Arnold Kling  

The CBO Offers Hansonian Charts

Why Are Bartenders So Rude?... Fundamental Causes...

Ezra Klein highlights them. They show little difference in outcomes between genuine treatments and placebos.

The implication of course is that it's far cheaper to give someone a chest incision than an angina, and far cheaper to give them a sugar pill than an anti-depressant. There are ethical issues, of course. But still. Interesting.

Klein's post was highlighted by Tyler Cowen.

Comments and Sharing

COMMENTS (6 to date)
ajb writes:

I'm not sure it says placebos work. Rather it seems to show that usual treatments don't matter for the majority of people treated. This is not the same thing.

Simple Example: Assume that most people are unaffected by treatments for a condition (80% will improve on their own), but that with treatment, another quarter of the remainder get better. So total improvement 85%. With placebos less than 5% of remainder will improve. So total with placebos is under 81%. The total difference seems small, but conditional on being in the category that won't improve on their own, people clearly benefit from treatment.

The issue then becomes cost-benefit. How costly is the drug/treatment and can we improve who gets the drug?

In fact, given the right loss function, it may well pay to give a very cheap drug to all if a) it has no side effects and b) it clearly helps even 1% of the people.

RL writes:

Just a brief correction from an MD. As phrased, "angina" in the sentence makes no sense. I suspect it should read "angioplasty". Angina is a symptom (chest pain), while angioplasty is a treatment.

Ajay writes:

RL, the charts from the link correctly address this point but Ezra, despite making many and varied statements about health-care policy, probably doesn't really know much about medicine, or its funding for that matter, and is sloppy enough to simply grab words from charts like that without knowing or caring what they mean. Such is the sad state of health-care policy debate when such mediocrities as Ezra are allowed to take part.

MouseJunior writes:

Ezra might have a stronger case if he wasn't using a chart from a 1959 study that resulted in the treatment being used in the study being dropped because it was ineffective.

"This treatment no one uses is no more effective than a placebo, therefore it's interesting to think about prescribing placebos" kind of fails as an argument.

Dan Weber writes:

Looking at the first chart, if I was depressed, I'd rather have the pill that has an effectiveness of 12 rather than an effectiveness of 8.

Although I might not want to spend $200 a month for the former as opposed to $4 for the latter.

Of course, given that most of the drugs in that chart are now off-patent, the additional cost doesn't matter at all.

PS: link to Tyler is just another link to the article.

George writes:

Arnold, I have to agree with Ajay: why are you linking to Ezra Klein? Don't people usually cite others who know more than they do about a subject?

Or did you just want us to head over there and laugh at him (or perhaps "give him an angina")?

Comments for this entry have been closed
Return to top