Just what level of certainty should we be going for? I have written before about the quest for diagnostic certainty, and about how we – and by “we,” I mean both doctors and patients – should be asking ourselves whether it makes sense to devote a lot of time, effort, and money to additional testing to raise the level from 93% to 97%.

But remember, if it’s 93%, that means being wrong roughly one time out of fourteen. I’m OK with being wrong one time out of fourteen for some things. For example, if I think you have strep throat and want you to take penicillin, and I think it’s unnecessary to waste time and money on a confirmatory test, what happens if I’m wrong? The penicillin is very unlikely to cause any harm. Serious reactions are quite rare, and I don’t worry much about increasing bacterial resistance to it, because most bacteria are already resistant to it. The strep that causes throat infections just happens to be an especially dumb bug, so penicillin still works.

But I don’t think it’s OK to be wrong one time out of 14 about whether you have something that could cause severe, permanent disability or sudden death. Neither do most of my colleagues or most of my patients. I guess some people are OK with that level of risk, but then some people jump out of airplanes, too. Yes, I realize that the frequency of the parachutes not opening is far less than one out of 14, but when it happens, the results are … well, you know.

So we tend to be very cautious with ED chest pain patients. We have all sorts of “tools” at our disposal to try to figure out what’s wrong with them, and in particular whether it’s a heart problem. This is an area of intense clinical research: what is the best strategy for assuring that, if we send somebody home from the ED, the likelihood of something bad happening is vanishingly small? Or at least down in the 1% range, because no strategy that involves human beings is going to be right 100% of the time.

We have a standardized name for the “something bad happening,” too: MACE, which stands for Major Adverse Cardiovascular Events. Our medical journals have many published papers on strategies for evaluating ED chest pain patients to assure that the likelihood of MACE in the next 30 days (the most commonly used time period) is as low as possible. Oh, and just in case you are worried about day 31, which I am, the better studies keep track of the patients out to a year.