The rarity of autopsies in the US

[nancylebov]

Today, hospitals perform autopsies on only about 5 percent of patients who die, down from roughly 50 percent in the 1960s.

I'd heard something about the decline of autopsies in the context of fat acceptance-- we don't actually know as much as is commonly thought about the risks of fatness because we don't know what fat people actually die of. However, I had no idea that the autopsy rate had gone that low.

I don't know how often the cause of death is so obvious that autopsy might not make sense, but even then, I've heard that one of the fascinating things you learn as a pathologist is how much can be wrong with people that doesn't kill them. Having a large statistical base of such information would be very useful in determining what conditions which have few or no symptoms should be treated.

More details, if you like. In other words, it's a long somewhat technical article with a bit I find supports my thesis (that 44.4% of autopsies turned up information which would affected treatment), but which I'm not going to read in any detail. Unfortunately, even though it mentions that autopsy rates are declining world-wide, it doesn't have a handy chart of % by country.

First link found at Overcoming Bias.

Comments

[st-rev.livejournal.com]

8.8% of the autopsies showed the wrong treatment was used, that's 44.4% of the 19.8% of the cases where the premortem diagnosis was incorrect.

The sample base--91 autopsies actually performed--is pretty small.

[osewalrus.livejournal.com]

This appears to be a case where the obvious culprit is a desire to cut cost. The question is, "why do an autopsy?" The next question is "based on the perceived benefit of an autopsy, who pays?"

In criminal forensic cases, it makes sense because you need evidence. Therefore the state pays. If malpractice is suspected, the patient's family has incentive and will pay -- or the hospital may have incentive for defensive purposes and will pay. But if the goal is merely to improve the system overall? No one pays for that, because the benefit is too diffuse. Classic collective action problem.

Historically, a role played by government is to help overcome the collective action problem. But modern political theory appears to have abolished the notion of "collective action problem" along with other real world considerations such as information asymmetry.

[madfilkentist]

Government is very often the locus of the collective action problem, in the sense that you use the term. A special interest group can lobby for favors for itself and get them, and the opposition, though potentially widespread, is less committed because the benefit of fighting is diffuse.

I wonder to what extent the number of autopsies has gone down because there are more non-invasive tools to determine the cause of death. Don't know if this is a significant reason; I'm just wondering.

[osewalrus.livejournal.com]

"It's a poor blaster that won't point both ways." -- Salvore Hardin

The benefit of overcoming collection action is balanced by numerous risks. You quite correctly point out the standard political action problem that most commonly manifests itself in the presence of government authority. The question, in a general sense, is how do you like your odds and what cost are you willing to bear.

For the specific of autopsies, there is still an unanswered question of "what is the benefit of an autopsy." Or, more generally, why does anyone care about the cause of death? There may be reasons, and it may be -- as you suggest -- that less invasive tools permit whoever gets the benefit to capture sufficient benefit. But it is impossible to determine without knowing what the benefits of autopsies are.

[nancylebov.livejournal.com]

However, we don't know why autopsies were that common. Were they legally required or was it a matter of hospital accreditation, or what?

From what Thette says about the decline of autopsies in Sweden, I wonder whether a consensus has built up that we already know enough about causes of death.

[thette.livejournal.com]

The Swedish numbers are:
*A peak at about 50% in the mid-seventies
*33% 1988
*12% 1998
*7% 2008

The trend is the same, despite different sources of funding and different organisations.

[hummingwolf.livejournal.com]

When my father died, my family asked that his body be donated to the state anatomy board for use in educational autopsies (none of his organs were really worth anything by that point, so we thought this would be the best use of the corpse that would respect my father's wishes). One doctor thanked us profusely and said they wished more people would do that.