I share your concern about why more physicians do not follow the constrained vision. However, I have read much of Thomas Sowell in the past but never Conflict of visions. But the concept of the unconstrained versus constrained vision seems to be in the pattern of a several 1000-year old conflict between Plato and Aristotle. The idea of the deductive versus inductive reasoning. I have just finished reading Arthur Herman's The Cave and the Light and I am going to read it again. It covers 2500 years of the philosophy of the conflict of this vision and its impact on society, politics as well as science. I took this moment to upgrade my subscription to paid. After 30 years in medicine I am starting to understand what I see things so differently than many other physicians.
"Prevention is best left to person's smart choices and luck." I could not disagree more. Most people do not have the knowledge or motivation to make smart choices. Fortunately, I had a doctor 50 years ago who gave me both -- a plant based diet, daily walking and lifting things, and goodbye to bad habits like drinking red wine for your heart. My 5 older brothers, all meat and dairy consumers, died between 50 and 80 after years of bad health. My wife and I are 90, physically and mentally active, and in good health. We all will die. We are not trying to prevent death. We are eating to save the planet and prevent cruelty to animals, and avoid what is preventable. Not chasing health. And we have a doctor who encourages her patients to adopt a plant-based diet as the evidence for its benefits piles up. Come on Doc, don't be a pessimist.
1. I sent my cousin who had aortic aneurysm for carotid screening. 50% carotid, but was found to have thyroid cancer -cured
2. A man owned 11 McDonald’s work like a dog to get where he was. Scanned his aorta found a moderate size malignant renal cell cancer - 10 years later no more disease
3. Scanned a 64-year-old man for his aorta -very obese, had a stent. Ex smoker, Medicare age next year insurance wouldn’t pay for the aortic scan. 7 cm AAA. Sent me a photo postop scuba diving with a sign that said “go blue” . He was OSU alum
I have many more of these, of course, and they fall into better to “be lucky than smart” category
These were not random people screened. They had significant risk of atherosclerosis and the screening was lucky to find an unrelated treatable problem
The world’s about to change with PCCT. Treating people with statins while their underlying disease progresses over five years and they have negative treadmills till they don’t is not Ideal.
There’s nothing wrong with being lucky and finding unrelated malignancy if the testing was reasonable in the first place.
I think a total body scan in the absence of anything to give you a direction has an extraordinary low predictive value, and as you point out bad things can happen chasing incidental Omas-
A 40 year-old woman had ST changes getting anesthesia for rhinoplasty . A colleague did an “ergonovine” cath.
The images of her coronary tree being peeled away with each heartbeat is still burned into my consciousness.
People forget that we play in the street . They buy the lotto ticket and actually anxiously check their results. The risk of medical harm is extraordinarily higher and they don’t consider that possible.
Just a minor correction. Likelihood ratios are mutlipliers of odds rather than probabilities. The probability 0.1 has to be converted to odds 9:1 against (0.111) to be multiplied by the LR and the resultant post test odds have to then be converted back to probability. 0.111 x 9 in this case of sensitivity and specificity equal to 0.9. This post test odds then has to be converted back to probability. In this case 0.111 x 9 =0.999 which is post test odds of 0.4997 rather than 50:50. Pedantic I know.
It also is important for example with d-dimer testing as referenced to John’s latest medscape review of the NEJM ILR study to know what the acceptable miss rate is not just what tests do to pre test probabilities.
A patient with a low pretest probability of PE of 10% (about right for -ve Wells score) with a d-dimer of <500 with -LR of 0.05 would have post test probability of 0.5525% - easily good enough to exclude PE. It is estimated that a d-dimer of 1000-1500 has a LR of 1.0. A d-dimer of 800 has a -LR of about 0.5 and so even if is often called “positive” it actually reduces the post test probability. But the crucial question is does it reduce it enough? And what is the acceptable false -ve rate? The post test probability for 0.1 prevalence and LR of 0.5 is 5.3%. And I think most would not accept that threshold. Most I think accept 1%. LR0.05 gets you there easily for low probability patients so we have just learnt that heuristic and attributed a false dichotomised label to the continuous test.
Great article! I am a Christian so I do believe that everyone already has an appointment with death that has been set and pre-determined.
It doesn’t mean we don’t do the things in this world to improve your health. But no human endeavor can change the fact that your appointment with death is set.
Honestly, knowing that fact, helps me to work tirelessly for my patients since we as doctors have that calling but to also not be utterly crushed by one of my patients dying.
Total agreement on the arithmetic: at 1% prevalence a good test still drowns you in false positives, and the near-fatal biopsy of a benign nodule is the harm nobody counts.
Where I want to push, as respectfully as I can, is the slide from "whole-body scanning is folly" to "we die when it is our time, prevention is smart choices and good luck."
Your own example undercuts the fatalism: the cyclist who called his breathlessness asthma had a fixable LAD lesion. That is not luck, that is a missed modifiable cause.
I was that patient, a reassuring workup over an 80% right coronary blockage .
The honest line is not "more is better" versus "nothing is modifiable." It is that a few high-yield, cheap measures with real outcome data (blood pressure, ApoB, Lp(a) once) sit in a different Bayesian universe than a total-body MRI hunting 0.1% disease.
Would you accept that the constrained-vision case argues for a short list of good tests, not for testing nihilism?
Very well put John and I appreciate your passion. I don't see anything in this essay to disagree with. But I also know that you and I and anyone else who thinks this way lost this battle a long time ago. Trying to explain these concepts to my very smart but non-medical family members, for example, has never worked. They seem to understand the larger concept but can't apply it in the real world. My sister frequently says (in regards to screenings generally), "isn't more information just always good?" and I ask her "if that's true, then why not get a screening colonoscopy in your 5-year-old son?" She knows it's a ridiculous question but can't see how the same principle might apply to any medical test in any person: that there are tests that are likely to provide meaningful and actionable results and many more tests that are unhelpful, wasteful, and/or harmful. Knowing the difference or trying to figure it out is so much of what we do in medicine! But so many people really think medicine is just doing tests, the more the better.
I share your perspective. Do the right things for your health (exercise, sleep, avoid harmful habits, etc) and don’t ignore warning signals. I joke that when I’m very old and ready for what’s next, I’ll ware a “Do not resuscitate” tennis shirt every time I play.
Listened to the discussion and appreciated it. I also have read/read Thomas Sowell. Love his take on human behavior. I am not willing to go down the scan/genomic sequencing rabbit hole. For those who do "More power to them" None of us gets out of here alive, do the best you can in the present moment and try to remain healthy
John, thank you so much for this super post. Coincidentally I just covered the same topic with the same math, but from the standpoint of blood tests for early cancer detection, the so-called “liquid biopsies.“. You added the philosophy and I really appreciated it. https://jamesstein18.substack.com/p/multi-cancer-early-detection-tests?r=p5sxn&utm_medium=ios
A great post. And thanks for sharing the link. To any readers here, Dr Stein has a tremendous Substack!
Thanks, John!!
I share your concern about why more physicians do not follow the constrained vision. However, I have read much of Thomas Sowell in the past but never Conflict of visions. But the concept of the unconstrained versus constrained vision seems to be in the pattern of a several 1000-year old conflict between Plato and Aristotle. The idea of the deductive versus inductive reasoning. I have just finished reading Arthur Herman's The Cave and the Light and I am going to read it again. It covers 2500 years of the philosophy of the conflict of this vision and its impact on society, politics as well as science. I took this moment to upgrade my subscription to paid. After 30 years in medicine I am starting to understand what I see things so differently than many other physicians.
"Prevention is best left to person's smart choices and luck." I could not disagree more. Most people do not have the knowledge or motivation to make smart choices. Fortunately, I had a doctor 50 years ago who gave me both -- a plant based diet, daily walking and lifting things, and goodbye to bad habits like drinking red wine for your heart. My 5 older brothers, all meat and dairy consumers, died between 50 and 80 after years of bad health. My wife and I are 90, physically and mentally active, and in good health. We all will die. We are not trying to prevent death. We are eating to save the planet and prevent cruelty to animals, and avoid what is preventable. Not chasing health. And we have a doctor who encourages her patients to adopt a plant-based diet as the evidence for its benefits piles up. Come on Doc, don't be a pessimist.
There is a flipside
1. I sent my cousin who had aortic aneurysm for carotid screening. 50% carotid, but was found to have thyroid cancer -cured
2. A man owned 11 McDonald’s work like a dog to get where he was. Scanned his aorta found a moderate size malignant renal cell cancer - 10 years later no more disease
3. Scanned a 64-year-old man for his aorta -very obese, had a stent. Ex smoker, Medicare age next year insurance wouldn’t pay for the aortic scan. 7 cm AAA. Sent me a photo postop scuba diving with a sign that said “go blue” . He was OSU alum
I have many more of these, of course, and they fall into better to “be lucky than smart” category
These were not random people screened. They had significant risk of atherosclerosis and the screening was lucky to find an unrelated treatable problem
The world’s about to change with PCCT. Treating people with statins while their underlying disease progresses over five years and they have negative treadmills till they don’t is not Ideal.
There’s nothing wrong with being lucky and finding unrelated malignancy if the testing was reasonable in the first place.
I think a total body scan in the absence of anything to give you a direction has an extraordinary low predictive value, and as you point out bad things can happen chasing incidental Omas-
A 40 year-old woman had ST changes getting anesthesia for rhinoplasty . A colleague did an “ergonovine” cath.
The images of her coronary tree being peeled away with each heartbeat is still burned into my consciousness.
People forget that we play in the street . They buy the lotto ticket and actually anxiously check their results. The risk of medical harm is extraordinarily higher and they don’t consider that possible.
Just a minor correction. Likelihood ratios are mutlipliers of odds rather than probabilities. The probability 0.1 has to be converted to odds 9:1 against (0.111) to be multiplied by the LR and the resultant post test odds have to then be converted back to probability. 0.111 x 9 in this case of sensitivity and specificity equal to 0.9. This post test odds then has to be converted back to probability. In this case 0.111 x 9 =0.999 which is post test odds of 0.4997 rather than 50:50. Pedantic I know.
It also is important for example with d-dimer testing as referenced to John’s latest medscape review of the NEJM ILR study to know what the acceptable miss rate is not just what tests do to pre test probabilities.
A patient with a low pretest probability of PE of 10% (about right for -ve Wells score) with a d-dimer of <500 with -LR of 0.05 would have post test probability of 0.5525% - easily good enough to exclude PE. It is estimated that a d-dimer of 1000-1500 has a LR of 1.0. A d-dimer of 800 has a -LR of about 0.5 and so even if is often called “positive” it actually reduces the post test probability. But the crucial question is does it reduce it enough? And what is the acceptable false -ve rate? The post test probability for 0.1 prevalence and LR of 0.5 is 5.3%. And I think most would not accept that threshold. Most I think accept 1%. LR0.05 gets you there easily for low probability patients so we have just learnt that heuristic and attributed a false dichotomised label to the continuous test.
Great article! I am a Christian so I do believe that everyone already has an appointment with death that has been set and pre-determined.
It doesn’t mean we don’t do the things in this world to improve your health. But no human endeavor can change the fact that your appointment with death is set.
Honestly, knowing that fact, helps me to work tirelessly for my patients since we as doctors have that calling but to also not be utterly crushed by one of my patients dying.
Well said!
Total agreement on the arithmetic: at 1% prevalence a good test still drowns you in false positives, and the near-fatal biopsy of a benign nodule is the harm nobody counts.
Where I want to push, as respectfully as I can, is the slide from "whole-body scanning is folly" to "we die when it is our time, prevention is smart choices and good luck."
Your own example undercuts the fatalism: the cyclist who called his breathlessness asthma had a fixable LAD lesion. That is not luck, that is a missed modifiable cause.
I was that patient, a reassuring workup over an 80% right coronary blockage .
The honest line is not "more is better" versus "nothing is modifiable." It is that a few high-yield, cheap measures with real outcome data (blood pressure, ApoB, Lp(a) once) sit in a different Bayesian universe than a total-body MRI hunting 0.1% disease.
Would you accept that the constrained-vision case argues for a short list of good tests, not for testing nihilism?
Very well put John and I appreciate your passion. I don't see anything in this essay to disagree with. But I also know that you and I and anyone else who thinks this way lost this battle a long time ago. Trying to explain these concepts to my very smart but non-medical family members, for example, has never worked. They seem to understand the larger concept but can't apply it in the real world. My sister frequently says (in regards to screenings generally), "isn't more information just always good?" and I ask her "if that's true, then why not get a screening colonoscopy in your 5-year-old son?" She knows it's a ridiculous question but can't see how the same principle might apply to any medical test in any person: that there are tests that are likely to provide meaningful and actionable results and many more tests that are unhelpful, wasteful, and/or harmful. Knowing the difference or trying to figure it out is so much of what we do in medicine! But so many people really think medicine is just doing tests, the more the better.
I share your perspective. Do the right things for your health (exercise, sleep, avoid harmful habits, etc) and don’t ignore warning signals. I joke that when I’m very old and ready for what’s next, I’ll ware a “Do not resuscitate” tennis shirt every time I play.
Definition of a healthy patient: Someone who hasn't had a proper workup.
Tremendous Dispatch!
Clinical importance often follows hype. A test result you didn't have a plan for may end up killing you. And that's not hype.
Listened to the discussion and appreciated it. I also have read/read Thomas Sowell. Love his take on human behavior. I am not willing to go down the scan/genomic sequencing rabbit hole. For those who do "More power to them" None of us gets out of here alive, do the best you can in the present moment and try to remain healthy
Exceptionally well thought out and written. Kudos.
One of your very best, Dr. Mandrola - hope it is widely read.