The problem with being too sensitive or too specific (created 2009-09-16)

This page is moving to a new website.

Somebody asked my opinion about cost effectiveness research. My bottom line is that I like it, but I understand why it is controversial. Here's the logic that I presented to draw that conclusion.

I've decided that all of life's controversies can be related to sensitivity and specificity. Bear with me here.

You can define sensitivity in terms of a formal diagnostic test, but I prefer to think of it in more general terms: sensitivity is one minus the false negative rate and specificity is one minus the false positive rate.

The problem with most of life's controversies is that half the people "too sensitive". They are arguing about how awful all the false negatives are and ignore the false positives completely.

The other half, of course, are "too specific". They argue about all the false positives and ignore the false negatives completely.

An example of the "too sensitive" group is those folks who advocate greater cancer screening with anecdotes of "I didn't get screened and now I'm going to die unnecessarily" (false negative) or the flip side of the same coin: "I got the screening test and it saved my life." Now I don't want to be too harsh, as greater cancer screening is often a good thing, but where is the balance of discussion of the unneeded biopsies, the unnecessary anxiety, etc. when a false positive result occurs?

This applies to other areas as well. There are people who argue that the U.S. Food and Drug Administration (FDA) has been making it too difficult and too expensive for drugs to get tested and approved. These are the "too sensitive" people, who worry about false negatives (good drugs that are kept off the market). Others argue that the FDA is too beholden to industry and needs to tighten up their rules. These are the "too specific" people who worry about false positives (useless drugs getting approval).

Tort reform falls in the same category. The critics of the U.S. legal system for malpractice claims are "too specific" because they worry about the false positive cases like the woman who spilled coffee in her lap while driving and sued McDonald's. Yet any effort to make it more difficult to sue and prevail in our court system is almost certainly going to increase the false negative rate, those people who were truly harmed by the negligence of an individual or corporation and are unable to seek adequate compensation. Of course, the "too sensitive" groups will highlight all those evil corporations and look at how they are getting off without having to pay a serious penalty for their wicked deeds (false negatives).

A false positive in healthcare is a patient that is needlessly treated. A false negative in healthcare is a patient that misses out on a helpful treatment. It seems to me that cost effectiveness research is a careful attempt to balance the costs of false positives and false negatives and their respective probabilities.

The big problem, though, is that the parties that bear the bulk of the cost of a false positive are usually the insurance companies and the parties that bear the bulk of the cost of a false negative are the individual patients. Although I do realize that higher insurance costs will eventually back to me, I tend to discount that if I am an individual patient with the prospect of a serious disease staring me in the face.

That's why there are no easy answers. Consider prophylactic treatment with antibiotics. If you treat a patient with a viral infection unnecessarily with antibiotics, who pays the cost? The patient does to some extent because of the exposure to unnecessary side effects of the antibiotics, but the greater cost is to society in general as antiobiotic resistant organisms gain a more solid foothold in our community. Society doesn't pay as much of a price if a bacterial infection is left untreated--I'm the one who was to miss work and suffer the misery of an infection that could have been cleared up a whole lot sooner with a few pills. I suppose society suffers in some sense that I won't be able to compose long webpages like this when I'm sick, but somehow I think that society will survive with one less webpage about Statistics.

Should we close a school when a report of an H1N1 infection appears? There's a cost to society when H1N1 spreads quickly, but there is a cost to the individual parents when they have to miss work or pay a babysitter for their kids because they don't have a school to go to during the work week.

In the broad sense, society suffers when an individual suffers and an individual suffers when society suffers. But it's hard to take the broad perspective when it's your body on the line.

So I'm all in favor of cost-effectiveness research, partly because the research demands the use of lots of complicated statistics. Seriously, cost effectiveness research avoid the trap of the "false positive" anecdotes dueling with the "false negative" anecdotes, and it tries to avoid being "too sensitive" or "too specific". But as long as there is a disparity in who pays which bills, I can understand why cost effectiveness research is controversial.