From my friend Ross Levatter, MD.
I’ve had several friends, both on Facebook and in real life, diagnosed with cancer over the last few years. And often it is higher stage cancer as it is typically true that symptoms often develop only after cancer has advanced or spread.
One method of picking up early cancer (or even pre-cancerous lesions in the breast and colon) is via screening. And yet screening is quite a controversial issue in medicine. Some people are concerned that it’s too expensive: in economics terms, that it doesn’t pass a cost-benefit test; that there are too many false positive and false negative work ups, leading to increased expense, unnecessary biopsies, undue patient worry, etc. Many well-respected medical statisticians and epidemiologists argue against almost all screening studies, whether they be via imaging (as in breast cancer screening with mammography) or invasive assessment (as with colonoscopy for colon cancer screening) or even simple blood tests (as with prostate cancer screening).
Given my knowledge of economics, such as it is [DRH note: it’s quite good], I’ve always held these statistical evaluations as somewhat dubious. That’s because value is subjective and determined by willingness to pay; yet virtually no cost-benefit analysis is done from the perspective of a paying patient (strictly, “customer,” because by definition screenings are performed only on asymptomatic people with no known disease.) The question instead is, “Should the government force insurers to cover this screening study for everyone, or for everyone in certain groups [e.g., women over 40 for mammography; smokers with more than a 40 pack-year smoking history for lung CT screens for early lung cancer] or is it too expensive for society?” No one explores: What happens when entrepreneurial physicians offer screening and those people who think it worth the cost, even with risk of false positive and false negative workups, etc. choose to get it done and pay for it themselves?
Because in our society screening is offered only en masse when the government decrees mandatory insurance coverage, doctors spend far more time lobbying Congress regarding the value of screening than advertising to individuals, whose values, after all should count way more than those of members of Congress, regarding the value of screening.
But there ARE screens available for many cancers, including the most common (breast, lung, colon, prostate.) Some radiology groups offer whole body screening with either CT or (without radiation) MRI. These can pick up not only cancers but also other problems like vascular aneurysms or stenoses. Each of us should consider whether the possibility of early cancer detection, even if we have to pay out of pocket, is worth it to us.
Ross is simply applying the main insight from the marginal revolution.
My own personal story. About a year ago, a friend died from a ruptured aortic aneurysm. A number of us, including Ross, were discussing it on Facebook. Ross pointed out that there are often no symptoms before it occurs, so when it occurs, it’s too late. He also noted that the test for an aortic aneurysm is straightforward and that if you have a problem, a surgeon can do something about it. (I told another friend who told me that a doctor acquaintance of his had been flying into Boston when he felt his aorta rupture, and that he knew right then that he would die, which he did.)
I sat with that knowledge for a few months and then arranged to see my doctor. I told her my concern. She asked if I had ever smoked. Answer: no. She asked if there was any history of it in my family, (I think she asked that.) Answer: no. Then she felt around and said she couldn’t detect any problem.
“If you want to proceed,” she said, “I can order the test, but your insurance might not cover it and it costs hundreds of dollars out of pocket.” “Fine,” I said. I got the test and my insurance company did cover it. I didn’t have a problem. But I’m still glad it did it and would have been glad even if I had paid the whole tab.
READER COMMENTS
Jon Murphy
Nov 15 2023 at 8:21am
Ross’s point is important and I think a vital insight of Public Choice (and, more broadly speaking, the LSE school of thought regarding cost that comes from Robbins, Hayek, Buchanan, Coase, and Thirlby). Whenever public policy is discussed, economists tend to say some version of “just calculate the npv and, if positive, do it.”* But the problem with that normative position is that not all costs are measurable in such a manner; they’re subjective and psychological. Indeed, one could reasonably argue that the majority of costs are unseen at the collective level, which suggests that the npv approach (while not wholly unreasonable) biases action toward the “do something” side of things rather than the “let individuals choose” side of things.
Incidentally, I think this reason is why Coase’s paper The Problem of Social Costs is both brilliant and misunderstood. Many read just the first four pages and say “oh, government intervention is only necessary if transaction costs are high.” But that’s only true in the Pigouvian framework. I’m the broader institutional framework that Coase discusses (and takes Pigou to task for missing), various rules, laws, rights, agreements, etc all developed to take care of various problems while taking into account the subjectivity of costs. Government intervention, even in a supposed market failure, often overrides these institutions and can make the problem worse because the intervention doesn’t account for subjective costs.
*I’m being a little glib here, but only a little.
Thomas L Hutcheson
Nov 17 2023 at 11:01am
But does not the problem of hard-to-measure “subjective” benefits and costs plague any consequentialist decision rule?
More narrowly, isn’t the NPV rule itself an example of an evolved response to “If it feels good, do it!” 🙂
Kevin Corcoran
Nov 15 2023 at 11:29am
Very good post.
My mother was diagnosed with pancreatic cancer earlier this year. Like most cases of pancreatic cancer, it was not detected until quite late, because it causes somewhere between no symptoms and minimal symptoms that are easily overlooked, until it’s very far advanced. This has changed my own evaluation of various kinds of screenings discussed here, particularly since there can be a genetic component to these diseases. My own kids are quite young (particularly compared to most guys my age), and to me, it would be worth quite a bit of out of pocket expense to keep a proactive eye out for things like this, to ensure that if anything has the potential to go wrong in my case it’s caught as early as possible.
But that’s my individual calculation, based on my family history, personal preferences, and subjective evaluations for my own case. My own interest in more proactive screening doesn’t mean more screening is better for the population in general, or anyone else in particular.
David Henderson
Nov 15 2023 at 12:07pm
Thanks, Kevin.
Good comment by you.
But I can’t resist a little humor. You wrote, “My own kids are quite young (particularly compared to most guys my age).” That has to be true.
Kevin Corcoran
Nov 15 2023 at 12:47pm
I’ll admit, the humor definitely made me chuckle out loud!
steve
Nov 15 2023 at 11:46am
The executive physical/screening tests has become pretty popular. Anyone can do it and IIRC most of those getting them are not executives. Many hospital and screening companies offer these. Also while some stats people think screenings are not cost effective others do, so it’s not clear who is correct. Some of it has to do with QALY calculations and there are big disagreements about those. My suspicion is that those individuals for whom QALY is low probably dont go to screening anyway. (When doing econ papers people usually use a uniform QALY value for everyone but I think its pretty clear its different for individuals.)
Steve
Joel N Pollen
Nov 15 2023 at 12:59pm
I am among those who usually argue against cancer screening, but I think the points raised here are correct. In a sense, all efforts to quantify the value of a screening one way or the other are useless. We should just offer them for sale on the open market and whatever happens happens. I think uptake will be fairly low, but certainly some people will do it.
Joel N Pollen
Nov 15 2023 at 1:03pm
I think it’s worth adding that most studies purporting to show at cost effectiveness, benefit of cancer screening or fundamentally flawed because most cancer screening techniques have not been shown to extend lifespan. They have only been shown to reduce cause specific mortality, but not all cause mortality. The main exception would be flexible sigmoidoscopy for colon cancer screening.
I believe the main reason for this failure is the problem of competing risk. As you age, your chance of dying from many, many different things goes up roughly commensurately. Preventing just one of those causes of death has very little effect on your lifespan.
Jose Pablo
Nov 15 2023 at 2:08pm
Preventing just one of those causes of death has very little effect on your lifespan.
That’s comforting, thank you!
Since I am, most definitely, aging and so, quitting wine (“just one of those causes”) has very little effect in my lifespan, I am going back to it today. First order of business!
Jose Pablo
Nov 15 2023 at 2:13pm
That would make “screening”, by the way, a wonderful example of the “illusion of control” bias, that sure have an effect on our “subjective evaluation of value”.
By the way, what’s the difference between a “subjective evaluation of value” and a “wrong / biased evaluation of value”?
Ross Levatter
Nov 15 2023 at 2:41pm
Jose asks: By the way, what’s the difference between a “subjective evaluation of value” and a “wrong / biased evaluation of value”?
Here’s an example of my understanding of the difference:
Subjective evaluation of value: Your family has a strong history of colon cancer. You talk with your sister about the value of getting a screening colonoscopy. You both look at all the objective factors (cost, risk of complications, etc.). You decide it is worthwhile to get the study. Your sister, just diagnosed with stage 4 breast cancer, decides it is not worthwhile for her. Values that change based on the subject are subjective values.
Wrong/biased evaluation of belief: You have a smooth stone you polish every day. You value it highly, because you have become convinced polishing the stone keeps you alive. You had been rubbing it once when you and a friend you were walking were mugged. He was shot dead but you were fine. This is of course an emotional and irrational belief. Values based not based on reasons or preferences, but on whim, hope, desire are wrong/biased values. They still motivate, of course. You continue to polish the stone.
Jose Pablo
Nov 15 2023 at 3:29pm
Thank you, Ross.
But then, by your understanding and taking into account what Joel said (which makes a lot of sense), screening for cancer is closer to “rubbing the stone”: it has no significant influence on your lifespan. Which should be the key metric since I am pretty much indifferent to my actual cause of death.
The “when” is the all relevant variable and it doesn’t seem to change due to screening much more than it changes due to rubbing the stone (if I understand Joel right).
[In my opinion “subjective” and “biased” are empirically indiscernible and so, following Leibniz, should be considered ontologically the same and theories based on their difference rejected]
Ross Levatter
Nov 15 2023 at 5:23pm
Well, Jose. I’d have to disagree that “subjective” and ‘biased” are practically indiscernible. For one thing, there are frank preferences. I prefer vanilla. You prefer chocolate. This is perfectly understandable and it would be strange to refer to either of our subjective preferences as “biased.” Then consider the brother/sister example I gave above. It is perfectly clear why a woman with stage 4 breast cancer doesn’t see value in getting screened for colon cancer. That is not bias on her part. But it IS subjective. Subjective simply means, as economists use the term, related to the deciding subject.
Joel’s point about overall effect on your lifespan, I think, actually proves my argument. In a large population, that may well be true that lifespan is unaffected. But if YOU have a colonoscopy at age 50 and they find a stage 1 colon cancer that can be surgically excised, you may well live to your normal expected age of 80 or more. If you had waited until it symptomatically presented to treat it (probably before you were 60), it would likely be a stage 3 or 4, and you could be dead within 5 years. What is true for large populations statistically need not be true for your specific situation individually. Is it worth paying to find out? It’s subjective. My point.
Finally, I should note that Joel’s point (I think) refers to specific screening protocols. It may well be true that if a woman religiously gets her mammogram every year and avoids breast cancer, she may still die young from colon or lung cancer. That’s one reason it may be true there is potential benefit of whole body imaging for screening—studying the lungs, the liver, the kidneys, the pancreas, the colon, and other organs all in one exam. Having read CTs for several decades, I cannot tell you how often I did an exam to look for one thing and ended up incidentally finding another thing. Chest CT for cough and the chest is fine but I note a cancer on the top of the left kidney on the lowest slice. Mammogram for palpable mass in the right breast. The palpable mass is on imaging a clearly benign calcified fibroadenoma, but there’s an obvious though non-palpable cancer in the left breast. These may well not be statistically significant. Just like it’s statistically unlikely you’ll win the lottery. But some people do demonstrate they value buying lottery tickets. And SOMEONE wins.
Jose Pablo
Nov 15 2023 at 7:04pm
Subjective simply means, as economists use the term, related to the deciding subject.
Then, Ross, by this definition, all “biased” evaluations are subjective (even if, as you point out, not all subjective evaluations are biased).
For both things to be indiscernible is enough with the set of biased decisions being indiscernible from a subset of all the subjective decisions.
The “illusion of control” bias sure plays a role when individuals decide if/when to do a screening. Even if there is a chance of this screening expanding their lifespan (as you argue), the individual will very likely overestimate that chance.
While, at the same time, he/she will underestimate the additional suffering of anticipating the diagnosis. Never do the screening just before your daughter’s wedding or just before your trip to Bali for your silver wedding anniversary.
We, humans, have very well documented biases when making economic and political decisions.
My general point is, I think, that saying that “subjective (as in “related to the deciding subject“) value” “should” be the key force in decision making, is the same as saying that human biases (as in “perception mistakes“) “should” be the key force behind human decision making.
Thomas L Hutcheson
Nov 16 2023 at 7:47am
But “very little” is not zero. Maybe the problem is getting cheaper screening tests and doctors who will decide with the patient whether to do the screening or not.
Joel N Pollen
Nov 16 2023 at 1:00pm
You’re right, in principle, that very little is not zero. It’s just that there are many other things we could be doing with our time and money. If screening has such a small effect on mortality that it can’t be detected in a study with hundreds of thousands of participants, it’s probably not the best use of our limited resources.
For the more, it’s probably not something very many people would choose to spend their money on. The list of things that you could do which would have a tiny protective effect on your risk of death is nearly infinite, and many of them are clearly not worth doing. The ones that are worth doing tend to be the ones that are extremely low cost (e.g. buckling your seatbelt). Mammograms, CTs, and colonoscopies are not going to be as convenient as a seat belt any time soon.
David Henderson
Nov 16 2023 at 2:21pm
You write:
Your language may be steering you away from Ross’s point. It’s your use of the word “our.” Ross’s point is that people should be free to spend their own money as they wish, no matter how small the probability.
Thomas L Hutcheson
Nov 17 2023 at 11:10am
I agree with your counter point. The question is how to use large scale statistical results for the individual patient. And I think the role of the doctor in a) understanding and b) conveying the costs and benefits to the patient is crucial.
Also, the more that screening is done on the basis of individual cost-benefit analysis the greater the market incentive will be to find less costly screening tests.
Philo
Nov 15 2023 at 2:15pm
Value is subjective? Well, on any plausible definition of ‘subjective’, your ex ante estimate of value-to-you is “subjective,” and goes to determining your willingness to pay. (You don’t usually pay based on your estimate of value-to-others.) But your ex post judgment may be quite different, and–since you are benefiting from hindsight–the latter judgment is likely much closer to the actual value(-to you) than is the former.
So is actual value–as distinct from your various opinions about it–subjective”? To answer this, we may need a precise definition of this rather vague term.
The distinction between intrinsic and instrumental value may be relevant here, since medical diagnostic procedures are chosen primarily on the basis of their instrumental value. There is probably a plausible sense of ‘subjective’ in which instrinsic value is purely “subjective,” but instrumental value, because of its causal component, is only partly so.
Ahmed Fares
Nov 15 2023 at 4:26pm
Joel N Pollen,
Your comment nailed it, and for exactly the reason you gave.
The following quote is from a site titled: “Harding Center For Risk Literacy”.
https://www.hardingcenter.de/en/transfer-and-impact/fact-boxes/early-detection-of-cancer/early-detection-of-breast-cancer-by-mammography-screening
For other diseases from the same site, the link to the main page. (The links are on the right side of the page.)
https://www.hardingcenter.de/en/transfer-and-impact/fact-boxes
Andrea Mays
Nov 15 2023 at 5:20pm
Interesting. I missed your FB discussion entirely, somI am glad to have seen this. My data are out of date, but my recollection is that under National Health Service in the UK, health screenings stop at a relatively young age— by fiat. You won’t even be aware (via advertising or PSAs for example) that such screenings are available. At 75 my father was well past the age at which NHS screened for cancers he would routinely have been screened for in the US. As in the US, subjective value is not taken into account in that nationalized system. Is it the same in Canada?
BS
Nov 16 2023 at 1:56pm
In Canada, I suppose it depends on procedure. For example, colonoscopy might not be done past 80, unless you are in good physical health. Canada’s problem right now is access to the front line – GPs. Governments are essentially rationing care and imposing wage controls by limiting private enterprise.
Thomas L Hutcheson
Nov 16 2023 at 7:37am
Just because insurance will cover it does not mean that a doctor can’t decide with his patient if the costs and benefits make sense or not. The fundamental problem is that doctors are not trained in thinking in cost-benefit terms.
Jon Murphy
Nov 16 2023 at 9:14am
I don’t think that’s the fundamental problem here. It seems to me the problem is even more Econ 101 than that: when an individual doesn’t face all the costs and benefits of an action, there is not a fully aligned incentive to get the choice correct. When insurance covers some (or all) of a procedure, neither the doctor nor the patient have the full incentive to make an optimal decision. Since some of the cost is borne by a third party, too much of the good (in this case, screenings) come to the market.
Jon Murphy
Nov 16 2023 at 9:17am
One other comment: most people are not trained in cost-benefit analysis (and, I’d argue per my comment above, many who are are trained incorrectly) but that is irrelevant. You do not need to be trained in cba before you do it. Rather, economic cba is a description of how people act
Jose Pablo
Nov 17 2023 at 6:42pm
cba is a description of how people act
then it is a very bad description.
That’s a great opportunity for a book: “The Myth of the Cba Performing Consumer”
or even better: “The Myth of the Cba Performing Political Leader”
Jose Pablo
Nov 17 2023 at 6:51pm
doctors are not trained in thinking in cost-benefit terms.
That sounds like Taleb’s concept of “teaching birds how to fly”.
Cost-benefit analysis pretends to be a description of how consumers behave … and then, you use it to teach consumers how to behave … it doesn’t make any sense!
[It would be patients / consumers, no doctors, the ones in charge of doing the cost benefit analysis but the “teaching birds how to fly” idea applies the same]
Peter
Nov 19 2023 at 2:19am
Just a comment but the real problem is the fact doctors have any say at all. Like in this case it’s nice David’s doctor was willing to order the test as long as David was willing to eat the cost of insurance didn’t cover but in my experience that is a rare doctor. My subjective experience is doctors actively fight patients on anything from tests to medication that they doctor themselves didn’t spontaneously come up with in their own volition.
I shouldn’t have to fly to Mexico to get a flu test, a skin biopsy, or get the HPV vaccine because my doctor arbitrarily disagrees.
David Marsh
Nov 17 2023 at 7:23am
Great Post! The acknowledgment that value is subjective has far-reaching implications across various disciplines. It underscores the importance of individual perspectives, cultural diversity, and the dynamic nature of human preferences in shaping how we assign value to objects, experiences, and ideas.
National Jester
Nov 17 2023 at 5:32pm
I have had three colonoscopies so far. Pre-cancerous growths were found and removed each time. I’m on the 5year plan. If you don’t have them, you get checked every ten years. But there are a lot of screening tests that are not expensive, such as PSA for prostate cancer. I think that we are going to see more blood marker tests in the future. BTW- A friend was recently diagnosed with stage four lung cancer. His cost for the treatment is $3600 a month. Don’t know what his insurance and Medicare pay.
Comments are closed.