The Doctors’ Lounge
Where scalpels meet systems — and physicians say what they really think.
Co-hosted by Anish Koka, MD & Anthony DiGiorgio, DO. Candid talks on healthcare policy, reform, physician autonomy & patient care.
The Doctors’ Lounge
Gaming the System: LTACHs, Guidelines, and the Evidence Problem in American Medicine
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Episode Summary
Dr. Anil Makam — hospitalist, health services researcher at UCSF, and faculty at Zuckerberg San Francisco General — joins Drs. Koka and DiGiorgio for a wide-ranging conversation on the hidden mechanics of American healthcare. Makam breaks down long-term acute care hospitals (LTACHs): what they're for, how regional variation and perverse payment incentives have distorted their use, and what the 2016 site-neutral payment reforms actually did to the market. The conversation then shifts to Makam's research on clinical practice guidelines — specifically his 2018 study showing that the majority of ATS recommendations were grounded in low-quality evidence, many carrying strong designations anyway — and what that means for how clinicians should read and apply guidelines at the bedside. The episode closes on the FDA, indication creep, the limits of central planning in quality measurement, and what it actually means to be a good doctor in a system where you can't buy your way to better medicine.
Chapter Markers
00:00 Introduction — Dr. Anil Makam, UCSF hospitalist and health services researcher
02:09 What is an LTACH? Origins, optimal use cases, and the vent-weaning niche
08:09 How clinical practice led Makam to study LTACH utilization
10:08 Geographic variation in LTACH use — decomposing what drives it
14:16 Post-acute care economics: DRGs, payment systems, and perverse incentives
19:11 Medicare Advantage denial rates and the two-tier access problem
23:06 Market access vs. total closures: what the 100 LTACH closures actually mean
24:04 Short-stay outlier rules and the "magical recovery" at the payment threshold
26:07 Site-neutral payment reform and its effects on the LTACH market
31:51 Moving to guidelines: evidence vs. recommendations
33:38 The ATS guidelines study — what they found and the Twitter fallout
39:34 How to practice when most of what we do lacks strong evidence
43:38 Why guidelines are getting more confident on less evidence
47:10 The generalist vs. specialist lens on evidence appraisal
53:47 How do you measure what makes a doctor good?
56:41 Three buckets of physician quality: technical, relational, cognitive
01:00:06 Running a trial vs. appraising a trial — two different skills
01:05:16 Indication creep and applying trial evidence to the wrong patients
01:09:24 The FDA, Vinay Prasad, Marty McCary, and why reform failed
01:13:45 Wrap-up and where to find Makam
Co-Host Handles
@anish_koka and @drdigiorgio
Show Handle
@drsloungepod
Subscribe Links
Spotify: https://open.spotify.com/show/44vw8eirsKKnjgNIrdDvrR
Apple Podcasts: https://podcasts.apple.com/us/podcast/the-doctors-lounge/id1832097658
YouTube: https://www.youtube.com/@TheDoctorsLoungePod
🔗 Connect with the Hosts:
If you like your healthcare plan, you keep your healthcare plan. Nobody knew that healthcare could be so complicated. It finally beat Medicare.
SPEAKER_03This is the Doctor's Lounge. Alright, welcome to this recent edition of the Doctor's Lounge, the the latest. Hot off the press. We have Dr. Makham. Anil Makham is a hospitalist, health service researcher at UCSF. He is on X. And as we love having opinionated people on X on here. That's kind of like the vetting. Like how do you how do you figure out who to get on here? Well, you gotta show us what you uh you know what you're doing uh uh on X. And so anyway, Anil, thanks thanks for coming on. He he actually trained uh right next to me. He trained at Drexel, uh, then did his uh which is in uh Philadelphia. Um am I getting that right, Anil?
SPEAKER_00Yes? Yeah, I was uh Hahneman when I accepted, and I became uh MCP Hahneman when I accepted, and it became uh Drexel when I started.
SPEAKER_03Uh okay he did it, he did his residency at UT Southwestern, did a research fellowship in UCSF, got a master's in clinical research, and you spent your career at the Zuckerberg San Francisco General Hospital, right? It's a safety net hospital, which is the you know, which is where Dr. DeGiorgio is as well. Uh he cares for patients who are uninsured, undocumented, living with addictions, really a you know, as I said, a safety net hospital. He's published a fair amount as well. Uh he's uh his he has this framework for practicing EBM uh that was in circulation, a cardiology journal, like how do we let a hospitalist publish in circulation? What's going on here? He peer-reviews MedPAC reports to Congress and also contributes as an evidence author for ICER. Um he also has a substack called A No Times 2 that looks like he started in uh in February, which has some pretty interesting uh posts on it as as well. So Anil, welcome to the Doctor's Lounge. Thanks for having me. Anil, tell me a little bit about when I was when I was prepping for this, I saw that you had done some really interesting work when it relates to LTACs. Um can you can you talk to us a little bit about that and how you got interested in that? Yeah, so my I guess we should we should tell people what LTACs are as well.
SPEAKER_00Yeah, I was gonna say, let me give uh most people don't know, but um depends where you practice, actually. So I split my career in um Dallas, which is uh is probably still is uh an LTAC hotbed. This stands for long-term Ecuare Hospitals, and then I've spent uh probably a little bit less of my time here in San Francisco where there's very few, and we probably underuse LTECs um because there's only one in the city, and there's uh one uh close to where you live, Anthony, Kentfield and Marin. And there's one in the East Bay, so kind of servicing the entire uh Bay Area. Um there's like three LTECs. And when I trained in uh Dallas practice in Dallas, there was 25. Um, and they were a normal uh part of post-acute care. It was just a normal thing where people went to go get their antibiotics, to get their dialysis, to get wound care, um, and uh sometimes to go for um critical care and uh event weaning, which is kind of their uh clinical niche. Um and then as you said at the beginning, I trained in um Philly. Um you you probably know more, Anish. Uh I don't remember LTACs being a major part, but it was also a naive uh third and fourth year. I honestly don't even know if I remember what the heck medicine was back then.
SPEAKER_03No, yeah, there are there are some actually uh my first experience uh with an LTAC, but I'd never heard of it, was when I was a resident at Temple Hospital. Temple is is like fantastic pulmonary program headed by the still still going strong Jerry Kriner. And he was one of the first people to introduce this concept of um folks that were had had a tracheostomy, and the traditional model was hey, just keep them in the ICU until until, you know, uh and then tracheostomy and then go to a some type of skilled nursing facility. And he he was one of the uh one of the first people to kind of introduce this idea of hey, we should we should have a a facility where we can actually have a lot of respiratory, you know, high, high uh acuity ability to deliver respiratory type care so we can actually wean people uh from you know from other chronic vent situations at this LTAC type facilities. So that that was my experience.
SPEAKER_00Can I give folks a quick uh 30,000 foot view of what the heck is an LTAC? So it's a long-term acute care hospital. It's kind of a doxymora. Uh you got long-term and you got acute in the same uh phrase, but it kind of makes sense. So these are people typically who've been in an acute care hospital. So your traditional hospital, where I practice, Anthony, you practice Anish, if you see patients in the hospital, they're usually in the hospital for two to four weeks, usually before they're even going to an LTEC. Um so they survive this really critical illness phase and sometimes subacute critical illness. And then they're in this recovery phase, but they still have like a ton of inpatient uh needs. And I'm talking about like the optimal use case, and we'll we'll talk about the more suboptimal use case and kind of what I've been studying for the past uh decade or so. Um so in the optimal case, you have someone really sick, um, often have uh multiple inpatient needs. They're usually multiple infusions, those are antibiotics, they're getting wound care. Oftentimes, um, you know, the clinical scenario is people um with prolonged mechanical ventilation for respiratory failure. Um, they're on a vent, um, often through a tracheostomy through your neck, and they're getting weaned, and um, that's kind of their niche. And then you go to an LTEC, they have a usually an ICU, they have respiratory therapists, they have uh subspecialists who round every day. People like me often staff an LTEC, a hospital medicine physician, they have their own imaging on site, usually, CT, uh, they can do procedures like bronx and uh EGDs, and you're getting multidisciplinary care. And uh the strength of an LTEC is you're getting um wraparound PT, OT, speech therapy, RT. They talk about you in this multi-cool multidisciplinary rounds every week that we don't do in the hospital very well. And they're really focused on uh recovery, rehab, um, in addition to the clinical side. And that's their like intended purpose. They evolved out of like TB sanatoriums with respiratory failure and then kind of been morphed into more just complex care NOS in the past couple decades, like two or three decades.
SPEAKER_03So, yeah, like I said, it was it was kind of uh wondrous to see uh again in the optimal use case, you had these patients who really weren't served well from going from an ICU setting straight to a s a SNF setting. So this this kind of ability to deliver much higher level care than what you could at a SNF, you know, in terms of pulmonary rehab, in terms of respiratory therapists, in terms of uh, you know, pretty experienced pulmonary attendings kind of rounding on these on these patients. And uh and it was really remarkable to see how many uh of these patients were able to be weaned uh actually. A lot of patients were, you know, at trachs and um um were were on chronically on vents who were going to the the this LTAC. And uh really remarkable stuff. I mean uh at the time, you know, this idea of rehab for for this group of folks wasn't really a well-developed uh uh concept. It's like what do you what do you mean you have a trach and you're you're trying to rehab? Like what what what does that mean? And so that of course uh now it's kind of standard, but but it was really, really remarkable uh to see. Sound sounds great, Anil. Tell me tell me what what kind of observations that you had that kind of made you decide to study this a little more with regards to how LTACs were being utilized.
SPEAKER_00Um clinical practice. Honestly, most of my questions come just from like what I'm seeing and what I'm doing, sometimes what I'm reading. But um, and uh this is often while I tell uh trainees or other faculty, like it's right in front of you. The questions, like what to ask, how to answer them is a different issue, but um at least what to explore. Um so I I practiced in Texas, North Texas, Parkland is a large safety net, four times the size of a Zuckerberg hospital, the San Francisco General is what I call it. I de Zuckify it, but uh huge hospital, but it was just a normal component. Like you would just be like, this person recovered from their hospitalization, they need six weeks of IV antibiotics, let's send them to an LTEC. This person needs, you know, a uh negative pressure wound therapy, wound vac is what we use for short, send them to an LTEC. It was serving almost similarly to what you would uh send people for uh skilled nursing facilities. Um then I did my research fellowship in San Francisco and they were non-existent. Um, didn't really think much about it. Um, came back to Dallas for faculty, so I ping-ponged. Um and kind of early on, I was just like, oh yeah, these LTACs exist and we're sending all our patients here. Uh why? Um I was taking care of pretty similarly sick older adults, is usually um who's going to these, but younger people go too. And it made no rhyme or reason why one market we practice a certain way and another market we practiced a completely different way. And it wasn't clear to me that there was much difference because there's fewer of these kind of more optimal situations, like the person on event, critically sick, often getting multiple infusions. They exist, but there's just fewer of those types of patients. Uh but it was just a really expanded use in Dallas. Um, and that kind of uh led to a series of studies.
SPEAKER_03So then this this is kind of harkens back to Elliot Spitzer's work in the Dartmouth Institute, right? Like this this interesting observation that there's these really large geographic variations in utilization of healthcare services. And the question is, what exactly is is driving that? So what what did you so what what were your I'm curious about how you set up that. Like what what did you think based on you being on the ground and you know being in these two places, right? Like what was your hypothesis? And then how did you do you come up with designing something to try to answer that question?
SPEAKER_00Yeah. So I think initially I thought if you don't have an LTEC, it's hard to use an LTEC. So if you just look at the map of where LTECs are, it kind of tells you that certain places use them and certain places don't. So I kind of are and then clinically knowing you know my experience in two different regions, um, but obviously there's many regions across the country. And my hypothesis was it was very regional. And so what I did was one of my first studies, and uh we published this one in JM Internal Medicine, just try to decompose the variation, like how much of the variation is because people are sick, how much of it is what hospital you're at, and then how much is just what region of the country you practice in. So we set up a multi-level, we call it a multi-level or hierarchical regression model, you know, use Medicare claims. And the advantage of the decomposition is you don't have to adjust for all the factors that I care about as a clinician, like how sick I don't need their Apache scores, I don't need this or that. I can just look at how much of why LTECs are used is clustered at these three levels. Whether I can explain it is a different issue. And I think we were able to explain some of it. And um, you know, in the variation world, LTAC use is probably one of the most regionally varying practices in medic, all of medicine. So we found about half of the variation is just what region of the country you're in. Uh, I mean, uh sorry, half is the patient how sick they are, which is really low for most things. So when I'm telling July interns or junior faculty who just joined our group, over 90% of why a uh your patient in the hospital will die has nothing to do with you. It's just how sick they are. That last five or 10% really matters. And we'll talk about that later. But most of why people have bad outcomes is not related to necessarily what we can change as hospitals and clinicians, which is humbling, but we also have a lot of power in in changing those outcomes, especially in screwing things up. But so we found about half of that variation was between patients. So the reason why someone goes to, and and our comparison here is going to a skilled nursing facility, which is the principal post-acute care comparison. The actual real comparison, and this one is really hard to study, is um you don't go anywhere, you just stay in the hospital longer. So it's an unobserved counterfactual. Your hospitalization is usually a week or two longer. But if you do go somewhere like in Dallas, like these people would have just gone to a sniff otherwise. Um half is how sick the patient is, a third is what what region you practice in the country. So I practice in Dallas, you practice in Philly, now I practice in in Anthony practice in San Francisco. That explains a third of it. And then another 15%, which is not trivial, is what hospital you're in. And we have a nice waterfall plot where we look at hospital variation within the same markets and pretty big differences. And this has to honestly do with culture relationships of one doctor. And often the doctors who are practicing in the hospital often are rot are rotating or staffing patients in the LTEC. They have a relationship, then relationship with the case managers, and it just takes a couple of those people to shift uh practice culture. Um, that's my guess. I couldn't tell you for sure, but uh there's a ton of like variation within the same market, just what hospital you practice at.
SPEAKER_02I noticed that same geographical variation coming from Louisiana. Um in Louisiana, the the big thing we had was a lot of inpatient rehab, you know, dealing in the in the traumatic brain injury and spinal cord injury world. A lot of my patients uh go, you know, either SNF, the nursing home, or inpatient rehab, or occasionally to LTAC. So we kind of see them go all three. And I noticed that when I came to San Francisco, just like you did, that it's very SNF dominant, right? So most people go to their nursing homes. Um, so I I've also tried to look into this. I I totally agree with you that it it's so much of it is local relationships. If there is, you know, one of these post-acute care facilities in a vertically integrated hospital system, they tend to, you know, prefer sending them there. Uh, I think certain managed care organizations may prefer uh sending their patients to certain facilities they have relationships with. Um, how much have you looked into how the finances are? Because a lot of this, the the payment for these post-acute care facilities is very complex because a lot of these patients are dual eligible Medicare and Medicaid. So you have the regional variability of Medicaid, and then you have the sort of flat fee setting from Medicare. So have you looked into any of that? I know you and I have talked about this kind of on the side.
SPEAKER_00But you look into that with the LTEC. Um when I first got into this, I wasn't reading academic papers, I was sifting through the federal register. Um, if you want to understand how healthcare is done, you don't read uh an NEJM perspective. You read the Federal Register of like what are the policies, what are the incentives. Um, and there's very little published on LTEC. So I was sitting here reading MedPAC reports, I'm reading white papers by RTI, I'm reading the Federal Register because you really have to understand how they're reimbursed. Fee for service Medicare is the primary driver and has shaped the field in terms of how other payers follow, but they're very different. So a lot of my work is in Medicare fee for service because of convenience, not because they're the only shop in town. Um we have the best data, we have national data, and they're the reason LTACs are expensive is they get a separate bundled DRG payment like a hospital gets, and they're separated from the hospital. So you you go to the hospital, you get one DRG lump sum payment. Obviously, uh there's like local market share corrections and outlier corrections, but you basically get a bundled payment that you expect, and then you get discharge.
SPEAKER_02And if you go to an LTEC, based on diagnosis, right? For those that don't know, diagnosis.
SPEAKER_00You get you get lumped in this diagnosis-related code. They have exactly they have like every year they update like how much that DRG is worth, and then there's uh variation by your market cost and you know, if the patient wasn't a short or long stay outlier. And LTECs get their own DRG system payment. So the reason it's so expensive to Medicare is now you're paying two lumpsum payments and it's not shared across. So you can see how that gets expensive. It's not expensive if you think about your substituting a day in the hospital versus a day in the LTEC. An LTEC is going to be cheaper than acute care, but still fundamentally much more expensive than a SNF because you have daily physician care, you have ICU, you have the staffing, the medicines procedures that you're doing, but it's less expensive than an acute care hospital. But by the nature of Medicare reimbursement, is what makes it especially uh expensive. So it's weird. LTACs and fee for service is one of the most preference sensitive decisions and all of post-acute care is. Post-acute care as a whole, so we're talking SNFs, home health, inpatient rehab. And this was a little bit dated. I think it's a little bit attenuated, but it's still the number that people cite is about three-quarters of all variation of Medicare spending is post-acute care, which is a dramatic, dramatic amount. If you think about all decisions made in outpatient, inpatient, all different settings. More three quarters of why a different market is a high spender, low spender is their post-acute care use. LTACs are small. They're one to two percent of all post-acute care stays. They're the most expensive, though. They're like 30 to 50K a stay on average from Medicare fee for service. But if you look at Medicare Advantage, it is one of the most managed uh benefits in all of Medicare Advantage plans. Um, there's not a lot published on denial rates and prior authorization um uh outcomes, but um there was a recent uh Senate uh OIG investigation and um they had the three largest Medicare Advantage plans. And they got some of that data and they looked into it and um across, and it's more for uh LTAX and inpatient rehab hospitals because they're the most expensive. Um, they get denial rates of like 60 to 80 percent. Uh, and the overturn appeal rates are often not successful. And the most common reasons are not indicated for this level of care, essentially. Um, they're not able to do that kind of rehab, they don't have that rehab potential if you're going to a rehab facility, or they're not sick enough to warrant going to an LTEC. So it's weird. You have like you can have an older adult if fee for service versus Medicare Advantage have completely different um insurance um pathways to the care they're getting um in the same same market, same region.
SPEAKER_03So just just so just to rewind, just to summarize what what you guys have said, because it's uh you know, I I I don't I would I don't want people to get lost, and I think it's really, really, really, really important. So LTACs are paid under their own prospective payment system, right? Um it's structured around DRGs, but there's specific what's called uh long-term care DRGs, right? So these are these are di uh diagnosis related groups that are specific to long stay patients. How do you how do you differentiate long stay from not long stay patients? Well, Medicare says if it's greater than 25 days, right? And you know, this this that matters because the the LTAC prospective payment system pays higher rates than the standard inpatient prospective payment system. So there they're that so that that that creates a certain incentive, a certain pervert perverse incentive, right? Uh so take patients who don't actually need that significant level of care, keep them long enough to capture the full DRG payment, then then discharge them. Am I getting that?
SPEAKER_00And then there's another DRG uh payment. And there's not a lot of clinical criteria. One, it's hard to define. So what you said was the only real criteria before 2016 was does an LTEC's average length of stay 25 days? Right. Not that patient. That patient could be there for two or three weeks, but on average their population 25 days as a result of like rapid growth in the LTEC market and um some of these situations that I described before, like people going who could have gone to a sniff. I I I I did a study in uh in an ACO in North Texas, and we actually had that electronic heart re health record uh data, and we actually looked at the reasons why people are going. Some people like went for really absurd reasons, and you're getting this massive like DRG lump sum payments. Um, but that's like one market that has a ton of LTECs. I think in other places like San Francisco, the people who go to an LTEC are really, really sick, and we probably underuse them for people who are not on event but also sick, because I think you get better care than sitting in an acute care hospital getting seen by a therapist for like 30 minutes, two or three times a week. Um they're way better focused at um kind of that rehab, I mean that uh late phase of an acute illness, illness. Um, but anyways, in 2016, they um implemented some uh basic uh clinical criteria, which I think are pretty reasonable. You had to either survive an ICU stay for three or more days, um, or you had to be on a ventilator um and you couldn't go for a rehab or psych diagnosis as your primary diagnosis. Um and that was it. And that since 2016 has led to probably a hundred out of four hundred-ish LTEX closing in the last couple of years. That alone, uh, that change. And uh there was uh recently uh I was quoted in an Axios, Axios, is that how you say that article about uh there's uh a couple lobby LTEC organizations. One I work with who are really great, uh, the National Uh Association of LTEX, they're mostly um nonprofits and really thoughtful uh group of folks who who are trying to use LTEX more appropriately and kind of lobbying uh Congress to repeal that uh policy um because of uh the impact it's had. And I, you know, I think the main thing is not total closures, it's uh market access. I think do you have access to an LTEC if for the people who need it rather than oh, a hundred LTEX closed nationwide? Because if you're in Dallas and 10 of those LTEX closed out of 25, you're fine. You probably still have five too five to ten too many LTEX. But if you're in San Francisco and one of one or two of the three LTEX closed, that's a whole different story. So it it depends on that regional uh availability.
SPEAKER_03Aaron Powell So it's a little bit I just want to it's a little bit of a game of whack-a-mole, right? Where you have these clear perverse incentives, people reacting to those perverse incentives, and then you have the um rules that Medicare comes up with. So one of the rules was the short stay outlier rule, correct?
SPEAKER_00Uh one of them, it's uh essentially you get a lower payment. So there's uh there's this incentive to keep people a certain amount of time. And and we did this work, and other people have done this work, and there's this magical recovery where you cross and each DRG has its own threshold magically. You know, I'm sure this is how God intended uh human bodies to recover. Say the the outlier for a diagnosis is day 10, you know, there's a steady discharge rate. There's people going up, you know, they can't prevent people from not leaving, but then like day 10 happens and boom, this spike happens, and like not everybody, but it's like a noticeable spike where all of a sudden these people get discharged so that you get the full payment and not a short stay outlier payment, which is much, much uh less.
SPEAKER_03Right. Right. And then and then there was the 25% threshold rule, correct?
SPEAKER_00Say that threshold one again. So this is the same thing.
SPEAKER_03It's separate from the SSA rule, SSO rule, but it cuts LTAC Medicare payments to an equivalent amount under the inpatient BPS for patients. Oh, for the short if you're a short stay. Right. Right. Yeah.
SPEAKER_00It's the lower of of the per diem cost or what the I think the inpatient stay costs. Uh don't quote me on that.
SPEAKER_03So if a single if a single acute hospital is sending more than 25% of its discharges to a co-located or affiliated LTAC, that arrangement looks Oh yeah, yeah, yeah.
SPEAKER_00That's right. That thing. I'm not sorry. You can't self-refer to a facility. So Anthony, to your earlier point, like vertical integration, they actually have to have separate management. So you could be half of the LTECs are co-located in an acute care hospital. There's just like a hospital in a hospital, they call it literally a hospital within a hospital. It's like a floor. But they have a separate management structure. But you know how that works when you're actually physically located. You could have a separate management, but often the same consultants are walking across, taking the elevator and coming and seeing your patients. So you can't send, you can't have a concentrated amount of patients from one acute care hospital.
SPEAKER_03Right. So they came up with a special saying, if it's more than 28%, that looks shady and we're gonna we're gonna have a have an issue with that. And then what you were talking about is the site neutral payment layer, right?
SPEAKER_00And that's that's that's what you're talking about in terms of and that's been the most dramatic policy change in the last decade.
SPEAKER_03Correct, correct. And so if if um uh you know so basically you get the lower site neutral rate if immediately preceding acute care that has at least three days in ICU or that required mechanical inhalation for at least 96 hours. So criteria, right?
SPEAKER_00Not enough to make it profitable for LTEX when you lose uh 40% of your patients uh no longer getting that higher bundle payment. Right.
SPEAKER_03So, Dr. DeGiorgio, site neutral payments in this particular case, was this a good thing or a bad thing? I mean, this is what we this is what we we bang on about for a long time. This potentially resulted in a lot of LTACs closing. Anil has already said that, you know, if it's in Dallas, no big deal, but if it's in some smaller market areas where you needed you needed an LTAC, now folks are going without. So is this the market winning or is this a failure of central planning?
SPEAKER_02Well, the whole thing is a failure of central planning. I mean, we have it's just to like take a step back, right? So you have these rather arbitrarily defined categories of inpatient care, right? Acute inpatient hospitalization, LTAC, SNF, and ERF or in inpatient rehab. These are all inpatient facilities taking care of sick people who cannot go home, but they're they're all arbitrarily defined by essentially by Medicare. Like you said, Anil, you learn a lot more reading the Federal Register, and they all get paid differently based on what sort of facility they are, right? So the the acute care and the LTAC both get paid on this DRG. LTAC has a different DRG system. The SNFs get paid on um a per diem, if I'm not wrong, right? Yeah.
SPEAKER_01Or per diem, up to 21 days before co-kicking.
SPEAKER_02Right. For Medicare. And then if you're covered by Medicaid, then Medicaid will pay longer for SNF, but they pay uh for every day that you're there. And then inpatient rehab is paid on a uh a rehab DRG, basically, what they're what they're there getting their rehab for. So again, spinal cord injury or if somebody had a hip surgery and they need some rehab, they get paid on that DRG. And so, but this is all completely arbitrary based on on how these institutions kind of grew up. And there's a lot of historical, a lot of historical nuance into how these these distinctions were made. But there's nothing that says that you should have a SNF paid on a per diem versus an L TAC paid on a DRG. And then you get managed care, which enters the whole mix. And like you said, now managed care wants to deny the LTAC because for managed care, now they got to pay a separate DRG. So you could see why Managed Care would have an 80% denial rate because they don't want to pay two DRGs for one patient. They'd rather just pay the one DRG, keep them in the acute care hospital longer because they're going to get the one sum that they pay the hospital instead of paying two DRGs or getting them to a SNF where the per DM rate is going to be lower than LTEC. So again, it it's all centrally planned and it's all a failure. And it's really hard to even conceptualize what a free market would look like in this because this really is a market failure, right? These are people that can't really shop around. There's yeah, there's very little, very little ways for them to compare and shop, and and there's not going to be that much price sensitivity, although you know, once you start in eating into people's savings and assets, then maybe there is a little bit. But I don't know. This is just such a big mess. And this, I think, is one of the most difficult parts of healthcare to manage from a pre-mark pre-market perspective because it's just it's so convoluted.
SPEAKER_03Yeah. So you had you know, you you you studied this, uh, Anil, in a 2018 paper um looking at the effect of um set mutual payments uh you know, with this criteria. I mean, the results are kind of somewhat stunning, right? You know, you showed 40% of you know, the policy hit 40% of all LTEC emissions. That's not like five percent.
SPEAKER_00It's forty percent all take a lot of technology. That's why a hundred LTACs closed.
SPEAKER_03Right, right. But as I was saying, the the policy kind of hurt, you know, it yes, so it hit the high the uh high uh the higher use markets, but it also hit the markets where there weren't a lot of folks that uh there weren't a lot of LTACs that existed. So potentially folks were were hurt by this.
SPEAKER_00Is that would that be your possibly I think I haven't looked into this recently, but uh the med the med pack analyses access looks reasonable. Distance to your closest LTEC seems to be largely preserved. You know, it's probably not a hundred percent. There's probably some places where um one LTEC was um kind of serving a region and they couldn't maintain it. But you can kind of think that there's only so many of these kind of patients in a market. We're not seeing people on a vent for weeks and weeks, and like this is not a normal, regular kind of population that's like running through hospitals, um, even through COVID, you know. Like most people recover. You either die or you get better enough. And uh there's a small number of people who are kind of in this uh intermediate state of just prolonged bad illness. All right.
SPEAKER_03So LTACs are are a net positive, but it's still it's still it's extremely challenging to sort out how best to do it. And and it just goes to show you the extreme complexity that exists when you have third parties paying because you get tons and tons of gaming of the uh of the system. Um all right, so moving moving to moving moving on from policy and prospective payment systems, I wanted to talk to you about one of your other passions, which is guidelines. Do you love guidelines? I love evidence. I will not say I love guidelines on this on the show. Yes, exactly. So you've made you've made the point that uh that guidelines and evidence are not the same. You know, in uh uh when was it? It was 2019, right? Uh in 2019 2019, that's right. 2018 or 2019?
SPEAKER_00Uh I can't even remember. It might have been 20, I think it was 2018. It was when I joined uh I was finishing it up when I was uh I was mentoring uh pulmonary fellow and I was at Southwestern. And then I was published when I joined uh um uh San Francisco General Hospital. So it would have been 2018, I think.
SPEAKER_03I see. So you just you'd just maybe started. And so you published, you were clearly again, uh, you know, informed by your what you were seeing clinically, you were frustrated, you must have been frustrated by the American Thoracic Society guidelines. ATS is you know the one of the pulmonary pulmonary guide, one of the main pulmonary societies, the American Thoracic Society, and uh you published something that irked pulmonary pulmonologists. Can you tell me, tell us about what it was about the ATS guidelines and what you what you published?
SPEAKER_00Yeah, it wasn't actually any skin in the game against uh ATS. It was uh I had a fellow who wanted to work with me and I was thinking about this idea of uh what's the evidence underpinning um guidelines in general, and it was kind of a fortuitous thing. And it's and it's a field I think we don't have a lot of great evidence. Um honestly, how I practiced critical care two plus decades ago is not that different to how critical care is practiced today. And how COPD and asthma are managed are also not that different to how it's managed today. Um so it was kind of the confluence of not a lot of good evidence to guide us, just the question of uh what's the evidence underpinning guidelines in general, and then uh um a fellow who I warned, I was like, if you want if you publish this paper, you're gonna be a pariah in your field when you go on the interview trail. And uh he said, sure. So I warned him and he became a pariah, but uh there's a long uh Twitter thread instigated by the chair of ATS at the time. That's a fun story, but we can get there after.
SPEAKER_03Yeah, so anyway, so tell me what tell me about the paper, or the paper that you publish uh that you published with the fellow. What you looked at the American Thras Thrash Society guidelines, you looked at the evidence underpinning those guideline recommendations, and you were looking at the strength of the uh evidence that underpinned the recommendation. What did you find?
SPEAKER_00Yeah, so most guidelines to understand the grade, they make a recommendation and they grade the strength of their recommendation, like strong, weak is often, or conditional is often the language you'll see in different guidelines. Um and then the other dimension that a guideline does is what's the evidence, what's the quality of the evidence base? And different different uh guidelines use different uh scoring strategies. And uh You know, I don't probably know better than I do, but the generality was a lot of evidence has a lot of guidelines have limited to no evidence. Their expert opinion, or maybe even worse than expert opinion, is highly confounded observational studies, which I think we do more harm than good. And I say that as an epidemiologist who uses observational research methods. And then the other uh main finding was uh a lot of the strongly recommended. So these are like you should do absolutely for all your patients unless there's exceptions. That's what like strong means. Like you should really, really consider this. Like one in 10 had like good evidence behind those. Some of them are common sense, but a lot of them are, I don't know if we should be doing this for all patients. And those were kind of the the two uh main findings that we published in J M internal medicine, and it led to um you know, at both positive and critical uh reception, I think, when we published.
SPEAKER_03Um, you know, some of the things that you found was that of the 222 recommendations, you know, 63% were based on low quality evidence, and only 8% were based on high quality evidence. Yet, you know, 40% were labeled as strong recommendations, which is what you're you're saying. What was the what was the institutional response uh to that? Um from ATS. Yes.
SPEAKER_00Pretty strong. So my fellow, he presented this at the conference. Not much, you know, it was an abstract. Uh some people came, thought it was interesting. In the process of doing the research, we emailed the chair of ATS. Uh, we shared our study findings, our poster, preliminary findings for input. Crickets, you know, we're not big names in the field. I'm not a pulmonologist. My fellow is a my mentee was a fellow. So, you know, you get this email. Probably got ignored. I don't know the, you know, maybe, maybe they're busy. I I don't know the intention. Published it. And then uh the next morning, I'm on the West Coast time. So my fellow mentee, my mentee, who was a fellow at the time, texted me and he was like, Oh my gosh, um the chair of ATS got on Twitter and started uh tagging people and just went on a tirade, wasn't on Twitter. I don't think he understood how what Twitter was, how it worked, understood that people read tweets, like it wasn't a private conversation, um, and just had this long diatribe of various critiques of how we're wrong. We didn't have a control group because obviously you would read the American Cardiology Association guidelines to help you manage uh uh uh ARDS and and COPD. So uh I don't really understand the rationale there, but basically it was like, this is wrong. How dare you attack us? Uh that was the tone. Other people, I will say, other people in the ATS community were receptive. It was that says it wasn't all uh black, you know, there were there was a mixed reaction, and some people found it really humbling. And um, honestly, this is not new to ATS either. This is true of all guidelines. This is not like an ATS issue. This is true of even cardiology, which has the best evidence base in all of medicine. You know this better than I do, but I think half of their guidelines are based on half of their recommendations are based on low quality evidence. And this is the pinnacle of our of medicine. So this is not an ATS issue, even though it's an ATS study.
SPEAKER_02What so I was gonna make a quip that coming from neurosurgery, I think we have like two randomized controlled trials in our entire specialty. And so we're like the complete opposite. It's all it's all just observational studies and expert opinion on our guidelines. So I I see where you're coming from. So what do you what do you do in the in in the case where you know how how does this affect your clinical decision making when you realize that there's very low evidence for most of the stuff we do?
SPEAKER_00It makes you very humble. One, you have to understand it's not a hard science. There's not truth and untruth. A lot of it is murky. We don't know. And so you have to accept that. A lot of people are not good at accepting certainty. I think the way we train select doctors and the way we teach medicine, the way it's uh incentivized and CME courses are taught, there's just a right thing to do. You have this patient do these five things, here's your algorithm. So one, you have to like free your mind. Like, there's like I use the analysis analogy of the matrix. There's the blue pill and the red pill, and I always mix them up. I think you you want to take the red pill, which is the inconvenient truth, the uncomfortable truth, but it is the truth. And the truth is medicine is hard to practice because one, we don't know a lot of what we're doing is helping people. So that you have to start there. The second thing you have to do is uh you actually should, as a physician, you should aspire to be able to read a paper on your own and make a conclusion. Um, I did that with research training. I don't think everyone can, but at least approach things a little bit more skeptically than just taking whatever the abstract conclusion or now the TikTok video from that eye doctor tells you to do. Like it's not these, like the people making these decisions are not like it, they're not the gospel. There's a lot of opportunity and room for interpretation of what the quality of evidence, how you should apply. And then assuming the evidence is right, then there's a whole separate like how should you apply that evidence to your patient? And that's a whole separate issue. So um the ATS guideline study is just what's the synthesis of a guideline and what's the quality of the evidence for a field. Like you said, for neurosurgery, this is true of every discipline in medicine. This is not a pulmonary issue.
SPEAKER_03So, but how do you respond to the critique of look, you're not a pulmonologist, you're not taking care of these folks? Well, I guess there are two issues. One one is, of course, I guess the there is little argument that the guidelines perhaps there's no reason for them to come out with these strong recommendations. Just say it's class 2B. You know what I'm meaning, say that it's not a class one recommendation that is strongly supported by the evidence, right? But there's this need for a variety of different reasons to communicate that this is what everyone must do. I I don't I guess I understand it, but I also I also don't. I feel like the guidelines over time have seemed to have shifted over time to be less and less more and more certain with less and less supporting evidence. What what do you think? Why do you think that is? I you know that I this comes up, I didn't actually know you had missed this whole ATS brew haha that you had in 2018. But when the lipid guidelines just came out, I the I had a very similar reaction because I was like, what the heck? Where what what has changed that now we're suddenly recommending everyone get an LP little A? Like, why is that? Why are we doing that? Do we have an LPA guided strategy to that shows benefit? Like I don't even I don't know, like why why are you why is that why are this strong why is calcium scoring suddenly giving getting the strong reclinations, right? Because as far as I know, as a cardiologist who's relatively involved in clinical practice and stuff, nothing there's no there hasn't been some earth shattering new data that's come out on that. But yet the guidelines shifted. And so I, you know, I did this little thing on Twitter and uh which showed what the guidelines were supported, what the guide what the prior lipid guidelines are, what the current lipid guidelines are, and how they've changed in terms of you know um what the strength of recommendations are. And clearly we've gotten much more certain with much less evidence. Why why do you think that is, Anil?
SPEAKER_00That's a hard one. I think this comes down to a philosophical divide. Some people think you need to make recommendations. You actually need to make recommendations where we don't have evidence to guide people. But why is that happening now?
SPEAKER_03I mean, it's the same thing. Why now? Like why now versus I mean, it seems like we're getting worse. Why is that? Is that because the academics are not as good?
SPEAKER_00I think uh medicine has gotten more complex. There's more, there's sub sub specialties, there's just more, there's just more we know. And then when you know more, then it opens up other guidelines for topics. And again, you don't have to have a strong evidence base to have a guideline. You can just make your guidelines off observational or expert opinion largely. I don't know. I think it's just we we know more now than we did three decades ago. Um more about diseases, but I don't I honestly don't know. It could just be a shift in medicine.
SPEAKER_03I think it's because there I think there's an increase in capture of key opinion leaders by by industry, honestly. I think I think everything is driven. I think our understanding of changed though? I think so. I think I think it's much worse now than it was before. I think I think there was a time when you had academics that were that had some type of Chinese wall between them and industry. And you know, a lot of guidelines weren't a lot of folks who were on the guideline committees weren't consultants and speakers and and whatnot. That has always existed.
SPEAKER_00Yeah, you can that's an empirical question. You can look at conflicts of interest uh between guideline writers now than 10, 20, 30 years ago.
SPEAKER_03Yeah, so lipid the lipid guidelines, for instance, that just came out, right? One of the things for the lipid guidelines is that they dropped the requirement that you could not have any, you know, interaction with with industry. Because there's been big big thing about, well, you know, just working with industry doesn't mean that I can I can't have an opinion or my opinion doesn't have weight or whatnot. So which is which is a fair point, but but that is something that obviously ha has definitely uh uh changed, I think. So I think it's because over and over again, what what I see uh at least in cardiology is is that you have you have a disease, right? You have you have a therapy that's created for that disease, then suddenly the therapy becomes the reason why we create disease. Like, I mean, aortic synosis has been around for forever, right? We have Taver, and now suddenly we have like 20 different ways you can get a Taver.
SPEAKER_00We have low flow, low gradient, and hammer and nail problem, or the mitus effectal shoe.
SPEAKER_03Yeah, and it's once you got it, you gotta use it. Yep. Decation creep is nuts. Like, you know, the the the the new drug, the new drug for Hokum, right? I mean, how many patients do I have that have symptomatic Hocum, right? And now, yes, it's obviously the case that there's always going to be some level of underdiagnosis when there's nothing to do for a patient, right? But you have this massive incentive now with you know, any patient that any any aging patient that's having shortness of breath that has a sigmoid septum with a with a you know thick interventricular septum and has a gradient, it's like wow, this is this is clearly this this is symptomatic hokum. This is obstructive hyper obstructive hypertrophic cardiomyopathy. They need they need they need this cyber expensive drug. I mean, it's remarkable uh to see. It's always existed, but I think it's just it's just boosted to you know to the to the tenth power at this point. But but so that you know, big agreement on. How do you counter the fact that look, we live in a world, as Anthony is saying, of, you know, where we don't have R C D evidence for everything, but most of the stuff we practice is not R C D based. That doesn't make doesn't make it the case that most of the stuff we do isn't is wrong. There's gonna be some uh but uh so you as a hospitalist compared to a pulmonary critical care person or a pulmonary person who is managing COPD, you know, you're always gonna have this tension between the specialist who's like, well, I've treated a bajillion people like this, and my observations are this. And he has his he or she are gonna have his biases, and you're gonna have your biases, but the issue with you and your biases are technically they're less informed. So, you know, there are competing biases, of course, the subspecialty bias that exists to have whatever therapeutic um optimism, but you always see that the more generalist folks have therapeutic uh therapeutic nihilism bias. So how does one how does one kind of make those make sense of those two things? Like what is objective truth, what's objective?
SPEAKER_00Um I'll push back on some of that as I'm the only generalist here. So um I guess you're a are you a generalist neurosurgeon? Such a thing exists?
SPEAKER_02Technically, I'm a trauma neurosurgeon, so that's about as close as you can get.
SPEAKER_00You know, I think there's credentialism, and there is a lot of gatekeeping. I often think the best guidelines are um, and I still have a problem with the guidelines like the Rex, but the US Preventive Task Force, USPS TF, USPS, PTF, something. I really added it one to you. And I think they're they're when you look at it without skin in the game and you just want, you know, I care about human thriving. Um, I don't care how you thrive. Like I don't care if it's a drug, it's a procedure, it's um, you know, palliative care, whatever it is, that's what I care about. So I don't have a like you said, I don't have a mitus effect. I'm not a I'm not a pulmonologist. To my tools are broader than the field of pulmon pulmonary. So I approach it with a different lens. I have a different clinical expertise than a pulmonologist has. Um, and this gets to indication creep. If you think of all people hospitalized with COPD, hospital medicine physician has a better understanding of that denominator than a pulmonologist. Um, maybe if you have some there are pre-community hospitals, I've seen these where they just literally COPD reflexive consult pulmonary. But you know, if you're doing this somewhat uh sensibly, you're only consulting a pulmonologist when it's a really tough case of COPD in the hospital. That's a huge selection bias of who a subspecialist might see in their clinic and in the hospital. So you're seeing the sickest of the sickest, which don't represent the monolith of the disease of which your recommendations are now applying to. But I see way more general heart heart failure, COPD than any subspecialist will like the representativeness of that kind of patient. And that's if you looked at the proportion of that patient in their in that disease, that's like that's the underneath the water iceberg is what a hospital medicine or a primary care or an emergency department physician, whatever, or a ICU critical care doctor will see. The subspecialist gets that tip of the iceberg. And I think that's what's seen. That's what's visible. And that's what gives them credentialism of like, oh, this is my area because I see this. But you're seeing a selection bias of the most sick, the most complicated. So what you what you do, practice, see, recommend in that population might look very different in a different population. That's one issue. The second issue is as a generalist, you're thinking about the whole patient, how COPD fits with their heart failure, with their cancer, with their frailty. And I can understand that, you know, this medicine, I can read the evidence base and understand this medicine has a time lag to benefit of years. While what you're recommending is true for that disease, it's not true for this patient who has a bad underlying disease and poor prognosis who's not going to live long enough to derive any benefit. Again, so what you have to view it, and I think a good, I've seen good sub-specialists who can do that, who can practice their specialty with the larger, like 30,000 foot view of what's going on with this patient. But I think that's honestly really hard. I've been in really great academic medical centers for most of my career. And uh it's not a trait I see commonly uh where people can step outside of their specialty and say, my specialty is not the most important here. I could recommend this, but I'm not because this patient's other stuff. And you know, it's not black or white, people do this all the time, but I'm often seeing expensive cancer drugs prescribed to people who are ECOG4, they're bedridden, can't get out of bed because they're they're interpreting indication creep. Evidence for people that they see in clinic where the evidence is really strong, but you apply it to a hospitalized population who can't who are so frail and weak, they're not they don't have the same prognosis as the people in the clinic. And that's the distinction. And the other part is I can read the evidence too. I got eyes, I got a brain, and I have a methodology training that most subspecialists don't have. And I I think if you're gonna appraise evidence, epidemiology matters as much or more than the clinical context. You need to have clinical context, but you don't have to do a three-year fellowship and spend your career seeing that disease solely to understand. And there's things I will not appreciate as a generalist, like I'm not gonna understand the prescribing uh uh strategies of like, oh, this side effect is potentially really bad and this is how I manage it. And that's the nuance that I really depend on my subspecialist. And I'm still consulting oftentimes for really complex patients. Like I I often consult um uh folks who have that expertise and uh that context, but it's not a consult for what's the answer. It's a consult to give me an opinion and then we talk and and come to some kind of uh understanding.
SPEAKER_02Aaron Powell That's that's why you never consult neurosurgery, I suppose. Realize that's the same.
SPEAKER_00I consult neurosurgery because I have no idea what to do with uh anything related to neurosurgery. There are fields I have no idea what to do. Neurosurgery, radiation oncology. I'm like, come come tell me what we should do. Optho, I can't even read their note. I can't even read an ophthalmologist's note. I have no idea what they're even saying.
SPEAKER_02We definitely like to keep it that way. I'm glad you're not saying that we're doing too many unnecessary surgeries. This is something that you and I have gone back and forth on a bit is how do you measure the quality of a doctor? Like what how do you how do you suss out those doctors that can do that next level thinking and really take into account um all these patient factors? Is that going to be something that shows up in any sort of quality metric, you know, that's reported to CMS or to your local health system? Or is that just something that, you know, you just have to kind of see it with your own eyes to recognize it?
SPEAKER_00Uh great loaded question with five minutes on the timer. You know, I'll start with where we likely agree. I think the central planning, here's metrics, has probably done more harm than good overall. It's a ton of compliance, it's not measuring the right thing, um, changes our focus from what we should be doing, um, which is actual quality improvement than reporting a lot of these metrics. A lot of the metrics going back to the evidence aren't based on strong evidence. So I think that approach has largely been poorer.
SPEAKER_02You're saying the 30-day readmission rate after COPD is not the most important thing that your patient wants.
SPEAKER_00Nor something we know how to prevent, nor something we know how to identify, nor something that's attributable to the care in the hospital necessarily. I just had the last couple stints of service, I had somebody get readmitted. They were recovering bad uh cold, got better, had COPD exacerbation from their cold, literally got home the next day and got readmitted for respiratory failure. And if you look at any of these metrics, you're like, terrible care, came back in one day. What was Dr. Mockham doing? She got the flu. Like somehow between her recovery from a cold and the one or two days she was home and came back with a whole different viral infection, unrelated to everything that we did in the hospital, and she was getting better.
SPEAKER_02How dare you let her get the flu?
SPEAKER_00How dare I not prevent the flu in the hospital? But so I think central, you know, I think there's examples. I think door to balloon time is a good one. Aspirin for a heart attack used to be a good one, and now the ceiling effects are so pointless that you can't differentiate any hospital apart from any other hospital because it's like 97.8% versus 98.3%.
SPEAKER_02Aaron Powell, I I think the the door to needle time, like a process measure like that, I agree with you. I think that that's that's probably important.
SPEAKER_00Um important because one, it has strong evidence, two, you can measure it. Uh three, you can intervene upon it. That's all related to your systems of care. And I I think very few metrics are like that.
SPEAKER_03Do we do we are we okay allowing patients to decide who is and is not a good doctor?
SPEAKER_00So this is like my thesis is you can't if you're a billionaire and if you guys have any do you have any billionaire listeners on your on your podcast that you're allowed? Yeah, we can only hope. Um, I tell I say this on Twitter. I say this on Twitter and I have a couple billionaire followers who follow me, um, and I hope they read it. But if you have a billion dollars, you can't can't buy good medicine. Like it's not identifiable. There's no market signal. I I think the central planning has done more harm than good, but uh, there's just it's a broken market. I think there's different aspects. So my co-editor for our Substack Ano Times 2 kind of laid it out in a post. What is good doctoring? And she put it in three buckets. I might differ a little bit, but generally there's and this applies more to you, maybe Anisha and definitely you, Anthony, but the technical competence. Can you operate well? I can't tell whether Anthony is a good neurosurgeon. I don't have that capability. I'm not a neurosurgeon. I can't I have no idea if Anthony is a good neurosurgeon on a day. Like I had no clue. Two is relational skills. And this is like bedside rapport. This is where I think patients definitely have the best understanding of who's who's good or not. Does my doctor listen to me? Do they care? Do they ask open-ended questions? Do they follow up with me? I think patients are probably in the best position. As colleagues, like I can't tell whether my colleagues are good. Like, doctors are like free-range chicken. You are your own practice. Maybe it depends on your practice if you're seeing a lot of your own colleagues' patients, but honestly, like I'm not in the room with my colleagues, like saying, I can tell by the care when I inherit their service, but I don't know what they're like at the bedside. Like, I can kind of get a gestalled if someone's terrible or really good when the patient talks about it, but I don't have that expertise. Nurses probably have that that understanding a lot better than uh than doctors do, honestly. What's their relational uh skills and empathy and communication like? Patients have a great understanding. Um, but it's kind of the the tropism. Um I'm blanking, what's that book about uh medicine? Really kind doctor at the bedside who's just really incompetent, technically, knowledge-wise, really bad. And uh I've seen I've seen those doctors. So I think you have the relational skills.
SPEAKER_02The the really, really compassionate surgeon that just can't operate their way out of a paper bag.
SPEAKER_00Yeah. And then the cogniz then I would put the cognitive skills. So do you have a fund of knowledge? Uh which you know I'm not did you memorize a bunch of stuff? I don't care about that, but do you know where to find the knowledge uh that you need to know? Do you have enough framework and foundation where you can build on that? And then do you understand the evidence base? Do you understand how to tailor the evidence base to your patient? And that's a big one. Uh that I think medicine as a field has has missed the boat on. Uh, don't you think that's very I think it's hard to identify that person. Yeah. But that's a really hard job.
SPEAKER_03Don't you think that's a that's a subjective like how who like who like obviously there's a bunch of folks who run clinical trials who I guess we can say have a reasonable grasp in terms of appraisal of evidence. I know I know those those two skills may not necessarily go together, but you know, unrelated. I consider that.
SPEAKER_00Completely unrelated. The people who run the trials are not the same people I trust to to appraise the trials. It's a Venn diagram. There are people who do both.
SPEAKER_02But and Anil, you you did forget the the most important bucket is is does your doctor have a podcast or not?
SPEAKER_00Have a podcast? What credentials?
SPEAKER_03No, but Anil Twitter followers. Running a trial and appraising a trial are two very different things, okay? But running and design, I should say designing a trial. Running and designing a trial just to figure out whether something something works or not. Wouldn't you say that's like an extremely hard skill that extremely few physicians have? Like, you know, like I run around appraising evidence all the time. Um could I design a trial? I've I've never actually designed a trial and and run it. I've never run an actual perspective randomly. No, no, it's not not about taking care of a patient. We're talking about, we're talking about we're talking about how about epistemology. We're talking about whether or not like how does one get to objective truth, right? So, you know, what I'm pointing out is that the folks who and will praise evidence a lot, who you, me, fall in those, in those buckets, some some better than others, what have you. But there is something, uh there's something to be said about folks that that have a clinical question, then go design a prospective randomized control trial to figure out whether it works. And because there's of course no perfect RCT, right? Like we can tear apart any RCT. Um, you know, and so why do we why do we choose not to tear apart some RCTs, but we tear choose to tear apart we choose to uh you know amplify other RCTs? And a lot of that, in my in my view, ends up being what your subjective frame is, right? Like so somebody who is therapeutically a nihilist, a nihilist, sorry, uh, will will I put my name in there. I see what you're doing. Yeah, we'll will have a will have a different approach to appraising evidence than somebody that is a therapeutic optimist. So so some of this in terms of who isn't is not a good doctor, and if it comes down to appraising evidence where and Euclid that that is that is extremely subjective, not at the tales. At the tales, it's pretty obvious that you know Eric Topel is like Jim Kramer, or or that, or that uh you know, people that talk about uh I love that analogy. That you know, that you know, uh Andre Huberman is uh and Peter Atia are just entirely too confident about the effects of you know tanning or what what are they talking about? Cold plunges, hot plunges, yes, cold plunge, right? You know, so clearly that's uh that's what the tales. But there's this big group in the middle. So for instance, uh you know, urologists, right? So if you're a urologist who uh who's who's whose his bias is that he's referred a bunch of a bunch of young males, relatively young males that have metastatic prostate cancer, right? So when he approaches, he's gonna approach PSAs and the the cascade of diagnostic testing that results from that. He's gonna have a very different approach from that because in his mind he has the list of like, you know, seven people, you know, his age in the last six months who, you know, died horribly of metastatic prostate cancer, correct? So my point is I don't know that he's necessarily he or she is necessarily wrong. I think you you both are correct in some ways, but you're making a subjective, a very subjective decision about what is and is not is not good. So can we all kind of like can can we all understand each other? Meaning, like what is the right thing for one individual patient, right? We don't, as an oncologist, right, even if somebody has, even if some somebody walks in and has some metastatic whatever, you never they are specifically trained not to be like, okay, we're gonna treat you as the average person like that. They're like, no, we're we're gonna you have a cancer that's growing, you're here for me to try to kill that cancer, and I'm gonna, I'm gonna do anything everything I can to try to do that. The other, the other perspective they have is of course, you know, and my wife's an oncologist, so I I get this, I get this a bunch, is that she has patients that have been doing really well for years and years and years. They get a pneumonia, they end up in the ICU, they look sick as crap. You know, it's they have metastatic breast cancer and they have this pneumonia and they're sick as crap. And the ICU physicians immediately are like, are like, why is this patient not have like, why is the oncologist not talking to them? Like, why was why did we intubate this person? It's like, okay, you perhaps don't have the perspective of the fact that she's been to see my wife for the last seven years with this metastatic cancer that's been stable because they have these wonderful drugs now. Perhaps you need that perspective. And instead of popping off about how, you know, well, metastatic cancer, and I can't believe that this person is intubated with a with a pneumonia. And of course, from the oncologist's perspective, it's going to be like, well, treat the damn pneumonia and see if you can get her back to where she was two weeks ago.
SPEAKER_00Yeah, I I probably agree with 90% on what you just said. Um and I think where we have to draw the line of what's subjective, what isn't, is um not values, not individual patient values, not what the doctors are seeing in their perspective. First of all, we have to start with, is this study even believable? And there's there is sub there is some subjectivity to that, but it's a lot less subjective than people think if you have those skills. And I debate whether we I debate more about what we should do with studies we believe more than should we believe this study among people who know how to read studies. Like that's like that's where I start. I think a lot of the issue is people don't know how to read studies. So they're bringing these really confounded or horribly done studies and using it as justification to do certain things. And like that, that to me is less subjective. If you take a huge EHR data set, run every single association with a GOP and say it cuts breast cancer by 50%. I think this is the latest uh buzz and on the on the Twitters. I still call it Twitter. That's the latest confounded study, or um that like we shouldn't be even having like our debates are around what is truth as opposed to how you apply truth. Um, so I think there's a ton of what's claimed as subjectivity in this bucket where there is actually less subjectivity than people think there is. Then when you s you have a settled science, there will never be a right answer for a lot of things that we do because it depends on how you look at it. Glass half full, glass half empty, what's your values, what's your perspective, how much shit you want to put up with if I'm allowed to cuss on here, how much uh suffering you're willing to live with for treatment, how much you just want to say, I don't want to see another doctor again. That is there is no answer to that. And that's that's where evidence-based medicine lives. And taking ground truths, as Eric Topel likes to say, but he can't decide what's a ground truth. Uh, actual ground truths and then applying it in practice are two different things. And that to me is like, I'm fine having that conversation and I'm fine tailor. I will treat 50 dia people with diabetes 50 different ways based on their circumstances, what they care, prognosis, side effects, all of that. And that's what we should be doing. But like the example I gave you earlier is not the person that you gave, like was well in the outpatient setting, got a terrible acute illness in the ICU, but before that illness, they were functional status, really good, cognitively intact, talking to you. I'm talking about the person who has been living back and forth between a hospital and a nursing facility, hasn't gone out of bed in a wheelchair. Then you have an oncologist who's reading these trials in ECOG stage one and two, where their prognoses are like six times better. So even if you believe this six-week difference in mortality, when you apply it to a prognosis now of weeks to months, not months to years, the absolute difference is non-existent. And we have that research. I see this regularly in safety net settings. And I imagine this is true in non-safety net settings, is we're applying these amazing new immunotherapy drugs and new targeted therapy drugs to people who should not be getting them because they don't, they're not even close to the trial inclusion population. And to me, that's not subjectivity, that is blatant indication creep and failure to recognize um how prognosis changes uh for different patients. And that's where I differ. Then that's the part that I differ in. And I think we we use values where we're actually the conversation is is the evidence actually applicable in this situation? Cancer screening is a great one where we'll never agree if one in a thousand people saved by breast cancer um with mammography is a good thing or not a good thing. It just depends on how you look at it. And I'm fine with that, and you just have to live with that.
SPEAKER_03That's a good point. Related and the last topic that I I have to get your two cents on is um the uh FDA, how it evaluates data. Do you has your opinion on on that changed given what happens, what's happened with Marty McCary and Vinette Prasad and you know, this string of of uh reje uh this string of different um therapeutics that you know the FDA did not approve, which then resulted in this massive media firestorm. What what's your take on it? Have you been surprised by by all of this?
SPEAKER_00Um surprised. You know, I think where people trying to reform went wrong was in process bureaucracy, um, politicking, stakeholder alliance, um, and a and a large management. If you can't do that, you're not gonna succeed if even if you have good ideas. But um I wasn't the long tail history of the FDA is I don't think they've been very successful at evaluating and appraising evidence. You know, it your bar is gonna be different than what a guideline would recommend. Approving a drug is different than whether you should give a drug. It just means there's a there's an option on the table. Um I think some problems come to like linking FDA approval to mandatory CMS coverage. Um, like because we don't bifurcate or separate those, that that process, the FDA approval has a huge clinical implication. When they approve it, now everyone has to pay for it, which which distorts like insurance and payments, and we pay for it through higher premiums and such. But there's a long history of basic appraisal errors. And you know, I think the Duchenne dystrophy drug was a good one. There's clear, and I work for ICER and the CMO for ICER, who did that report, wrote a scathing like JAMA perspective article at the time. So I'm not this is not hindsight bias of what we know now. This was like at the time of what we knew, improved some random surrogate. We have no idea if it does, none of the clinical outcomes really improved. Huge like red flag of liver toxicity that showed up in small trials. And guess what? That all got worse. So, like, great example of like at the time Peter Marx overruled his own advisory commission, didn't get the uproar that like a Vinay Vinay Prasad gets or someone else gets, but um, you know, Vinay's bombastic, he has strong views and he doesn't follow playbooks. And I thought that was his and Marty's uh downfall as much as the pharmaceutical decisions they made, which didn't win them any allies. You could probably have no allies across the board. They didn't win pharmaceutical allies, they didn't win institutional allies, they didn't win political allies on either. It's really hard when you're trying to actually make major reform to to succeed in that way without having alliances. Trevor Burrus, Jr.
SPEAKER_03Well, it seems like there's this disconnect in the political class and the folks that are on the ground. Meaning you essentially have a uniparty if if you don't have political backing. Um you're talking about staffers in the FDA, or are you talking about like Yeah, no, no, I'm talking about I'm talking about the you know the White House. Um so if the White House is is not going to is not gonna back you, then yeah, then you're you're you're dead, right? And the question is, okay, then why does the White House not back you? Well, if the White House is is is is being directly contacted by whatever interests, um then yeah, then then it's gonna be impossible. Then then basically it's gonna be. Right. Right all right. Well, we've taken up a lot of your time. They uh I know it was fantastic uh uh having you. Um so thanks for all your wisdom. We'll we'll have to have you back. You have a co-host um who uh you know it's hard to coordinate everything, but we'll have to see if she can come back on as well.
SPEAKER_00Your Substack You gotta do this this in uh West Coast hours, Anthony. Exactly.
SPEAKER_02This is West Coast hours. This is uh we're not up as late as well.
SPEAKER_03Not neurosurgery hours.substack.com. Am I getting that right? More or less.
SPEAKER_00I can't even remember.
SPEAKER_03A no times two. Just Google it. Google it. Yeah, Google it. We'll attach it to the show notes. And then uh your Twitter handle is at Anil Macham. So you can follow him there for all his riveting insights as well. So Anil, thanks again. Appreciate you uh coming on. Thanks.
SPEAKER_00Thanks for having me, guys. That was fun. All right, stay on.