I've just discovered Sensible Medicine and am enjoying it very much. I'm a general and vascular surgeon who practiced for thirty-six years in rural Wisconsin, and I did my formative schooling in England, right up through high school. Your common-sense approach, questioning whether tests truly help patients, is something I've preached my whole career: a careful history and a good physical exam still matter.
I've tried to write about the human side of surgery in two memoirs, Life in the Pits and Life on the Cutting Edge, and I've recently started a Substack at nicholasmaxwellmd.substack.com.
I'd love to stay in touch and would be delighted to compare notes.
I just looked up “Ending Medical Reversal” on Amazon. It looks like an incredibly important book packaged uglier than an Amazon delivery, and honestly the title sucks, too. No wonder footnote 3 was a joke. Give it a snappy consumer-friendly makeover and it might land, as it needs to, on the best-seller list. I'm a retired medical writer, clinical and consumer, I'd come out of retirement to ghost it with you.
It's interesting that the saying, 'absence of evidence is not evidence of absence' is often stated as if it's axiomatic, when it's actually incorrect.
Absence of evidence isn't proof of absence, which may be where the confusion arises. But if there's a reasonable expectation that evidence would have been found if something did actually work, then absence of evidence becomes evidence of absence. It might only be weak evidence, depending on the circumstances, but it's evidence all the same. Any good Bayesian would use it to lower their prior probability that a treatment works.
What happened to the good ol CASH brace? I use to put a lot of those on in the 90s and early 2000s before The Plasties. Little hard to do a Double Blind for a Brace like that. They are not very comfortable. Difficult to put on for some.
If my memory serves me well they had a track record of either giving immediate relief or immediate pain. Likely depended on the location of the fracture and where the posterior strap contacted the back and therefore the fracture site and any soft tissue trauma from the injury. Also not tolerated well but the 70+ crowd especially if they already had severe kyphosis before their fracture. Would like to know opinions on bracing for compression fractures.
Another factor is that patients may feel improvement (perhaps placebo, but at any rate improvement), but later suffer another fracture above or below because the cement allows no movement and stresses the vertebra above or below.
I wonder if there are studies discussing the number of subsequent fractures that later occur.
The barriers to doing an RCT for this procedure are multiple:
Lead investigator: this has to be someone whose income does not depend on surgical volume, and who has training and interest in performing clinical studies with good statistical support. Most likely a radiologist at a large academic center with an active biostatistics department.
Funding - because the study is randomized, neither patients nor insurance can be billed for any part of either arm of the procedure. Therefore, each participating hospital must be reimbursed for the costs of all personnel, room usage, materials, etc. The only funding agency with those kinds of dollars is the NIH. So the lead investigator must learn how to write a large grant and convince the NIH that they have the expertise and support to do the study. This will likely take at least 3 submissions, so probably two years to get a fundable score.
Patient selection - for a first trial, you need patients who are ideal candidates, and with the ability to provide the follow up data required. Not an easy task when patients are likely quite elderly, have multiple co-morbidities, and have to survive long enough to provide data at, say, 1 year after the procedure. And for a definitive trial, you need a LOT of patients, so the generalizability of the study can be wide. A study with enough patients will be very expensive indeed.
It is frustrating and sad that the structural problems of our health care system combine to make necessary studies like the one Adam proposes so difficult to do. We must look to more advanced countries, like Canada, Australia, or Denmark to carry such a study out
Those who do it are doing it for the money. Period. Full stop. Next. The only place where the patient factors into their decision-making is as the vector to their bank account.
Sure, absence of evidence of benefit is not evidence of absence of benefit. But it sure as bloody hell is not evidence of presence of benefit either.
This is a tough one. I almost never refer people for invasive treatment for compression fractures. These are typically extreme elderly women. The times I've seen this invasive intervention pushed for is by a zealous adult child who does not believe that pain cannot be relieved or fixed easily. As pointed out, the focus on "do something!" that affects at least half the population will be at play. Every physician has their own personal bias/approach as a more aggressive "doer" or a "less is more" type approach and each patient and their family same thing. The best guide in the care of the extreme elderly is a very experienced clinician who has watched all of this play out for decades and understands the supports needed to get someone through this. The acute pain throws the elderly person completely out of their normal balance.It will take a fair amount of monitoring and support to assist with pain meds, mitigating side effects, supporting general function and lack of independence. Calcitonin is probably only used by those of us old enough to have practiced when it was a "thing" - I'm not sure a younger physician would even know of its existence. I agree with the commenter that this is not likely a significant portion of any Medicare "spend" and other areas would be much more impactful to consider. And, yes, wouldn't it be truly lovely to have actual evidence as we practice EBM?
I wish we were able to educate patients about available therapeutic options.
I was hoping AI would help the medical literacy of non-medical people.
Out of curiosity, I asked the Google Gemini - " I am 82 year old and have a fracture in my backbone. Should I have a balloon kyphoplasty?" and it answered,
"Balloon kyphoplasty is widely considered safe and effective for patients in their 80s, though whether it is right for you depends on your pain level and how recently the fracture happened. Being 82 does not rule you out, as the procedure is minimally invasive and routinely performed on older adults with osteoporosis." . This was the answer on the top and the source was Hospital of Special Surgery in NYC - the major orthopedic surgery hospital.
AI may even increase the health care cost, as it just pushes the existing medically inaccurate narrative that a lot of health care marketing people working for the hospital and the industry are propagating.
I am hoping a new group of senates will shed the light to this practice and empower FDA to oversight the clinical studies run by the industry, and make sure only the effective intervention/treatments are reimbursed at a negotiated price.
I thought we threw this out 20 years ago. In its heyday we tried it on a retired neurosurgeon colleague. The kyphoplasty blew out the vertebrae above and below the "treated" one! I have't seen it being done in years. Well, I guess 20 years is enough time for a new generation to experiment with failed technology. Sorta like Watchman.
Medical devices are evaluated/ approved through a FDA process prior to Medicare certification for payment. There are several pathways, but the process and evaluation must be followed.
Pharmaceutical agents may be approved by the FDA for one indication supported by clinical evidence, but then may be used in off label conditions or at doses that have not been evaluated.
I do not find the FDA to be a reliable screen, since it seems to be an example of regulatory capture. It is beholden to and deeply intertwined with the industries it regulates.
Going from vertebroplasty to kyphoplasty without a single sham-controlled trial is striking, after the 2009 trials showed placebo doing just as well. A procedure that makes physiological sense is not the same as a procedure that works.
Pair the do something bias with broad spectrum financial incentives, and you get the juggernaut of ever-rising medical insurance costs with justified cynicism to boot.
That pairing explains a lot. When doing something is both what the patient hopes for and what gets paid, the sham-controlled trial that might say no is the one nobody rushes to run.
The problem here is not whether kyphoplasty works, it can and does produce dramatic clinical improvement in pain and patient function. The key, like so much in medicine, is to identify specific patients who will benefit from the procedure from those who will not. There is also wide agreement that the procedure is safer than vertebroplasty not just only because the placement of cement can be better controlled, but also because the bone cement used has a lower concentration of “free monomer”, a component that can cause cardiovascular collapse when leaked into blood vessels.
Unfortunately, clinical trials often sink on patient selection criteria because these are used to select entrants into the trial. We typically perform review the data after all enrollments and treatment are completed to see if particular subgroups demonstrate differences in outcome.
A large government funded study of epidural steroid injections that was published in The New England Journal of Medicine (Friedly, et al) used clinical selection criteria that did an extremely poor job of identifying patients who could ever be successfully treated with this modality, and as one would expect, the outcomes were terrible in all groups. The principal investigator for the study was an expert in upper extremity prosthetics, but had no prior expertise or other publications in this area (treatment of lumbar radiculopathies). When politics influence The conduct of medical studies, Truth suffers.
How are you - from an "all comers" standpoint, determining who is most likely to benefit from the procedure? As a primary care internist, I see a lot of compression fractures. Majority are in the extreme elderly. We are supposed to give conservative care a good 6 week trial. Where and when would you possibly be filtering "best candidates"? The population itself - women > 85 years old - is high risk for complications from "procedures". The adage I live by is the less we "do" to 90+ year olds, the better. And generally I've been proven correct. Where do you have a robust sorting protocol to tease out who would potentially benefit more than someone else?
Patient selection is a fair point, and probably the most useful thing a good trial could answer: not whether it works for everyone, but which fractures and which people actually get the benefit.
Dear Sensible Medicine team,
I've just discovered Sensible Medicine and am enjoying it very much. I'm a general and vascular surgeon who practiced for thirty-six years in rural Wisconsin, and I did my formative schooling in England, right up through high school. Your common-sense approach, questioning whether tests truly help patients, is something I've preached my whole career: a careful history and a good physical exam still matter.
I've tried to write about the human side of surgery in two memoirs, Life in the Pits and Life on the Cutting Edge, and I've recently started a Substack at nicholasmaxwellmd.substack.com.
I'd love to stay in touch and would be delighted to compare notes.
With warm regards,
Nicholas Maxwell, MD
General and Vascular Surgeon
I just looked up “Ending Medical Reversal” on Amazon. It looks like an incredibly important book packaged uglier than an Amazon delivery, and honestly the title sucks, too. No wonder footnote 3 was a joke. Give it a snappy consumer-friendly makeover and it might land, as it needs to, on the best-seller list. I'm a retired medical writer, clinical and consumer, I'd come out of retirement to ghost it with you.
It's interesting that the saying, 'absence of evidence is not evidence of absence' is often stated as if it's axiomatic, when it's actually incorrect.
Absence of evidence isn't proof of absence, which may be where the confusion arises. But if there's a reasonable expectation that evidence would have been found if something did actually work, then absence of evidence becomes evidence of absence. It might only be weak evidence, depending on the circumstances, but it's evidence all the same. Any good Bayesian would use it to lower their prior probability that a treatment works.
You articulated exactly the response I was going to make upon reading the 'absence' statement. A case of failure to distinguish evidence from proof.
What happened to the good ol CASH brace? I use to put a lot of those on in the 90s and early 2000s before The Plasties. Little hard to do a Double Blind for a Brace like that. They are not very comfortable. Difficult to put on for some.
If my memory serves me well they had a track record of either giving immediate relief or immediate pain. Likely depended on the location of the fracture and where the posterior strap contacted the back and therefore the fracture site and any soft tissue trauma from the injury. Also not tolerated well but the 70+ crowd especially if they already had severe kyphosis before their fracture. Would like to know opinions on bracing for compression fractures.
Thank you, again, for your thoughtful reviews.
Another factor is that patients may feel improvement (perhaps placebo, but at any rate improvement), but later suffer another fracture above or below because the cement allows no movement and stresses the vertebra above or below.
I wonder if there are studies discussing the number of subsequent fractures that later occur.
The barriers to doing an RCT for this procedure are multiple:
Lead investigator: this has to be someone whose income does not depend on surgical volume, and who has training and interest in performing clinical studies with good statistical support. Most likely a radiologist at a large academic center with an active biostatistics department.
Funding - because the study is randomized, neither patients nor insurance can be billed for any part of either arm of the procedure. Therefore, each participating hospital must be reimbursed for the costs of all personnel, room usage, materials, etc. The only funding agency with those kinds of dollars is the NIH. So the lead investigator must learn how to write a large grant and convince the NIH that they have the expertise and support to do the study. This will likely take at least 3 submissions, so probably two years to get a fundable score.
Patient selection - for a first trial, you need patients who are ideal candidates, and with the ability to provide the follow up data required. Not an easy task when patients are likely quite elderly, have multiple co-morbidities, and have to survive long enough to provide data at, say, 1 year after the procedure. And for a definitive trial, you need a LOT of patients, so the generalizability of the study can be wide. A study with enough patients will be very expensive indeed.
It is frustrating and sad that the structural problems of our health care system combine to make necessary studies like the one Adam proposes so difficult to do. We must look to more advanced countries, like Canada, Australia, or Denmark to carry such a study out
Call me a pessimist and cynic at this point.
Those who do it are doing it for the money. Period. Full stop. Next. The only place where the patient factors into their decision-making is as the vector to their bank account.
Sure, absence of evidence of benefit is not evidence of absence of benefit. But it sure as bloody hell is not evidence of presence of benefit either.
This is a tough one. I almost never refer people for invasive treatment for compression fractures. These are typically extreme elderly women. The times I've seen this invasive intervention pushed for is by a zealous adult child who does not believe that pain cannot be relieved or fixed easily. As pointed out, the focus on "do something!" that affects at least half the population will be at play. Every physician has their own personal bias/approach as a more aggressive "doer" or a "less is more" type approach and each patient and their family same thing. The best guide in the care of the extreme elderly is a very experienced clinician who has watched all of this play out for decades and understands the supports needed to get someone through this. The acute pain throws the elderly person completely out of their normal balance.It will take a fair amount of monitoring and support to assist with pain meds, mitigating side effects, supporting general function and lack of independence. Calcitonin is probably only used by those of us old enough to have practiced when it was a "thing" - I'm not sure a younger physician would even know of its existence. I agree with the commenter that this is not likely a significant portion of any Medicare "spend" and other areas would be much more impactful to consider. And, yes, wouldn't it be truly lovely to have actual evidence as we practice EBM?
I wish we were able to educate patients about available therapeutic options.
I was hoping AI would help the medical literacy of non-medical people.
Out of curiosity, I asked the Google Gemini - " I am 82 year old and have a fracture in my backbone. Should I have a balloon kyphoplasty?" and it answered,
"Balloon kyphoplasty is widely considered safe and effective for patients in their 80s, though whether it is right for you depends on your pain level and how recently the fracture happened. Being 82 does not rule you out, as the procedure is minimally invasive and routinely performed on older adults with osteoporosis." . This was the answer on the top and the source was Hospital of Special Surgery in NYC - the major orthopedic surgery hospital.
AI may even increase the health care cost, as it just pushes the existing medically inaccurate narrative that a lot of health care marketing people working for the hospital and the industry are propagating.
I am hoping a new group of senates will shed the light to this practice and empower FDA to oversight the clinical studies run by the industry, and make sure only the effective intervention/treatments are reimbursed at a negotiated price.
Good perspective. Thanks.
I thought we threw this out 20 years ago. In its heyday we tried it on a retired neurosurgeon colleague. The kyphoplasty blew out the vertebrae above and below the "treated" one! I have't seen it being done in years. Well, I guess 20 years is enough time for a new generation to experiment with failed technology. Sorta like Watchman.
Medical devices are evaluated/ approved through a FDA process prior to Medicare certification for payment. There are several pathways, but the process and evaluation must be followed.
Pharmaceutical agents may be approved by the FDA for one indication supported by clinical evidence, but then may be used in off label conditions or at doses that have not been evaluated.
I do not find the FDA to be a reliable screen, since it seems to be an example of regulatory capture. It is beholden to and deeply intertwined with the industries it regulates.
Going from vertebroplasty to kyphoplasty without a single sham-controlled trial is striking, after the 2009 trials showed placebo doing just as well. A procedure that makes physiological sense is not the same as a procedure that works.
Oooooh back surgery…once that door is open…😱
Regarding why our medical spend is so high...I'll bring up the MASSIVE Medicaid and Medicare fraud before BK.
These doctors, pharmacists, therapists, Learing centers, and patients need to be locked up and denaturalized (if applicable) and deported.
Do you think Patients also should be locked up, since they demand DO something?
If they knowingly participated in the criminal enterprise, hell yes.
Is it possible that Medicaid and Medicare fraud are rampant nationwide and not just where you think it is and by who you think it is?
Pair the do something bias with broad spectrum financial incentives, and you get the juggernaut of ever-rising medical insurance costs with justified cynicism to boot.
That pairing explains a lot. When doing something is both what the patient hopes for and what gets paid, the sham-controlled trial that might say no is the one nobody rushes to run.
The problem here is not whether kyphoplasty works, it can and does produce dramatic clinical improvement in pain and patient function. The key, like so much in medicine, is to identify specific patients who will benefit from the procedure from those who will not. There is also wide agreement that the procedure is safer than vertebroplasty not just only because the placement of cement can be better controlled, but also because the bone cement used has a lower concentration of “free monomer”, a component that can cause cardiovascular collapse when leaked into blood vessels.
Unfortunately, clinical trials often sink on patient selection criteria because these are used to select entrants into the trial. We typically perform review the data after all enrollments and treatment are completed to see if particular subgroups demonstrate differences in outcome.
A large government funded study of epidural steroid injections that was published in The New England Journal of Medicine (Friedly, et al) used clinical selection criteria that did an extremely poor job of identifying patients who could ever be successfully treated with this modality, and as one would expect, the outcomes were terrible in all groups. The principal investigator for the study was an expert in upper extremity prosthetics, but had no prior expertise or other publications in this area (treatment of lumbar radiculopathies). When politics influence The conduct of medical studies, Truth suffers.
How are you - from an "all comers" standpoint, determining who is most likely to benefit from the procedure? As a primary care internist, I see a lot of compression fractures. Majority are in the extreme elderly. We are supposed to give conservative care a good 6 week trial. Where and when would you possibly be filtering "best candidates"? The population itself - women > 85 years old - is high risk for complications from "procedures". The adage I live by is the less we "do" to 90+ year olds, the better. And generally I've been proven correct. Where do you have a robust sorting protocol to tease out who would potentially benefit more than someone else?
Patient selection is a fair point, and probably the most useful thing a good trial could answer: not whether it works for everyone, but which fractures and which people actually get the benefit.