Arnold Kling  

Cancer Screening Costs and Benefits

PRINT
End the IMF... Greg Clark at the Movies...

According to Robin Hanson, the latter are statistically small.


While cancer screening does consistently lead to more cancer detection and more cancer treatment, it consistently doesn't lead to lower mortality.

This is based on studies of breast cancer, colon cancer, prostate cancer, and lung cancer.

Keep this in mind when someone tells you that other countries have shown how to reduce the cost of health care. No other country is as fanatical as we are about doing cancer screening. My guess is that it is not the magical efficiency of single-payer health care that holds down spending in other countries. Instead, they have chosen not to budget for cancer screening to the extent that we have.

If we adopt a single-payer system here, my guess is that we will not cut back on cancer screening. If anything, we may do more of it.

In Massachusetts, health care reform that was promised to be cost-reducing in fact raised health care spending. I predict that will be true for any health care reform that is not market-oriented. The reformers will project a lower path for spending, and what they will get will be a higher path.


Comments and Sharing





COMMENTS (15 to date)
Thucydides writes:

From the individual point of view, it is worth protecting against a small risk of a catastrophic outcome by finding cancer early when it is treatable. From the state's budget point of view, since few will be saved, why bother?

In the UK, I have heard PSA tests are not given until age 70. As a result, they have a high number of advanced prostate cancers found, and a much higher death rate than we do. But it is much cheaper. There is no point in operating on these advanced cancers (expensive); they just give hormone therapy which may work for a while.

Tom Ault writes:

I imagine that if the US adopts a single-payer healthcare policy, it will become one more benefit people expect to receive without paying taxes for (or more accurately, expect someone else - either in the present or in the future - to pay taxes for), leading to a rapid increase in the national debt and an eventual solvency crisis.

Thomas DeMeo writes:

This is very interesting information, but clearly something is missing in our understanding here.

These are not facts which are ready for use. They tell us we need to look more deeply and find out what is going on. Anyone who makes any judgements from this is making a mistake.

Dan Weber writes:
From the individual point of view, it is worth protecting against a small risk of a catastrophic outcome by finding cancer early when it is treatable

Is it? I don't think that follows. As Robin was quoted above, it doesn't lead to lower mortality. The expected extension of your life by finding and treating the cancer is offset by the expected shortening of your life by the screening and/or side effects of procedures done in response to false positives.

It does cost more money, though.

Shangwen writes:

Go to all the advocates of "prevention" and early detection, and ask them if they've heard of Bayesian probability, or indeed any of the dismal economic reviews that most prevention and screening programs have. They will just look confused or irritated.

I do agree with Thomas DeMeo that the information as reported shouldn't be treated as a slam-dunk against screening, but there is plenty of better literature out there on mammograms, prenatal ultrasounds, psychological screening, etc that essentially supports the arguments that (1) screening is over-used in most cases, and (2) prevention rarely lowers costs.

ajb writes:

Shangwen is right. I am often surprised by the ignorance even of professional researchers. I remember hearing a medical professor citing research about the improved five year survival statistics from early cancer detection (I think it was breast cancer). When I asked him if they corrected these 5 yr numbers for the fact that a greater number of these cancers were being detected at an earlier stage he just looked puzzled. It took over an hour to convince his coauthor/junior researcher that there was a problem with his reported stats.

Floccina writes:

Thucydides Wrote
In the UK, I have heard PSA tests are not given until age 70. As a result, they have a high number of advanced prostate cancers found, and a much higher death rate than we do.

Or it could be that we find more prostate cancers that would not kill treated or untreated.

Kevin Driscoll writes:

I don't think I understand. How do we measure mortality? From reading Hanson's post it seems like they're saying that with and without screening, the same % of people die from cancer. I don't know that that is terribly surprising, because it seems like if you pull in a random group that hasn't been screened, the breakdown of what % is healthy, what % is stage I, what % is stage II, etc. will be the same as it was before. Assuming there are no advances in treatment, you're just as likely to die of Stage IV cancer as before, the screening just tells you that you have it.

That said, I believed that it was relatively well supported that you are more likely to die after being diagnosed with Stage IV cancer than from being diagnosed with Stage I cancer (for just about any kind of cancer). So, these studies compared people who were screened and diagnosed with people who found out they had cancer through some other method (ie not preventative screening). Basically, it seems like that means that you are just as likely to be screened and have Stage II cancer as you are to develop symptoms and then be diagnosed with Stage II cancer because of those symptoms. I still don't understand why I shouldn't want to get screened. Screening decreases my mortality rate if it discovers cancer at an earlier stage than it would have otherwise been discovered. In all other cases it does nothing to help me.

What we really need is a study of a bunch of people who we test and confirm are healthy. Then, we divide them into two groups randomly. We screen both groups at the recommended intervals. In the first group we inform and treat them if they test positive from their screening. In the second group we record their illness, but we don't tell them or treat them until they otherwise figure out that they have cancer. Obviously, we can't tell them what we're screening for or that they're in a cancer study. Compare the mortality rate due to cancer (or the all-mortality rate). I don't understand how the people who get screened and told won't die at a lower rate than those that we don't tell (which is analogous to not getting screened).

Can anyone help explain this to me?

Dan Weber writes:

See also http://en.wikipedia.org/wiki/Will_Rogers_phenomenon

When the Okies left Oklahoma and moved to California, they raised the average intelligence level in both states.

ThomasL writes:

@DanWeber

Nice reference.

Tyler writes:
My guess is that it is not the magical efficiency of single-payer health care that holds down spending in other countries. Instead, they have chosen not to budget for cancer screening to the extent that we have.

While I agree that a single-payer system is not the main cause of the differences in healthcare spending between the US and other countries, the suggesting that we spend twice as much as other countries because of cancer screening is absurd. I highly doubt half our healthcare bill is spent on cancer in general, let alone on the difference between what we spend of screening and treatment of those screened and found positive.

Shangwen writes:

On a related note, Austin Frakt has two posts today (here and here) about the impact of cost sharing as discussed in two recent papers, and notes that consumers are as likely to cut back on low-value health care (e.g., screening) as they are on high-value health care.

Information about what procedures are low-value is not kept in a locked chest under Mount Doom. What matters--what drives consumption--is biases and expectations about procedures. Screening is seen as cheap, preventive, population-based, thoughtful, and solace-giving, so it must be good. When governments in Canada cut back severely on health care spending in the 90s, they drastically cut back on the amount of time people could stay in a hospital bed after elective surgery. This was decried as inhumane, with lots of anecdotes about patients going home with fresh sutures, etc. Yet today no one has an issue with it, and post-surgical complications in outpatients are not a major cause of death, emerg visits, or medical error.

I'm no expert on Romneycare, but my guess would be that he tackled the financing but was unwilling to tackle popular mythology.

Dan Weber writes:

the suggesting that we spend twice as much as other countries because of cancer screening is absurd

It's not cancer screenings; it's our general tendency to spend money on things that have minimal, no, or negative effects on outcomes.

Cancer screenings are just an excellent example of the problem. The random person on the street thinks they must reduce costs, because they're prevention, and prevention saves money. Right? Right?

Tyler writes:
Cancer screenings are just an excellent example of the problem. The random person on the street thinks they must reduce costs, because they're prevention, and prevention saves money. Right? Right?


I would agree that our inability to efficiently allocate where our healthcare dollars are spent, both in the public and private programs, is the primary reason our costs are out of control. If his argument was merely that our healthcare system inefficiently allocated healthcare, I would not have a problem. However, using this as a defense of privatization when this problem is not solely in the public sphere is bogus. Preventative healthcare in general can increase life expectancy (obesity, high-blood pressure, etc.) and, in general, would be under supplied by a privatized healthcare system (whether overspending in a public one is any better is certainly up for debate).

Kyle writes:

Arnold is making an argument I make to people all the time - how you finance health care delivery (single gov't (UK), multi gov't (Canada), private-ish (US) etc.) may be less important than whether the system is intelligently designed and whether the population is making intelligent health choices.

There's not much reason to think changing who pays for healthcare is going to make the populace suddenly less obese or less obsessed with end of life care. It will, however, make it less expensive for most of the population, since most taxes are paid by a small percentage of the total. A lot of the obsession with single-payer here in the US is less about making health care smarter or better and more about getting someone else to pay for their stuff.

Comments for this entry have been closed
Return to top