This is the most honest thing I have read on the scanning question, and the South Korea thyroid natural experiment is the cleanest proof anyone could ask for: diagnoses up fifteen-fold, mortality flat, and the fix was to stop looking.
Your pre-test-probability point is the hinge, and it is also where I would draw the one line that keeps this from being read as "never image." A whole-body MRI on an asymptomatic 30-something and a calcium score in a 44-year-old with a family history and a high ApoB are not the same Bayesian object: the second one starts from an elevated pre-test probability, which is the exact variable your essay says everyone ignores.
I am the guy who argued his way into an indicated scan that found an 80% blockage, so I am sympathetic to imaging, but the worried-well spa scan is the strawman version of what I did.
The externality argument is sharp. Would you accept "indicated, pre-test-probability-raised testing" as the carve-out, or do you think even that leaks into the cascade?
Your suggestion for "indicated, pre-test probability raised testing" is, at least in theory, what we already have, but that is falling apart, because patients are fundamentally unsatisfied with the current system and looking for alternatives, and who could blame them? The collectivized insurance system has grown so bloated and so expensive from middle men and regulators that patients are going to be willing to pay cash to go outside the system.
The problem now is that when they do, all the findings from their cash-pay care will then get absorbed by the insurance system, because now there will be real indications for follow up testing, with ICD codes that justify it and force the insurance system to pay for it. Costs will go up even more and the collapse of American healthcare will accelerate. I fear it will collapse not into free market healthcare but into single payer, which will not make the care better or cheaper but perhaps more rationed. I wonder how a single payer system would treat such an upstart as Midjourney Medical. Will it simply be regulated out of existence?
Really thought-provoking perspective. Thoughtfully questioning whether an intervention is truly indicated is an essential part of high-value, patient-centered care. At the same time, it’s also very important to remind ourselves that “unindicated” does not necessarily mean “unhelpful” in every circumstance. Clinical decisions depend on the patient’s history, symptoms, risk factors, preferences, and the balance of potential benefits and harms. Avoiding unnecessary interventions is just as important as ensuring appropriate care is not withheld. The challenge is finding that balance through evidence-based, individualized decision-making. Perhaps one of the most important shifts in modern medicine is moving away from the assumption that more care is always better. The highest-quality care is often about choosing the right intervention for the right patient at the right time, not simply doing more. Thanks for sharing this insightful overview.
This is the most honest thing I have read on the scanning question, and the South Korea thyroid natural experiment is the cleanest proof anyone could ask for: diagnoses up fifteen-fold, mortality flat, and the fix was to stop looking.
Your pre-test-probability point is the hinge, and it is also where I would draw the one line that keeps this from being read as "never image." A whole-body MRI on an asymptomatic 30-something and a calcium score in a 44-year-old with a family history and a high ApoB are not the same Bayesian object: the second one starts from an elevated pre-test probability, which is the exact variable your essay says everyone ignores.
I am the guy who argued his way into an indicated scan that found an 80% blockage, so I am sympathetic to imaging, but the worried-well spa scan is the strawman version of what I did.
The externality argument is sharp. Would you accept "indicated, pre-test-probability-raised testing" as the carve-out, or do you think even that leaks into the cascade?
Your suggestion for "indicated, pre-test probability raised testing" is, at least in theory, what we already have, but that is falling apart, because patients are fundamentally unsatisfied with the current system and looking for alternatives, and who could blame them? The collectivized insurance system has grown so bloated and so expensive from middle men and regulators that patients are going to be willing to pay cash to go outside the system.
The problem now is that when they do, all the findings from their cash-pay care will then get absorbed by the insurance system, because now there will be real indications for follow up testing, with ICD codes that justify it and force the insurance system to pay for it. Costs will go up even more and the collapse of American healthcare will accelerate. I fear it will collapse not into free market healthcare but into single payer, which will not make the care better or cheaper but perhaps more rationed. I wonder how a single payer system would treat such an upstart as Midjourney Medical. Will it simply be regulated out of existence?
Really thought-provoking perspective. Thoughtfully questioning whether an intervention is truly indicated is an essential part of high-value, patient-centered care. At the same time, it’s also very important to remind ourselves that “unindicated” does not necessarily mean “unhelpful” in every circumstance. Clinical decisions depend on the patient’s history, symptoms, risk factors, preferences, and the balance of potential benefits and harms. Avoiding unnecessary interventions is just as important as ensuring appropriate care is not withheld. The challenge is finding that balance through evidence-based, individualized decision-making. Perhaps one of the most important shifts in modern medicine is moving away from the assumption that more care is always better. The highest-quality care is often about choosing the right intervention for the right patient at the right time, not simply doing more. Thanks for sharing this insightful overview.