 Welcome. This is Craig Thomas, your host on Much More in Medicine, part of the Think Tech Hawaii live stream series. And today we're assisted by our engineers Ray and Rich. Thank you. And joining us today in the studio is Tom Forney. Tom is an emergency physician in our group, Hawaii Emergency Physicians, and works on four islands. He's here today actually both for his perspective as an emergency physician, but also because his mission in our group and in the larger health systems is as care coordinator and police of variability. So Tom, we're delighted you're here. The theme recently has been where is the 18 cents out of every dollar generated in the U.S. that ends up in healthcare? Where is it going and is it working? Well, thank you for having me today. I'm honored to be here. Yeah, it's amazing that there's so much spent on healthcare in the U.S. and we're all trying hard to do the right thing and take care of our patients, but it's expensive. And the overwhelming theme is where is all this waste going? So one of my platforms within the group is trying to coordinate care for 90 doctors across four different islands and multiple different health systems. And it's a challenge. Like I said, we all try to do the right thing, but we all do it differently. Hurting cats. You've done a nice job. In fact, it's why we're here today. Historically, in medicine it was sort of what I would call eminence-based medicine. Namely, you went to medical school, somebody with gray hair like I have now would say, this is how it's done, son. And you'd be like, cool. And there wasn't a lot of evidence. There often had never been much evidence. And we were a guild. And that was how medicine was done. And pretty clearly, a lot of it was useful. For example, do I think that we should stop doing appendectomies? No, although interestingly, and you can pursue this or not, it looks like antibiotics work a fair amount of the time, which in my training days would have been just heresy. But many other things we've looked at don't work. And in addition, the amount of potential diagnostic tools, therapeutic interventions, medications, all sorts of things are just landing like crazy. In fact, I refer to this whole thing as we're part of the medical industrial complex. And that means we need to scrutinize what we do and shape it and look what works and what doesn't. Yeah, absolutely. It's fascinating. It was, it was the master craftsman apprentice model for a long time with little evidence of what I hear is even now, best case scenario, about 20% of what we do actually has strong scientific evidence. The rest of this is a little fuzzy. Yeah, knowledge is expanding. And it's, it's hard to hard to keep up with everything. And moving, trying to trying to take that next step on how how do we take what was done and improving it and making it better. And so you also have been leading the charge for this over the the year. So yeah, it's been a wonderful project and we are just starting. And the listeners, it may seem surprising how little evidence there is to support many of our practices and how when these practices get scrutinized, there may not be benefit in the intervention. And by the way, of course, if there isn't benefit, all you're left with is harm. Because essentially everything we do has potential for harm also. So for example, I'm an old guy, I've got old knees. And until relatively recently, the belief was that get a little arthroscopic clean out, you'll be fine. They did this great VA study, which I'm sure you're aware of, where they did a sham procedure. And it turned out, whether you had the sham procedure or the real thing, two years later, you couldn't tell them apart. So guess what? Shouldn't be doing that. There's lots of things like that. Yeah. And it plays out time after time after time of, you know, the old adage was 50% of what we learned in medical school is going to be wrong. We just don't know which have. I think that's too big of a number. I think it's probably more like 70, 80%. And so that's, that's where the lifelong learning comes in and trying to do. And it's so much of what we're trying to do with the standardization part is we work in a highly complex field. So take complexity plus high variability, that's going to lead to poor outcomes. And so if we can remove at least the highly variable part of the equation should lead to better outcomes. And this has been playing out, this already has played out across multiple different industries, whether you want to look at the automobile industry, the electronics industry, or honestly, most any other fields have already gotten this figured out. For instance, you know, back in the 1920s, if you went and bought a car, it would be customized for you. But if anything broke, well, you got to go to that same person to get that part. Whereas the Japanese automakers in the 70s and 80s change it to where it's highly reliable, you know what you're getting, you go to the dealership and it's done. Medicine is trying to get to that same place to where whether you see, you know, one of 10 different cardiologists, you'll get a relatively similar take on things or whether you see 10 different emergency physicians, a relatively standard. And again, that's our goal of, across our group, is whether you show up at one of nine different facilities that you'll get approximately the same care for a similar condition. Yes, exactly. And so it strikes me that we have both a growing challenge and fortunately, if we implemented some solutions. And the growing challenge is that information is increasing rapidly. We're actually learning the answers to some of the questions that we thought were settled long ago and turned out were not. And the question is how do you tap that knowledge to change practices? And then of course, the other piece is, and you have to track outcomes to see how you're doing. And so I know that's what you're working on. Let's talk about this challenge. What should we be looking at? How do we shape behavior? How hard is it to change behavior? It's it's hard. It's I think the saying goes, people like change, but they don't want to be changed. And fortunately, physicians for the most part are very intellectual and will respond to data. And so with the beauty of electronic medical records and the advent of data is we're getting more and more information to be able to then share with our doctors and say, you know, here's this variability or here's how we're doing as a group. And we're all competitive by nature as well. And nobody wants to be the outlier. And so sharing that data helps to move the needle, whether it's, you know, in a particular specific medical scenario or just general productivity or you name it. I think that the transition we've been going through is trying to figure out, we now have all this information. How do we use it and harness it to create a positive change? Exactly. And after the break, we're not close to that yet. But after the break, we're going to look at at least one specific example of what the data suggests, how to approach it, and why and what were the outcomes. So hold that thought. We'll also discuss after the break sort of what the initial variability in the group that led to wanting to do that in the first place. Let's talk first though about variability itself. Can't be good. Can't. And so Ed Deming is a famous person regarding just improving processes. And his saying is inappropriate variability is the enemy of quality. And specifically meaning you can't have a highly variable process and high quality. And so in any outliers or too much of anything is not good. So if you go to a coffee shop and it's too hot or it's too cold or it's too dark or it's too light, you're not happy. And coffee, I'm sorry, you're correct, of course. And so it's the same thing with with our industry as well. If there's, you know, again, you don't want to see four different emergency doctors and get five different answers. That's what we're working towards. Exactly. And it's tricky because every presentation, every individual is a little different. On the other hand, the concerns for a spectrum of presentations should be similar. So that implies that the overarching approach, trying to figure out whether the risk of a condition is low, medium or high. And by the way, of course, you want to look for the condition you can actually impact. So is it particularly important if I diagnose on the first visit somebody with terminal cancer? Well, it'd be nice to know. It would certainly help figure out how to chart the next course. But if the condition is terminal, and sadly, this is one of the, as you know, truly horrible events in emergency medicine. Somebody comes in with an ache. And you're like, you know what, that's a brain tumor. This is a life changing awful moment. But the truth is, it's not as important to find that as if I find meningitis or a QMI or something we can impact. So how, how should we be thinking about stratifying the relative importance of different possibilities on arrival? Absolutely. And as you mentioned, most most processes can be disease processes can be placed into low, moderate or high risk. I think we're all in agreement on generally what to do with high risk is you take care of it. Moderate risk, I think is where the lot a lot of the variability comes into play. The low risk, there's real chance for potential harm, in addition to trying to track down whatever emergency condition that there is. Perfect. The and with low risk, you probably shouldn't do anything more, or at least not much because as we said earlier, the if you can't benefit from the treatment, the only thing left is harm. This is Craig Thomas with Tom Forney program is much more on medicine, part of the Think Tech live stream series, and we'll be back after the break. This is Think Tech Hawaii, raising public awareness. I just walked by and I said, what's happening, guys? They told me they were making music. I'm Ethan Allen, host of likeable science on Think Tech Hawaii. Every Friday afternoon at 2pm. I hope you'll join me for likeable science. We'll dig into science, dig into the meat of science, dig into the joy and delight of science. We'll discover why science is indeed fun, why science is interesting, why people should care about science and care about the research that's being done out there. It's all great. It's all entertaining. It's all educational. So I hope to join me for likeable science. Welcome back. This is Craig Thomas, host of much more on medicine with Tom Forney from Hawaii Emergency Physicians. And we were discussing before the break the, where is our 18 cents going? Why is there variability? And what should we do about it? And Tom, project near and dear to both of our hearts has been the elusive chase of the pulmonary embolus. Would you mind telling our audience, A, what is the pulmonary embolus? Why we might be worried about this? And sort of what changed to diagnostically in about 2000 that kind of changed everything? And then maybe not so much after all? Absolutely. So pulmonary embolus is essentially a blood clot that has traveled to your lungs and has can range from a small nothing to something that can potentially kill you. And so for years, we would hear this is a silent killer. There's more people dying from pulmonary embolism than we thought. And as you mentioned, around 2000 CT imaging became much more widely available. And so we all of a sudden had an opportunity to diagnose this potentially deadly disease. And so we've ordered a lot more CT scans over the last 18 years. We have diagnosed a lot more pulmonary embolism over the last 18 years. But fascinatingly, the death rate has not changed at all from the disease. And so we're treating a lot more, we're diagnosing a lot more. We have not impacted mortality at all. Isn't that sobering? So just to sort of recap, first of all, this is a sneaky disease, not very common, but it can kill you. And in fact, there's a treatment. So this is just what we're looking for in the emergency department. As we said before the break, no point in looking for something if you can't fix it. Well, we can fix it. And before about 2000, our diagnostic options were terrible. But all of a sudden, they got good. The test is available, essentially every emergency department in the state. It's pretty easy to do. And it gives you an answer may not be correct, it turns out, but kind of over calls things a bit. And so we're all thrilled. And look at this, we're diagnosing tons more pulmonary embolism. And look, the case fatality rate is almost zero. God, we're great. Oh, wait. Same number of people die in every year. I guess we're giving people die, giving them radiation, putting them on blood thinners, which can also kill you for no benefit. So this was very sobering. It was something you didn't mention is in our group, there was nearly sixfold variation, our favorite topic on between individual providers at the same facility, with essentially the same patients in their rate of ordering these tests. So as Deming would say, variability is the enemy of quality. So this was clearly an enemy of quality. So what did you do? Absolutely. And just to kind of reiterate one or emphasize one point that you just made, there's real harm with the work upon this, whether it's the radiation or the overtreatment. So we were, we're there was real harm and looking into it. But yeah, that variation was phenomenal. And goes back to the point that we were talking about earlier on data and physicians respond to data, we didn't know we all think that we're doing the right thing for our patients. We all think we're doing what's best. But not until you see, wow, there's a sixfold variation and what what my partner next to me is doing. And so we took that data, we then did an extensive review of the literature. And part of the reason why we were pushing the pulmonary embolism is the the comments that you just made. But also, there's pretty clear best practices on how to handle it. And so we took the best practices that are out there, implemented into a protocol for our physicians to follow. And then we've tracked the outcomes. And so far, it's been a success. The overall number of CT scans that we're ordering is dropping. The percent positive of pulmonary embolism that we're diagnosing is going up. And so theoretically, based on my understanding of the data, we're causing less harm, and we're providing more benefit for our patients, along with reduced variation, increased quality. And overall, it helps the financial system. And it's we're costing the entire health system less money, which is important, because then you can spend the money on other things that impact health. So I'd like to just sort of reiterate a couple things. So actually, you describe it. The basic premise is, depends, and you alluded to this just before the break, of categorizing people by low, medium, and high risk. And this applies to pulmonary embolism, but many other conditions. Because if they're low risk, you can't get them to zero. None of our tests are perfect. And if the risk is low enough, you should just stop. Whereas if they're high risk, you should proceed with therapy, because again, our tests aren't perfect. And they can be falsely negative as well as falsely positive. So is that a fair characterization of your sort of testing and pathway to therapeutics process, the different buckets of risk? Absolutely. And as you mentioned, you can never get to zero risk. It's just that does not exist. And the closer you get to zero risk, the more real harm that you're causing our patients by finding false positives and kind of unnecessary extra testing. And so yeah, for pulmonary embolism, but the best we can do is in that 98% or 2% risk of missing this deadly disease. That's about the best that we can get. And if we try to take it any lower, we're causing real harm. Right. And in fact, you cause harm a couple of ways to test themselves are harmful. But in addition, they have false positives. So if I do a test that's not indicated on you, I irradiate you, and then I put you on blood thinners, if it's positive, but if you didn't have the disease, all I'm doing is hurting you. And so my sense, and I'm interested in your thoughts, is we doctors, and it's our fault, because we've often considered we can rule out disease or we can cure disease. And the answer is, we can find a low probability of disease and many of our therapies have a good chance of achieving a cure. But it's not absolute. So we've actually presented to the general public, which is understandably desirous of hearing this. Oh, yes, we can do this. And I think we need a new narrative. More shared discussion, things are gray. How would you proceed with that? You're absolutely, I mean, it's, yes, medicine is a very gray zone. There's some some areas where yes, we have the evidence we know what to do. That's about 20% of the time 80% of the time, it's gray. And everybody has a different risk threshold. So for instance, well, whether it's, we'll stick with the pulmonary embolism topic. If I talk to a patient say, I am 98 99% certain that you are fine. Most patients will say, great, I'm okay with that percentage. I don't want to take the risk of getting more aggressive. And so involving our patients in a discussion of saying, here's the evidence that we have. This is is my thoughts on your presentation. How do you how let's let's solve this problem together on deciding what we're going to do next. Difficult but clearly essential. So just a couple numbers to get people thinking about this in a little different way. The before your project started, I think that the positive study rate was right around or just under 4%. So in other words, it took 25 studies to get a positive one. Well, that's a lot of x-rays and a lot of dye to get one positive result. And now I think it's on the order of triple that, which is way better. It implies there's a third less studies done. I don't know what the right answer is. Should it be higher than 12% maybe shouldn't be 100%. That means you're missing things. But just to give people an idea. One out of 25 probably doing too much. The other thing is physicians have different risk tolerance for themselves and for their patients. In other words, I'm much less inclined to take risk on you than I am on me. And that's probably not right. Probably I need to let you help decide what level of diagnostic and intervention and risk to assume bearing in mind zero is never going to happen. And with that taking on the physician's perspective is historically with medicine, we had a very last case scenario approach or the heuristics. And that really clouds your judgment of well, I remember that one case that went well or that one case that didn't go well. And now I'm basing everything off of that one case as opposed to basing it off of the general medical knowledge of what's what's best overall. And kind of shared processes of best practice helps to take that part that that last case that last case bias away. And let's just say, Okay, this is the evidence let's let's treat our patients off of this evidence with reduced variability, as opposed to off of your your personal experience with that one case previously. Well, as they say, if you come up to me say, Craig, remember that case you saw last night? I'm probably not going to win the Nobel Prize for Medicine. Bad news travels fast. It seems like good news doesn't go anywhere. But what you described is exactly right. And it's partly because we all sort of still practice in silos. So I know about my bad case. I don't. And there are going to be some. This is, we can't reach certainty. Biologic conditions are complex. And there will be bad outcomes. Doesn't mean by anything wrong. Sometimes people did, but surely not. And the trouble is if it shapes my practice in a way that doesn't actually comport with care pathways, I'm actually hurting people. And so this is where it's nice to share data, see where you fit in the spectrum of providers. See that yes, there have been returns for unexpected returns for everyone else also. And I think it's all part of how we use data to help ourselves and each other. Yeah, that's true. And that's, again, we we respond well to data. And I think that's the next steps with where we're going with this is how how do we get data more more widely available so that it can help to shape my practice. And you know, if if I find out that something happened with the patients that I saw last week at a different facility, I then can find out about it. And that also helps to shape how we improve. You know, Tom, it's been great having you here today because we're just scratching the surface. So we'll have you back. It's exciting times. And we look forward to seeing how we can use data, air pathways, and more knowledge to shape our practices. Thanks for coming. I appreciate it. Again, thank you for watching. This is Craig Thomas. We're discussing with Tom Forney much on much more on medicine, how we're spending our money, how we're using it to impact health. And I appreciate your attention. Thank you.