Moneyball Rheumatology Save

Hi, everyone. I'm Jack Cush with RheumNow. Today, we're gonna talk about Moneyball, Moneyball rheumatology. This is based on the movie Moneyball starring Brad Pitt and Jonah Hill. It's also based on really the book written by the great Michael Lewis called Moneyball.

It's the story of a baseball team that is trying to do it differently. So let's get into moneyball rheumatology. The story really begins, with the book as written by Michael Lewis that became then the movie. I saw this about 10 times on the airplane, liked it, but I was bothered by it because behind the story was this issue of how do you find undervalued players? How do you find an edge in the situation you're in, whether it's business or baseball or medical care?

The subtitle to the book and movie is The Art of Winning an Unfair Game. It's all about the salaries and the budgets that Major League Baseball was using back in the year 2002. For many years, the New York Yankees, a very rich franchise, was spending the most on its payroll. In 2002, they spent over almost $130,000,000. By comparison, you can see on this graph that other baseball teams didn't really have that much money, and small market teams like the Oakland Athletics were only spending $38,000,000.

And the question is, how can they compete? Michael Lewis wrote this book because he wondered, why is it that the left fielder is making 200,000 and the right fielder is making $4,000,000? How does that work? And in basically trying to understand what was going on in baseball, he found that there was this movement afoot based on finding players of value that he could afford or that Oakland could afford. So tradition was that you chose the best talent, and then you just paid them what they were worth.

That metrics and statistics weren't a part of the game, and that was until Billy Bean and Bill James came around. So what they did was they instituted a system that found undervalued players at a much better price. And the question is, could you do something like this in taking care of your patients? How do you value your patients? How do you value your treatments?

How do you value the kind of care that you give? The story begins with Bill James, who in the 1970s, when the computers were just starting to be roll out, the Tandy computer came out. Bill James was a baseball geek. He was a college graduate who majored in literature, but he loved baseball and he started to count things and he used computers. And while he was a night watchman at a pork and beans factory, he published in 1977 something called Baseball Abstracts featuring 18 categories of statistical information you just can't find anywhere else.

It was an '88 page document. He was the inventor of something called sabermetrics. And what they did was they evaluated every player and every situation to give a numeric score to player performance. And then scouts and teams and GMs could use that. Again, this was heretical in the 1970s.

This was crazy talk in 1970s, but now it's standard common practice. Now all baseball, football, hockey, NBA, they spend millions of dollars every season on analytics to help them get their edge. This was evidence based baseball. Just like at the same time, we were going through something called evidence based medicine. It next begins with Billy Bean.

Billy Bean was drafted out of high school by the New York Mets because he had all five tools. He could hit. He could hit for power. He could field. He could run.

He could catch. You don't usually find that in a, high school candidate. Well, they drafted him. He passed on a baseball scholarship to Stanford. And you know what?

He never really found his way in baseball. He bounced around from the minors and the majors and never became a star. And it was because they overvalued Billy Bean. They didn't really get what he was about. Billy Bean loved baseball.

He stayed in baseball, and he ultimately worked his way up, through administration and became the general manager of the Oakland Athletics. And in his tenure, his performance relative to the budgets that he was given to spend was far, far, far better than any of his competitors throughout the league. I mean, he was way ahead of everyone else. And the question is why? Well, he started to believe in this the saber metrics of Bill James.

And it really came to a head in 2002 when the Oakland Athletics lost their three star players, Ingrahamy Ingrahamy, Donnie Damon, and Jason Giambi. Jason Giambi was a power hitting first baseman batting average three fourteen, 41 home runs, a 122 RBIs. And the question is, how do you replace Jason Giambi? They couldn't afford another Jason Giambi because at the time that he went, he was making a few 100,000, and now he was gonna go to the Yankees and make several million dollars. They just couldn't afford that with a 30 or $40,000,000 budget.

And what Billy Beane figured out was he was gonna replace Jason Giambi in aggregate. He was gonna get three players that were good, that were undervalued, that together, they would put together statistics like Giambi. And in fact, he did. He got his younger brother, Giambi, Dave Justice, and Scott Hatterberg. And these were undervalued players.

At that time, the scouts made all the decisions. The scouts, you know, found players who look like great players, had a great squint, swing, had a great arm. Even if they didn't perform badly, and they passed over these under recognized people as being fat guys, bad body guys, bad mechanics, you know, dating an ugly girlfriend, whatever it was, they didn't choose them. What you really needed, what they found out through Bill James and disciples of Bill James is that you needed runs, not players. You needed on base percentage.

You needed to acknowledge that a walk was as good as a hit. You needed to acknowledge that a bunt and a steal were exciting, but they actually were ways of losing. No small ball, no bunting, no stealing, big ball hits on base home runs. That and and then paying attention to the fact that baseball and hitting was more mental than physical at this level. The people that Billy Beane hired figured out to get in the playoffs, they needed 95 wins.

To get 95 wins, they needed to score a 135 more runs than they allowed. And how do you get there? You get there with statistics. They threw out tradition. They threw out biases, and they went with the data.

At the root of the Oakland A's experiment was a willingness to rethink baseball. Why am I talking about moneyball rheumatology? I think we're at a crossroads. I think you're all very, very good at what you do. I think that you could be better if you were to rethink the way you did your sport.

Bill James said, if you challenge the conventional wisdom, you'll find ways to do things much better than they're currently done. To me, this sounds like one of my quotes, which is if you keep doing more of what you've been doing, you're going to get more of what you got, meaning there's no room for improvement. And while you're doing great, what about that fifteen percent, twenty percent who are D2TRA difficult to treat RA or refractory PSA or refractory lupus. Again, the tradition was you relied on wizen crusty scouts who drove city to city smoking and drinking coffee and riding in station wagons. And again, they had these themes that they held onto.

They said things in the movie like, does he have an ugly girlfriend? I think he's got a fleshy kind of body. Yeah, but he's got a great swing. The player looks right. Sounds like when we say things like that therapy seems to work well for me and for most of my patients, when in fact it doesn't work well for all your patients.

Billy Beane said people don't know how to value baseball players correctly. You got to think about that. Do you really know how to evaluate patients? Yeah, you know when they're active, but do you really know how to value them, what they need and what the right therapy is? Billy Bean says, you can't change players.

You can teach skills, but you can't teach discipline. You can't change patients. You can educate them. I would say you can't change rheumatologists. You can give them skills, but we're really looking for here is a different kind of discipline that you need to employ when things get rough.

Again, the assistant GM who was the believer in the Bill James way of doing things was Paul V Podesta played in the movie by Jonah Hill under the name of a different name. Doesn't matter. He they said, you can buy wins by buying runs, not by buying the best baseball players. Again, we need 135 more runs, throughout the season to get to 95 wins. Billy Bean and Moneyball used a groundbreaking data driven method to actually analyze players.

That strategy changed baseball, sports, business. Why shouldn't it change medicine and rheumatology? Basically by upending traditional thinking. Why Moneyball rheumatology? Thirty percent of your RA patients do not only thirty percent of your RA patients will respond to methotrexate alone.

Thirty percent to forty percent of patients will fail to respond to your best therapy and best combo therapy. Treatment failure begets more failures. A patient who fails your best therapies, uh-oh, it's the beginning of a rocky, rocky road. They're more likely to fail the next time out and the time after that. So a Swedish registry also published that despite decreasing the mortality rates in RA, RA continues to be linked to an increased risk of death.

So yeah, we are doing better, but we're doing as good as the general population, but we still have more deaths. Moneyball is about the decisions you make and how you can be the architect of change. Again, if you look at the studies on what you do, this is the NORD STAR study, eight twelve early RA patients who got all of them got methotrexate and then one arm got cerdulizumab, another avatacep, another tocilizumab, another got traditional therapy, which was methotrexate and a combo DMAR with high dose steroids. And the first twenty four weeks, they all do the same. One is not better than the other.

And long term follow-up, they all do the same when you have a biologic and methotrexate. Here's another study that's using second line addition of, either TNF inhibitor, Avatacep, IL-six or JAK inhibitor. They all do the same when followed over three years. You don't know what your best drug is. I know it works for you and congratulations, but it doesn't work for everyone.

And the durability of these drugs is not as great as we want. In a sense, it just doesn't matter the way we're doing it. How can we make this matter? Well, look at the Ticora study, a study done in Scotland in 2,004, where they showed that when you use crummy weak DMARDs in combination, according to protocol and a treat to target fashion, that intensive treatment treat to target has an ACR 20 of ninety percent, whereas usual treatment has an ACR20 of sixty percent. But if you look at superlative outcomes, ACR70, DAS remission, it's four times the success when you treat the target.

This is a situation where what you do matters, not the drugs that you're using. That's what you're looking for here. Professor Nagy, who was a lead author on the difficult to treat EULAR guidelines, recently published an article in Lancet Rheumatology, where he says, despite considerable therapeutic progress in recent decades, treatment choices are often guided by a trial and error approach, inevitably bringing about a difficult to treat state in a substantial number of patients. It's a significant minority. Again, we have a very heterogeneous disease.

He's calling for stratifying patients into distinct phenotypes that could lead to more personalized management to treat and or prevent difficult to treat disease. So the questions we have in rheumatology, can you achieve better results without spending more money? What's your edge? What's your different approach? If you keep doing what you're doing, you're going to get more of what you've got.

Maybe you should use data analytics and AI in your practice to change the outcomes. Can data actually outperform your experience and intuition? Or can it add to your experience and intuition? Basically, we're doing hunch driven expertise. And we're saying a money ball, I'm saying a money ball approach will change that in the future for the drugs you choose, the patients you select and the healthcare you deliver.

So what is your performance metrics that you're relying on? You're obviously not doing ACR20, but you can collect data, you can understand the situation the patient's in. And what's the value proposition that the patient is looking for? Is it pain? Is it work?

Is it their avocation? In Moneyball, the book and the movie, the goal was to win games. They use data to better characterize patient performance or player performance. That was then used by the GM who passed it on to the coach, who then selected the right players and developed a strategy to win games. The net result being more wins at a lower payroll, benefiting the city, the fans, the owner.

How does that apply to you? The goal is remission, whether it's RA, PSA, lupus, spondylitis, doesn't matter. What are you gonna use as coach or and the GM to make better selections as far as the drugs, the players or the patients, the strategy. Is it data? Is it a systemic change in how you practice?

The net result has to be greater efficacy with better safety. It's going to benefit the owner who is the patient. Many, many years ago, IBM chose as their slogan, think, That's why you have think pads and think whatever. And then as a competitor to IBM, Apple came up with, that's right, think different. That's what you're being called to do.

So understand that one drug doesn't fit all, that we are now in the money ball era of EHRs, computers, and data collection. You should be a part of that. How can you reevaluate your patient's drugs, the clinical trials you look at and the endpoints? And where is the value in machine learning or AI? So where's your edge?

And I'm going to use a simple example here. I'm in Dallas. I'm driving around. I'm about to run out of gas in my car. I need to get gas, but I'm kind of frugal.

I wanna get the best deal. My cousin says I should go to the Shell station on this corner. And I'm driving around and, you know, I can go to that Shell station because it's not too far, or I can break out my app called GasBuddy. And I can see where I am and where the cheapest gas is mirror where I am. You need a gas buddy to help you in practice.

We should obviously, in the past, we've always looked at clinical trials, and clinical trials are great. But are we measuring the right things? Number one, a clinical trial is designed to prove a marketing intent that this drug works for that condition. Let me sell the drug. It's done on a population that's very unreal.

It's very homogenized. They have no comorbidities, and they have to adhere to the regimen. The disease that you're treating is systemic, but the outcome of the trial is often a specific endpoint that's organ specific. And the problem is that the measures you practice are not the same measures that are used in the clinical trials that came from Ted Pincus twenty years ago. We don't do trials that have real world data.

Globalization of trials have changed placebo rates and the results. And again, the priorities of the patients are different than the priorities of doctors and are different than the priorities of the companies trying to get a drug approved. So I like this study. It's using machine learning in the ceruleumumab drug development trials at four trials that were done for development. And what they did was they put together all the data that they collected in their trials and they tested it on one dataset and then proved that it worked on another dataset.

As you know, cerulium mAb is an IL-six inhibitor. IL-six is directly then related to CRP levels. But studies have shown that CRP levels don't predict who's going to respond to IL-six. When they did their machine learning analysis, they found the rule that predicted better outcomes was being ACPA positive and is having a CRP of greater than 12.3 milligrams per liter. In the trial without the rule, sixty seven percent of people responded with an ACR20 response.

But if they implied this rule only to the people who were rule positive, ACMA positive, high CRP, that sixty seven percent became eighty one percent response. So this is just an early exercise. There are many biomarkers out there, they're not very consistent in what they show. There are many studies out there that look at AI to predict better outcomes in clinical trials. And here's a bazillion of them.

And while they all made a point that XYZ was a good predictive variable to enhance responses, they're not consistent. Of course, are different studies done in different continents, done different ways with sometimes different outcome measures. But we still need more work on how to use data to help us get better outcomes. So in one study, R4RA study, they looked at synovial biopsies. And they tried to look at patients.

And in a study, patients were either going to get tocilizumab or rituximab. And their study that had forty percent non responders. You couldn't actually predict who was going to respond better based on the histology being rich in what looked like B cells or whatever. But when they did RNA sequencing and develop different pathotypes, they said that certain biopsies were pauci immune, mainly myeloid, some more lympho myeloid, and others that were not. So they then showed that the people who had no B cells, had no rituximab response, people had B cells, had equal rituximab and tocilizumab responses.

And then the ones who didn't respond had a predominance of fibroblasts and stromal cells. The point being was that the RNA seq analysis of synovial biopsies was limited in its ability to predict. And one of the but it did help some. But the real big finding here was that there's a big percentage of people who had these fibroblasts and stromal signals, and that's the one we're not treating. That was a lesson from here.

So again, it's refractory RA, difficult to treat RA. Is it the wrong target? Do you keep jumping around? Or do you recognize that it's fibroblast driven disease and that we need to treat differently? IL-six works well there, but we need other drugs.

Is it lifestyle and modification? Or is it damage done that can't be undone? So in a meta analysis of machine learning to predict responses, 89 studies, most of them being in RA, they looked at EMR data, biomarkers, genetics, proteomics, different omics, imaging. And they showed overall an accuracy of sixty percent to seventy percent with an AUC of 0.6 to 0.9. But the problem is tremendous heterogeneity, varying methodologies.

And the bottom line that you could take from this is that this will become a bigger thing in the future, but we're going to need an interdisciplinary approach from rheumatologists and data scientists to know what that edge is in any particular patient. I predict in ten years, you're going to have an AI tool, a robot, if you will, that will make you better at diagnostics, daily practice choices, better treatments that are better tailored to that patient situation, certainly managing your time and your education. Realize that we in medicine have been spending money as if we had the budget of the New York Yankees. When really the signs are that we are really operating more like the Oakland Athletics and we need an edge. So while Moneyball rheumatology could give us better high value care at lower costs, it may also give us smarter choices for individual patients.

And also along the way, give us better safety profiles. So is this I don't know and I don't care, also known as ignorance and apathy? I don't think it is. I think the time has come. I wrote a blog on this almost ten years ago, where I wanted you to acknowledge that rheumatology and baseball are an unfair game.

And that it should no longer be business as usual. It should be an adapt or die. Can you advocate for numbers and data driven decisions? Why aren't you a card counter at the blackjack table in Vegas? I don't think you need a new drug.

I think you need new practices and I think you need new data to help guide you. So I'm going to end by saying change is good. And here's some ideas that you can employ in a moneyball fashion that can give you better outcomes now while you're waiting for AI to make that a big leap for you. So find better patients, build better patients, educate your patients better. The main changes you should do is hire three nurse practitioners or physician assistants to deliver better care.

Number two, put everybody on hydroxychloroquine. I wrote a blog on that last week, look at it. Because it works in RA just like it works in lupus. Change the flow and the workflow in your office and how you evaluate patients before they get to see you. Use medical assistance efficiently, use forms efficiently.

So you spend more time talking and listening face to face with the patient. Collect data, metrics and outcomes. Don't worry, they may not be fruitful right now, but the day is soon coming when AI will step in and make you smarter at your job. You're already smart, it'll make you smarter. It'll make you more efficient.

You should have time goals, earlier referral visits. How are patients going get to see you sooner? Because right now, most of you have ungodly wait times for your next appointment. You should be doing shorter trials of the drugs that you start, not waiting three or six months, making decisions in four or six weeks. You should be risk stratifying your patients.

This is high risk right from the start. For instance, all women with SPA and PSA are high risk. They just don't respond as well as the men. What are you going to do differently? And then tailor therapy.

And then ultimately, you're going to hire an analytic computer crunching geek who's going to make you the better doctor that you already are. This is my approach with Moneyball rheumatology. A few quotes to end, Mickey Mantle said, It's unbelievable how much you don't know about the game you've been playing your whole life. Next, Sir William Osler said, Medicine is a science of uncertainty and the art of probability. I'm going to say the money ball is going to change William Osler's words to medicine is a science of probability and a science of uncertainty.

And then Margaret Mead said, Never doubt that a small group of thoughtful, committed citizens can change the world. Indeed, it's the only thing that ever has. Thanks very much. I hope you thought this was useful.