trustworthy and less trustworthy. And they, even the general public, if we got the message out, randomized trials are more trustworthy and observational studies are less trustworthy. The GRADE system is something that allows one to make these distinctions.

And if you’re listening, who is the expert? And what evidence are they citing? So, you have one of two things. Somehow you’ve got to identify the trustworthy and the untrustworthy experts. Easier in the medical world where we can tell doctors how to recognize the trustworthy and yeah, trustworthy pretty easily.

More difficult in the, more difficult for the general public because for instance, you would think you would, when the public health officials in the nutrition world who work for the government say, you’ll be healthier if you eat this way, it’s unfortunate. You’d think that was a trustworthy source and it is trustworthy to the extent, if they were only to qualify and say, we’re not really too sure.

Some suggestion that this might be better, then that would be accurate and appropriate. The problem is when they say, you will be healthier if you eat this way. So it is challenging, but the public in some way has to be able to identify what is trustworthy and less trustworthy and have, we need to educate to have some notion of what, of identifying what is trustworthy and what is trustworthy that will get through to the, that will get through the general public. I have one colleague who has spent the last part of his career looking at, and effectively trustworthy versus untrustworthy evidence.

And he uses the phrase making informed health choices. He thought, he believed you can teach school kids to do this. And extraordinary where you chose to do his first randomized trial, he did it in Uganda and he found in Ugandan schools with the low, obviously limited by resources, he randomized schools to be teaching informed health choices or not.

And he found you can teach kids to make informed health choices and the finding that I liked best from his randomized trial was, he then also tested the parents of the kids, and the parents of the kids in the intervention group who received the instruction in making informed health choices did better than the parents in the other, the kids were coming home and teaching the parents.

So, you can teach this, so I think these studies, but we don’t, if we really, the thing that would be really great is if we started teaching from grade school on up these principles of making informed health choices because my colleague has shown it’s possible to teach it even in the grade school.

It’s the latest, I find it somewhat amusing. He’s really pushing it. He’s now trying to teach it in nursery school anyway. He may be, maybe he’ll pull it off, but whether he succeeds in nursery school or not, he’s already shown in grade school you can do it. And certainly, in high school and certainly.

For adult education. It is possible. It, the, basic principles, it’s all seems very sophisticated, but the basic principles and, it’s what we do with physicians. We say, okay, you are not going to be able to make these great assessments, but here are the principles. You can recognize who’s following the principles.

Bret: 
Yeah. I love that idea of teaching kids younger how to do this, especially now in today’s day and age with social media where everybody has a voice and it’s even.. 

Gordon: 
Yes, absolutely, and the, one of the things that certainly happened to me with evidence-based medicine, you start to apply the principles elsewhere.

So in other words, yes, we did this. Things are more, you have trustworthy evidence. But you also have more and less trustworthy evidence about education, and you have more and less trustworthy, evidence about transportation policies, et cetera, et cetera. The whole world, it applies to any policy choices of any sort.

And so if you teach it, if you teach it for health, people may well get the message, Hey, maybe I should think about all other aspects of life in the same way? 

Bret: 
Yeah. And, then I guess the last point to make is bringing it back to the clinician, is for the clinician to be able to make that connection between evidence, quality of evidence, and what it means for the patient in front of me, rather than seeing everybody as a guideline or as everybody, applies to the same evidence.

And I really hope that your emphasis on the GRADE of the evidence, will help people do that. And do you think we need more of an emphasis on that as well to say, look, we know you only have five or 10 minutes with your patient. at least that’s how it is in the US. I don’t know how it is in Canada, but we still have to individualize the recommendations.

And, I guess, it’s going back to what you said before about adding that third aspect of evidence-based medicine about the patient’s wishes. I think we need to do more of that. So are you seeing that catch on or do you think we still need to push for that. 

Gordon: 
The lots of work still to do, but GRADE can be very helpful.

So as you have, as you have noted, physicians are very time constrained. And so that means we have to ration our time GRADE can be very helpful because it classifies recommendations as strong or weak or conditional, strong recommendations. You don’t need to spend a lot of time on shared decision making.

Somebody comes into the emergency room with a myocardial infarction. You don’t have to go into a long song and dance about whether they should take your aspirin pill. Say, hey, this reduces your risk of dying at the moment with very little side effects. Please take it, okay. Whereas, like the atrial fibrillation example that we talked, you talked about earlier, you want to get it right with the patient.

You have to explain the bleeding risk that goes along with anticoagulants and the decrease in stroke risk and the magnitude of the effect. And you have to, the physician has to, understand what the evidence shows. So in fact, what we try to teach physicians, we don’t expect them to do the, we don’t expect them to make the GRADE assessments.

We teach them enough that they respect the GRADE assessments and so they know high-certainty evidence means this low-certainty evidence means this. And what we focus on in teaching them is understanding the results. So understanding how much anticoagulation decreases the risk of stroke in your low-risk patients and your high risk patients.

How much it increases the risk of bleeding. That’s what you got. That’s what the physician has to understand to be able to do appropriate shared decision making. And then you ration your appropriate shared decision making to what we call value and preference-sensitive choices. Whether you take aspirin in the emergency when you’re having a heart attack is not a value and preference-sensitive choice.

Every informed individual would say, please give me the aspirin, whereas the atrial fibrillation in low-risk patients. It is a value and preference-sensitive choice, and that’s where you have to devote your time to shared decision making. 

Bret: 
Yeah. Great examples. Great examples. Thank you. This has been just a really enjoyable conversation.

I really appreciate you coming on and sharing your wisdom with everybody. If people wanted to follow you, learn more about you, learn more about the grade system, is there somewhere where you can direct them to go? 

Gordon: 
Yes. If they want to, you can, i’m okay to share my email address, which is just my last name at McMaster, the institution I’m at dot ca (guyatt@mcmaster.ca), and anybody who emails me, I can direct them to further information. 

Bret: 
Great. Thank you again for joining me and I look forward to hearing more from you in the future with all the work you’re doing. 

Gordon: 
Thank you very much. 

Bret: 
Thanks for listening to the Metabolic Mind Podcast. If you found this episode helpful, please leave a rating and comment as we’d love to hear from you.

And please click the subscribe button so you won’t miss any of our future episodes. And you can see full video episodes on our YouTube page at Metabolic Mind. Lastly, if you know someone who may benefit from this information, please share it as our goal is to spread this information to help as many people as possible.

Thanks again for listening, and we’ll see you here next time at The Metabolic Mind Podcast.