Good perspective. If one considers a major center in Calgary, there are likely several physicians doing the procedure who differ in years of experience and volume of procedures. An individual seeking care might easily match with a relatively young physician whose outcome measures are masked by the volume of his/her senior partners. Thus, it is not so simple. Overall, my experience in pediatric lung transplantation bears out the conclusions of your analysis. I worked in two high volume centers and it was evident that our outcomes, often with sicker patients, was superior to the low volume centers, none of which seemed to be upfront with their patients and families about their low volume and outcomes.
Does Procedural Volume Equal Better Care? Skill, Equity, and Outcomes in M-TEER
Dr Mandrola,
The recent analysis by Kumbhani et al. in JAMA Cardiology evaluating procedural volume and outcomes for transcatheter aortic valve replacement (TAVR) and transcatheter edge-to-edge mitral valve repair (M-TEER) provides an important opportunity to revisit your thoughtful reflections in Sensible Medicine [ *], on the tension between experience, convenience, and gradients of technical skill in contemporary cardiovascular care. With respect for your argument, I offer a complementary critical perspective.
Access to equitable, high-quality care for individuals living in rural or socioeconomically vulnerable settings should not be framed as an aspirational goal but as a minimal ethical obligation of modern health systems. Defining quality primarily through procedural expertise risks reinforcing structural inequities if it privileges volume-driven specialized centers over integrative models of care oriented toward sustaining health rather than maximizing procedural throughput.
Kumbhani et al. model cumulative annual procedural volume using multilevel random-effects logistic regression, examining 30-day all-cause mortality, a composite outcome, and in-hospital complications. Yet registry-based observational analyses inevitably rely on coded encounters rather than fully characterized patients and remain vulnerable to residual confounding, incomplete risk adjustment, and unmeasured selection biases that may influence both referral patterns and outcomes.
This limitation is particularly relevant for M-TEER, a costly and technologically complex intervention supported by relatively few randomized controlled trials, often restricted to selected subgroups. In the absence of randomized allocation to low- versus high-volume centers—an ethically and logistically unlikely design—robust adjustment for patient characteristics, institutional context, and operator-related variables becomes essential for causal inference.
Although slightly higher STS-PROM scores were observed among patients treated in lower-volume settings, the magnitude of these differences was modest. Nevertheless, they raise an important question: might structurally vulnerable patients—those with greater frailty, malnutrition, or socioeconomic disadvantage—be disproportionately concentrated in lower-volume centers because of systemic barriers to access? If so, apparent differences in outcomes may reflect inequities in healthcare delivery rather than solely differences in technical expertise.
Methodologically, categorizing procedural volume into tertiles may obscure non-linear exposure–response relationships. In complex biological systems, cumulative exposure often follows inverted-J or threshold patterns; such categorization risks masking early learning curves as well as potential declines associated with excessive procedural intensity. Moreover, annual procedural counts fail to capture institutional longevity and accumulated experiential capital. Performing 50 specialized procedures per year in a recently established center is not equivalent to the same volume within a mature institution shaped by decades of organizational learning.
While specialization may reduce technical errors, it may also narrow holistic clinical judgment regarding patient selection and procedural appropriateness. Observational evidence suggests that during national cardiology meetings—when highly specialized interventional cardiologists are absent—hospital mortality for acute cardiovascular events may paradoxically decrease despite fewer specialized procedures, highlighting the complex and non-linear relationship between procedural expertise and patient outcomes (Jena et al., 2018; Epstein, 2019).
Within the M-TEER cohort, baseline clinical and demographic differences between low- and high-volume strata appear modest and not consistently aligned with a clearly higher lethal risk profile. Interpretation is further constrained by the absence of key covariates, including comprehensive comorbidity indices, functional status, severity and etiology of mitral regurgitation, pulmonary hypertension, atrial fibrillation burden, and transparent clinical criteria guiding procedural selection. Additionally, the lack of adjudicated causes of 30-day mortality limits differentiation between procedural complications, underlying cardiac disease, and noncardiovascular comorbidities.
Under these conditions, attributing poorer outcomes in lower-volume centers solely to reduced operator skill remains clinically speculative. Robust adjustment strategies—such as propensity score modeling for baseline mortality risk—could meaningfully alter effect estimates and potentially attenuate observed associations between procedural volume and outcomes (Rosenbaum & Rubin, 1983; Austin, 2011).
From a salutogenic perspective, the central question is not how many procedures a system performs but how effectively it preserves and strengthens the conditions that sustain human health across the continuum of care. Procedural metrics, although operationally convenient, risk narrowing clinical focus toward technical throughput rather than toward the biological, social, and contextual determinants that shape patient trajectories long before and long after intervention. When volume becomes a proxy for quality without adequate adjustment for structural inequities, it may reflect accumulated access privilege as much as accumulated expertise.
The prevailing emphasis on procedural volume may therefore contribute to a gradual shift from a Health Business—centered on resilience, functional capacity, and well-being—to a Disease Business organized around reimbursable interventions and technological intensity. Such a shift risks redefining success through procedural efficiency rather than through meaningful health outcomes at individual and population levels.
Discussions of experience and outcomes in M-TEER should thus extend beyond operator proficiency to include the systemic incentives that shape patient selection, access to specialized centers, and the distribution of technological resources. Without this broader ethical and structural lens, procedural volume risks becoming less a marker of clinical mastery than an index of how health systems organize inequality. A genuinely salutogenic model of cardiovascular care would prioritize integrative clinical judgment, equitable access, and long-term functional health over sheer procedural intensity—understanding experience not merely as repetition of technical acts but as the cultivation of wisdom in the service of human health rather than the expansion of procedural markets.
Kumbhani DJ, et. Al. ( 2026). Contemporary Operator Procedural Volumes and Outcomes for TAVR and MTEER in the US. JAMA Cardiol. 2026 Jan 8:e255645. doi: 10.1001/jamacardio.2025.5645. Epub ahead of print. PMID: 41505119; PMCID: PMC12784256.
Epstein, D. J. (2019). Range: Why generalists triumph in a specialized world. Riverhead Books.
Jena AB, et al (2018). Acute Myocardial Infarction Mortality During Dates of National Interventional Cardiology Meetings. J Am Heart Assoc. Mar 9;7(6):e008230. doi: 10.1161/JAHA.117.008230. PMID: 29523525; PMCID: PMC5907570.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. https://doi.org/10.1093/biomet/70.1.41
Austin, P. C. (2011). An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behavioral Research, 46(3), 399–424. https://doi.org/10.1080/00273171.2011.568786
My only gripe with calling the Canadian System superior is the slight benefit you see from doing it at a high-experience center is most likely massively outweighed from the harm of less people getting the procedure in Canada because of the 'drive and the wait'.
The comparison group here should be 'someone getting the procedure in a low volume center' and 'someone not getting the procedure at all/12 months later'; and I'm sure the low volume center will win.
Nice post, JM. I'm onboard with driving for more experience. Problem is that the consuming public isn't really aware of the variance in skill found in medicine. My question: what will these organizations do with this information?
The piano in the lobby hits hard. Among amateurs I could be considered on the expert spectrum in piano playing, and yet I have been kicked out of playing Chopin in hospital lobbies. Instead what they prefer is an auto-play mechanical device doing a few chords here and there.
Interesting about Canada. However, how long is the post procedure treatment at the hospital? Do they go back home at or near baseline or improved?
It’s one thing to put them back in their car for a long distance drive fully tuned up and ready to go. It’s quite another to get them out the door as fast as possible, to a rehab facility as fast as possible, and home as fast as possible.
Because after they leave that long distance center of excellence it will be the local community hospital to pick up the pieces. It’s something I’ve dealt with in one of the most resource rich areas of the country.
The entire practice and structure of the medical profession has no mechanism to reward expertise, experience, or seniority. We are all cogs in the wheel - the newest grad gets in line, pumps out widgets alongside the physician with several decades of experience. We all get paid the same; we all work the same long hours. This whole set up shows it has no value for the experience mid- or later-career physicians provide. I recognize that procedures and surgeries bring this to a higher level, but again, it really is the entire structure of the medical profession. There are no mechanisms to reward (or value) the fact that experienced physicians are most assuredly doing things more cost effectively and likely better-outcomes. Where else can you work diligently for 30 years and still make as much as the early career doc and have little, if any, other nods to seniority? The RVU formula we've allowed to dominate and shape the practice of medicine is at the root of it.
When our community medical center began offering heart surgery in 1984 and for the next 30 years, 30-40% of elective cases still opted to go to the big city, 45 minutes away. We had two groups of surgeons, a very young group (2nd or 3rd inning) and a considerably older group (8th inning). After the first 5 or so years, I remember the more experienced group was questioned because of poorer absolute outcomes. However, it turned out that their cases were judged much higher relative surgical risk. By about 2000, the two groups had combined and Administration had put a piano in the general waiting area. By the time I retired in 2022, still 20% of elective cases went to the big city.
It’s an observational study, so of course there are limitations. But as noted, the fact that it was a mandatory registry means there was no issue with selection bias.
I agree the results are congruent with our Bayesian priors. Totally “makes sense” that results are better with more experienced operators. I also liked how it was NOT a dichotomous cutoff btw high vs low volume. It was in fact >2x volume for both procedure types. We are not arguing about the difference btw 15 vs 16 TAVI per year. I do wonder if somehow low volume operators were doing higher risk cases, which might make their worse outcomes more acceptable. I wonder if there were STS scores or some other measure of patient complexity factored into the analysis.
I would also say, even though low volume operators did statistically significantly worse (ie the effect is real), the absolute risk increase is pretty small. So this could be resolved from a patient preferences and values discussion.
I don't know about Calgary or Edmonton, but here, if you go to the university medical center, you may well not have the professor do your procedure. It may be a fellow with the professor looking over his shoulder. And it may be his first procedure.
I would submit answering this question (value/volume) is easy (relatively) but getting into the position to answer it is hard.
I may have faulty recall of the past, but I come from before large groups and networks.
Your family doctor knew who knew what they were doing. He (usually male) knew Fred drank too much to safely do surgery, even if the large hospital in town kept him on the board for his contribution$. Additionally, your family doctor knew Joe, an older surgeon that didn't always wear wing tip shoes, took all the toughest cases, including those with no money, and could pull back old people from the brink of death. He'd send you to Joe.
Arm primary care with outcomes data (because they have to face the patient or family when the outcome is bad). Strip them of CoI (a hard task indeed.) Set them as the primary/only arbiters of referrals. and see how the system works.
The solid organ transplant community in the United States has been tracking such outcome data for decades and there are a number of publications in peer reviewed journals outlining the effects of policy changes, implemented to address the “less than expected“ outcomes in lower volume centers. Here is an AI synopsis of those studies.
Over the past 40 years, higher-volume transplant centers consistently achieve better patient and graft survival than lower-volume centers across kidney, liver, heart, and lung transplantation . This volume–outcome relationship persists after risk adjustment and is most pronounced for complex cases and early post-transplant outcomes . In response, UNOS/OPTN and CMS have implemented public reporting, minimum volume thresholds, and risk-adjusted performance monitoring to incentivize quality improvement and reduce practice variation . Despite these policies, disparities persist, and ongoing efforts focus on refining metrics, expanding access, and addressing social determinants of health .
---
Evidence of volume–outcome relationship
Multiple longitudinal studies and registry analyses demonstrate a robust association between transplant center volume and outcomes:
- Kidney transplantation: High-volume centers have significantly lower observed-to-expected rates of 1-month and 1-year graft loss and patient death compared to low-volume centers, indicating better risk-adjusted outcomes .
- Liver transplantation: Although early studies suggested a volume–outcome relationship, more recent analyses in the MELD era indicate that center volume alone may not significantly predict survival after adjusting for disease severity and other factors . However, low-volume centers still exhibit higher complication rates and resource utilization.
- Heart transplantation: Low-volume centers have significantly higher 1-year mortality and graft failure rates compared to high-volume centers, with the relationship persisting across different risk strata .
- Lung transplantation: Low-volume centers demonstrate higher waitlist mortality and lower post-transplant survival, suggesting that experience and infrastructure significantly impact outcomes .
---
Policy changes and regulatory responses
Recognizing the volume–outcome relationship, several policy changes have been implemented:
- Public reporting: The Scientific Registry of Transplant Recipients (SRTR) began publicly reporting center-specific outcomes in 2001, aiming to improve transparency and incentivize performance improvement .
- Minimum volume requirements: The Centers for Medicare & Medicaid Services (CMS) established minimum volume thresholds for transplant programs, requiring centers to perform a minimum number of transplants annually to maintain Medicare approval.
- Risk-adjusted performance monitoring: UNOS/OPTN and CMS adopted risk-adjusted models to evaluate center performance, accounting for differences in patient and donor characteristics .
- Textbook outcome metrics: Recent initiatives have introduced composite quality metrics, such as "textbook outcome", which incorporate multiple post-transplant endpoints to provide a more comprehensive assessment of center performance .
---
Impact of policy changes on transplant outcomes
Policy changes have had mixed effects on transplant outcomes:
- Improved transparency: Public reporting has increased awareness of center performance variations and encouraged centers to address deficiencies.
- Reduced practice variation: Risk-adjusted monitoring has helped identify underperforming centers and promote standardization of care .
- Unintended consequences: Some centers have adopted risk-averse behaviors, avoiding high-risk patients or marginal organs to maintain favorable reported outcomes, potentially limiting access to transplantation .
- Persistent disparities: Despite policy interventions, significant geographic and center-level disparities in outcomes persist, indicating the need for ongoing evaluation and refinement of policies .
---
Current challenges and future directions
Several challenges remain in addressing the volume–outcome relationship and improving transplant outcomes:
- Balancing access and quality: Policies must balance the need to maintain high-quality outcomes while ensuring equitable access to transplantation, particularly for underserved populations .
- Refining quality metrics: Current metrics, such as 1-year survival, may not fully capture center performance. Longer-term outcomes and patient-reported quality of life should be incorporated into quality assessments .
- Addressing social determinants: Socioeconomic factors, insurance status, and geographic location significantly impact transplant access and outcomes, necessitating targeted interventions .
- Promoting innovation: Regulatory frameworks should support innovation and the adoption of new technologies and practices that improve outcomes without penalizing centers for taking on complex cases .
---
Over the past 40 years, higher-volume transplant centers consistently achieve better patient and graft survival than lower-volume centers. Policy changes, including public reporting, minimum volume requirements, and risk-adjusted performance monitoring, have been implemented to address these disparities. While these policies have improved transparency and reduced practice variation, persistent disparities underscore the need for continued evaluation and refinement of quality metrics and access policies.
1. 2.4% vs 2% may be statistically significant, but real world, it’s not. Do ANYthing a million times, and you can glean out mathematical differences. I don’t disagree that experience has clear value; this study however is weak sauce in its defense.
2. Oddly enough, very large university/referral centers always seem to favor high volumes at the expense of local access. Imagine that. Our community of about 60,000 has a busy TAVR program (I’m not a cardiologist); the ‘big city’ is 200 miles away. Some things need to go downtown. TAVR should not: the very point of the procedure is to provide broader access to those who are perhaps imperfect surgical candidates.
To burden them with long drives and long stays ‘downtown’, simply to pump up the tertiary facility’s volume and keep their parking lots filled? Thanks; no. Same argument they’ve made for 30 years for multiple types of cancer care. They were wrong then, and now.
The Canadian example is a poor one. Count the number of Americans flocking north for care. Now, do the reverse. If numbers are important, then you need to include those as well.
Few will argue against the value of ongoing experience. Let it not however become a stalking horse for the relative handful of massive centers whose organizational goal is to bury their heads in the feeding trough, to the exclusion of ‘mere’ community care.
In cardiology in 2025, for example, less than 1% of Canadians went out of country for procedures. And some of those will have gone somewhere other than the United States.
I wonder how many Americans went out of country for healthcare for better or more affordable access last year. The 1% number for Canadian medical tourism probably isn't unusual.
It's obvious why Americans don't "flock" North. Even if you could find some way to get on the list, paying full fare in Canada isn't any sort of advantage.
All good points. I thank you especially for #1. I tend to disregard conclusions based on the tiny differences in low incidence outcomes. Always look for the raw numbers and compute the percentages for each outcome. So many studies today try to obscure the common sense facts with misleading statistical methods.
Hit the send button too soon. :-). I meant to say the Feds transfer a certain amount of money each year for health to the provinces based on population.
Very interesting John, but not unexpected as you note. Just one point, Canadian health care delivery is a provincial jurisdiction, each transferring a certain amount of money each year to the provinces based on population, so what happens in Alberta may not be the case in Ontario, or BC, or Quebec, etc.
Good perspective. If one considers a major center in Calgary, there are likely several physicians doing the procedure who differ in years of experience and volume of procedures. An individual seeking care might easily match with a relatively young physician whose outcome measures are masked by the volume of his/her senior partners. Thus, it is not so simple. Overall, my experience in pediatric lung transplantation bears out the conclusions of your analysis. I worked in two high volume centers and it was evident that our outcomes, often with sicker patients, was superior to the low volume centers, none of which seemed to be upfront with their patients and families about their low volume and outcomes.
Does Procedural Volume Equal Better Care? Skill, Equity, and Outcomes in M-TEER
Dr Mandrola,
The recent analysis by Kumbhani et al. in JAMA Cardiology evaluating procedural volume and outcomes for transcatheter aortic valve replacement (TAVR) and transcatheter edge-to-edge mitral valve repair (M-TEER) provides an important opportunity to revisit your thoughtful reflections in Sensible Medicine [ *], on the tension between experience, convenience, and gradients of technical skill in contemporary cardiovascular care. With respect for your argument, I offer a complementary critical perspective.
Access to equitable, high-quality care for individuals living in rural or socioeconomically vulnerable settings should not be framed as an aspirational goal but as a minimal ethical obligation of modern health systems. Defining quality primarily through procedural expertise risks reinforcing structural inequities if it privileges volume-driven specialized centers over integrative models of care oriented toward sustaining health rather than maximizing procedural throughput.
Kumbhani et al. model cumulative annual procedural volume using multilevel random-effects logistic regression, examining 30-day all-cause mortality, a composite outcome, and in-hospital complications. Yet registry-based observational analyses inevitably rely on coded encounters rather than fully characterized patients and remain vulnerable to residual confounding, incomplete risk adjustment, and unmeasured selection biases that may influence both referral patterns and outcomes.
This limitation is particularly relevant for M-TEER, a costly and technologically complex intervention supported by relatively few randomized controlled trials, often restricted to selected subgroups. In the absence of randomized allocation to low- versus high-volume centers—an ethically and logistically unlikely design—robust adjustment for patient characteristics, institutional context, and operator-related variables becomes essential for causal inference.
Although slightly higher STS-PROM scores were observed among patients treated in lower-volume settings, the magnitude of these differences was modest. Nevertheless, they raise an important question: might structurally vulnerable patients—those with greater frailty, malnutrition, or socioeconomic disadvantage—be disproportionately concentrated in lower-volume centers because of systemic barriers to access? If so, apparent differences in outcomes may reflect inequities in healthcare delivery rather than solely differences in technical expertise.
Methodologically, categorizing procedural volume into tertiles may obscure non-linear exposure–response relationships. In complex biological systems, cumulative exposure often follows inverted-J or threshold patterns; such categorization risks masking early learning curves as well as potential declines associated with excessive procedural intensity. Moreover, annual procedural counts fail to capture institutional longevity and accumulated experiential capital. Performing 50 specialized procedures per year in a recently established center is not equivalent to the same volume within a mature institution shaped by decades of organizational learning.
While specialization may reduce technical errors, it may also narrow holistic clinical judgment regarding patient selection and procedural appropriateness. Observational evidence suggests that during national cardiology meetings—when highly specialized interventional cardiologists are absent—hospital mortality for acute cardiovascular events may paradoxically decrease despite fewer specialized procedures, highlighting the complex and non-linear relationship between procedural expertise and patient outcomes (Jena et al., 2018; Epstein, 2019).
Within the M-TEER cohort, baseline clinical and demographic differences between low- and high-volume strata appear modest and not consistently aligned with a clearly higher lethal risk profile. Interpretation is further constrained by the absence of key covariates, including comprehensive comorbidity indices, functional status, severity and etiology of mitral regurgitation, pulmonary hypertension, atrial fibrillation burden, and transparent clinical criteria guiding procedural selection. Additionally, the lack of adjudicated causes of 30-day mortality limits differentiation between procedural complications, underlying cardiac disease, and noncardiovascular comorbidities.
Under these conditions, attributing poorer outcomes in lower-volume centers solely to reduced operator skill remains clinically speculative. Robust adjustment strategies—such as propensity score modeling for baseline mortality risk—could meaningfully alter effect estimates and potentially attenuate observed associations between procedural volume and outcomes (Rosenbaum & Rubin, 1983; Austin, 2011).
From a salutogenic perspective, the central question is not how many procedures a system performs but how effectively it preserves and strengthens the conditions that sustain human health across the continuum of care. Procedural metrics, although operationally convenient, risk narrowing clinical focus toward technical throughput rather than toward the biological, social, and contextual determinants that shape patient trajectories long before and long after intervention. When volume becomes a proxy for quality without adequate adjustment for structural inequities, it may reflect accumulated access privilege as much as accumulated expertise.
The prevailing emphasis on procedural volume may therefore contribute to a gradual shift from a Health Business—centered on resilience, functional capacity, and well-being—to a Disease Business organized around reimbursable interventions and technological intensity. Such a shift risks redefining success through procedural efficiency rather than through meaningful health outcomes at individual and population levels.
Discussions of experience and outcomes in M-TEER should thus extend beyond operator proficiency to include the systemic incentives that shape patient selection, access to specialized centers, and the distribution of technological resources. Without this broader ethical and structural lens, procedural volume risks becoming less a marker of clinical mastery than an index of how health systems organize inequality. A genuinely salutogenic model of cardiovascular care would prioritize integrative clinical judgment, equitable access, and long-term functional health over sheer procedural intensity—understanding experience not merely as repetition of technical acts but as the cultivation of wisdom in the service of human health rather than the expansion of procedural markets.
Respectfully,
References
*[ https://substack.com/app-link/post?publication_id=1000397&post_id=186602716&utm_source=post-email-title&utm_campaign=email-post-title&isFreemail=true&r=1uh5ig&token=eyJ1c2VyX2lkIjoxMTE2NTQ5NTIsInBvc3RfaWQiOjE4NjYwMjcxNiwiaWF0IjoxNzcwMDM3ODY1LCJleHAiOjE3NzI2Mjk4NjUsImlzcyI6InB1Yi0xMDAwMzk3Iiwic3ViIjoicG9zdC1yZWFjdGlvbiJ9.PY1M7EyrgFanrEUhGCXO3azjKfNTzg7Bp7EyXsoVOAo ]
Kumbhani DJ, et. Al. ( 2026). Contemporary Operator Procedural Volumes and Outcomes for TAVR and MTEER in the US. JAMA Cardiol. 2026 Jan 8:e255645. doi: 10.1001/jamacardio.2025.5645. Epub ahead of print. PMID: 41505119; PMCID: PMC12784256.
Epstein, D. J. (2019). Range: Why generalists triumph in a specialized world. Riverhead Books.
Jena AB, et al (2018). Acute Myocardial Infarction Mortality During Dates of National Interventional Cardiology Meetings. J Am Heart Assoc. Mar 9;7(6):e008230. doi: 10.1161/JAHA.117.008230. PMID: 29523525; PMCID: PMC5907570.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. https://doi.org/10.1093/biomet/70.1.41
Austin, P. C. (2011). An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behavioral Research, 46(3), 399–424. https://doi.org/10.1080/00273171.2011.568786
My only gripe with calling the Canadian System superior is the slight benefit you see from doing it at a high-experience center is most likely massively outweighed from the harm of less people getting the procedure in Canada because of the 'drive and the wait'.
The comparison group here should be 'someone getting the procedure in a low volume center' and 'someone not getting the procedure at all/12 months later'; and I'm sure the low volume center will win.
Nice post, JM. I'm onboard with driving for more experience. Problem is that the consuming public isn't really aware of the variance in skill found in medicine. My question: what will these organizations do with this information?
The piano in the lobby hits hard. Among amateurs I could be considered on the expert spectrum in piano playing, and yet I have been kicked out of playing Chopin in hospital lobbies. Instead what they prefer is an auto-play mechanical device doing a few chords here and there.
Interesting about Canada. However, how long is the post procedure treatment at the hospital? Do they go back home at or near baseline or improved?
It’s one thing to put them back in their car for a long distance drive fully tuned up and ready to go. It’s quite another to get them out the door as fast as possible, to a rehab facility as fast as possible, and home as fast as possible.
Because after they leave that long distance center of excellence it will be the local community hospital to pick up the pieces. It’s something I’ve dealt with in one of the most resource rich areas of the country.
The entire practice and structure of the medical profession has no mechanism to reward expertise, experience, or seniority. We are all cogs in the wheel - the newest grad gets in line, pumps out widgets alongside the physician with several decades of experience. We all get paid the same; we all work the same long hours. This whole set up shows it has no value for the experience mid- or later-career physicians provide. I recognize that procedures and surgeries bring this to a higher level, but again, it really is the entire structure of the medical profession. There are no mechanisms to reward (or value) the fact that experienced physicians are most assuredly doing things more cost effectively and likely better-outcomes. Where else can you work diligently for 30 years and still make as much as the early career doc and have little, if any, other nods to seniority? The RVU formula we've allowed to dominate and shape the practice of medicine is at the root of it.
When our community medical center began offering heart surgery in 1984 and for the next 30 years, 30-40% of elective cases still opted to go to the big city, 45 minutes away. We had two groups of surgeons, a very young group (2nd or 3rd inning) and a considerably older group (8th inning). After the first 5 or so years, I remember the more experienced group was questioned because of poorer absolute outcomes. However, it turned out that their cases were judged much higher relative surgical risk. By about 2000, the two groups had combined and Administration had put a piano in the general waiting area. By the time I retired in 2022, still 20% of elective cases went to the big city.
It’s an observational study, so of course there are limitations. But as noted, the fact that it was a mandatory registry means there was no issue with selection bias.
I agree the results are congruent with our Bayesian priors. Totally “makes sense” that results are better with more experienced operators. I also liked how it was NOT a dichotomous cutoff btw high vs low volume. It was in fact >2x volume for both procedure types. We are not arguing about the difference btw 15 vs 16 TAVI per year. I do wonder if somehow low volume operators were doing higher risk cases, which might make their worse outcomes more acceptable. I wonder if there were STS scores or some other measure of patient complexity factored into the analysis.
I would also say, even though low volume operators did statistically significantly worse (ie the effect is real), the absolute risk increase is pretty small. So this could be resolved from a patient preferences and values discussion.
My point exactly. The number going north is approximately…zero.
I don't know about Calgary or Edmonton, but here, if you go to the university medical center, you may well not have the professor do your procedure. It may be a fellow with the professor looking over his shoulder. And it may be his first procedure.
I would submit answering this question (value/volume) is easy (relatively) but getting into the position to answer it is hard.
I may have faulty recall of the past, but I come from before large groups and networks.
Your family doctor knew who knew what they were doing. He (usually male) knew Fred drank too much to safely do surgery, even if the large hospital in town kept him on the board for his contribution$. Additionally, your family doctor knew Joe, an older surgeon that didn't always wear wing tip shoes, took all the toughest cases, including those with no money, and could pull back old people from the brink of death. He'd send you to Joe.
Arm primary care with outcomes data (because they have to face the patient or family when the outcome is bad). Strip them of CoI (a hard task indeed.) Set them as the primary/only arbiters of referrals. and see how the system works.
Is there a natural experiment out there?
The solid organ transplant community in the United States has been tracking such outcome data for decades and there are a number of publications in peer reviewed journals outlining the effects of policy changes, implemented to address the “less than expected“ outcomes in lower volume centers. Here is an AI synopsis of those studies.
Over the past 40 years, higher-volume transplant centers consistently achieve better patient and graft survival than lower-volume centers across kidney, liver, heart, and lung transplantation . This volume–outcome relationship persists after risk adjustment and is most pronounced for complex cases and early post-transplant outcomes . In response, UNOS/OPTN and CMS have implemented public reporting, minimum volume thresholds, and risk-adjusted performance monitoring to incentivize quality improvement and reduce practice variation . Despite these policies, disparities persist, and ongoing efforts focus on refining metrics, expanding access, and addressing social determinants of health .
---
Evidence of volume–outcome relationship
Multiple longitudinal studies and registry analyses demonstrate a robust association between transplant center volume and outcomes:
- Kidney transplantation: High-volume centers have significantly lower observed-to-expected rates of 1-month and 1-year graft loss and patient death compared to low-volume centers, indicating better risk-adjusted outcomes .
- Liver transplantation: Although early studies suggested a volume–outcome relationship, more recent analyses in the MELD era indicate that center volume alone may not significantly predict survival after adjusting for disease severity and other factors . However, low-volume centers still exhibit higher complication rates and resource utilization.
- Heart transplantation: Low-volume centers have significantly higher 1-year mortality and graft failure rates compared to high-volume centers, with the relationship persisting across different risk strata .
- Lung transplantation: Low-volume centers demonstrate higher waitlist mortality and lower post-transplant survival, suggesting that experience and infrastructure significantly impact outcomes .
---
Policy changes and regulatory responses
Recognizing the volume–outcome relationship, several policy changes have been implemented:
- Public reporting: The Scientific Registry of Transplant Recipients (SRTR) began publicly reporting center-specific outcomes in 2001, aiming to improve transparency and incentivize performance improvement .
- Minimum volume requirements: The Centers for Medicare & Medicaid Services (CMS) established minimum volume thresholds for transplant programs, requiring centers to perform a minimum number of transplants annually to maintain Medicare approval.
- Risk-adjusted performance monitoring: UNOS/OPTN and CMS adopted risk-adjusted models to evaluate center performance, accounting for differences in patient and donor characteristics .
- Textbook outcome metrics: Recent initiatives have introduced composite quality metrics, such as "textbook outcome", which incorporate multiple post-transplant endpoints to provide a more comprehensive assessment of center performance .
---
Impact of policy changes on transplant outcomes
Policy changes have had mixed effects on transplant outcomes:
- Improved transparency: Public reporting has increased awareness of center performance variations and encouraged centers to address deficiencies.
- Reduced practice variation: Risk-adjusted monitoring has helped identify underperforming centers and promote standardization of care .
- Unintended consequences: Some centers have adopted risk-averse behaviors, avoiding high-risk patients or marginal organs to maintain favorable reported outcomes, potentially limiting access to transplantation .
- Persistent disparities: Despite policy interventions, significant geographic and center-level disparities in outcomes persist, indicating the need for ongoing evaluation and refinement of policies .
---
Current challenges and future directions
Several challenges remain in addressing the volume–outcome relationship and improving transplant outcomes:
- Balancing access and quality: Policies must balance the need to maintain high-quality outcomes while ensuring equitable access to transplantation, particularly for underserved populations .
- Refining quality metrics: Current metrics, such as 1-year survival, may not fully capture center performance. Longer-term outcomes and patient-reported quality of life should be incorporated into quality assessments .
- Addressing social determinants: Socioeconomic factors, insurance status, and geographic location significantly impact transplant access and outcomes, necessitating targeted interventions .
- Promoting innovation: Regulatory frameworks should support innovation and the adoption of new technologies and practices that improve outcomes without penalizing centers for taking on complex cases .
---
Over the past 40 years, higher-volume transplant centers consistently achieve better patient and graft survival than lower-volume centers. Policy changes, including public reporting, minimum volume requirements, and risk-adjusted performance monitoring, have been implemented to address these disparities. While these policies have improved transparency and reduced practice variation, persistent disparities underscore the need for continued evaluation and refinement of quality metrics and access policies.
Sources
1. https://pubmed.ncbi.nlm.nih.gov/28650358/#111aeuUW. Annals of Surgery, 2018
2. https://pubmed.ncbi.nlm.nih.gov/16925567/#114geEcF. American Journal of Transplantation, 2006
3. https://pubmed.ncbi.nlm.nih.gov/20837931/#113rzKmj. Circulation, 2010
4. https://pubmed.ncbi.nlm.nih.gov/28862792/#116PTUqa. American Journal of Transplantation, 2018
5. https://pubmed.ncbi.nlm.nih.gov/22682114/#111y5b4p. American Journal of Transplantation, 2012
6. https://pubmed.ncbi.nlm.nih.gov/24786673/#113UciBY. American Journal of Transplantation, 2014
7. https://pubmed.ncbi.nlm.nih.gov/38809546/#115A3Pdc. JAMA Surgery, 2024
8. https://pubmed.ncbi.nlm.nih.gov/23132331/#116HJ8wV. Circulation: Cardiovascular Quality and Outcomes, 2012
9. https://pubmed.ncbi.nlm.nih.gov/21980182/#117W5X7y. Clinical Journal of the American Society of Nephrology, 2011
10. https://pubmed.ncbi.nlm.nih.gov/36102576/#1145FqJM. Journal of the American College of Surgeons, 2022
11. https://pubmed.ncbi.nlm.nih.gov/37120382/#114WLHTU. Surgery, 2023
12. https://pubmed.ncbi.nlm.nih.gov/33295970/#11658a5N. JAMA Network Open, 2020
13. https://pubmed.ncbi.nlm.nih.gov/32282985/#116R5HQk. American Journal of Transplantation, 2020
Two points:
1. 2.4% vs 2% may be statistically significant, but real world, it’s not. Do ANYthing a million times, and you can glean out mathematical differences. I don’t disagree that experience has clear value; this study however is weak sauce in its defense.
2. Oddly enough, very large university/referral centers always seem to favor high volumes at the expense of local access. Imagine that. Our community of about 60,000 has a busy TAVR program (I’m not a cardiologist); the ‘big city’ is 200 miles away. Some things need to go downtown. TAVR should not: the very point of the procedure is to provide broader access to those who are perhaps imperfect surgical candidates.
To burden them with long drives and long stays ‘downtown’, simply to pump up the tertiary facility’s volume and keep their parking lots filled? Thanks; no. Same argument they’ve made for 30 years for multiple types of cancer care. They were wrong then, and now.
The Canadian example is a poor one. Count the number of Americans flocking north for care. Now, do the reverse. If numbers are important, then you need to include those as well.
Few will argue against the value of ongoing experience. Let it not however become a stalking horse for the relative handful of massive centers whose organizational goal is to bury their heads in the feeding trough, to the exclusion of ‘mere’ community care.
Flocking is overdoing it.
In cardiology in 2025, for example, less than 1% of Canadians went out of country for procedures. And some of those will have gone somewhere other than the United States.
I wonder how many Americans went out of country for healthcare for better or more affordable access last year. The 1% number for Canadian medical tourism probably isn't unusual.
It's obvious why Americans don't "flock" North. Even if you could find some way to get on the list, paying full fare in Canada isn't any sort of advantage.
All good points. I thank you especially for #1. I tend to disregard conclusions based on the tiny differences in low incidence outcomes. Always look for the raw numbers and compute the percentages for each outcome. So many studies today try to obscure the common sense facts with misleading statistical methods.
Hit the send button too soon. :-). I meant to say the Feds transfer a certain amount of money each year for health to the provinces based on population.
Very interesting John, but not unexpected as you note. Just one point, Canadian health care delivery is a provincial jurisdiction, each transferring a certain amount of money each year to the provinces based on population, so what happens in Alberta may not be the case in Ontario, or BC, or Quebec, etc.