Medical Nemesis

You may have heard about this analysis in the BMJ that says medical error is the third leading cause of death in the U.S., after heart disease and cancer. This isn't actually supposed to be news, what the authors are really saying is that we need to keep track of it better and deal with it better.

A bit more about that anon, but let me also say that it isn't exactly, precisely true. Although they give as an example a case in which a botched medical procedure affirmatively caused a death, some of the cases that counted are more examples of medical error failing to prevent death from disease or an unrecognized lesion. It's sort of like saying that a faulty guardrail, rather than a driver who falls asleep at the wheel, is a cause of death. In other words we have an expectation that effective medical intervention will be provided when it is possible, and we're willing to call failure to prevent death a cause of death.

On the other hand, medical intervention can cause death when there is not any error per se, or at least it's not clear that there is. There is a certain risk associated with surgery or pharmaceutical treatment, which has to be weighed against the likely benefits. At what point less than perfect judgment or execution gets defined as an error is not usually clear, either. We don't know the details of Prince's case yet, but soon we might be asking, "Should his doctor have written that prescription for opioids after his surgery?" At what point did continuing to give him percocet become an error?

So what Makary and Daniel are saying is not that we should blame health care providers and that those incompetent, careless doctors are killing us all. What they are saying is precisely the opposite: that we need to create a culture that doesn't blame, but fixes problems. Medical professionals are human. They will make mistakes. The difference is that in most professions, mistakes are seldom or never fatal.

So they propose, first of all, that death certificates include a notification that medical error was a factor in the death. I'm not sure how well that would work -- the physician responsible would likely be the one filling out the certificate, there are liability issues, and as I say, how do we really know when what happened qualified as an "error"? This seems unrealistic to me.

On the other hand, we could have a culture and a practice in which avoidable death and injury are discussed openly within the provider institution, and procedures are put in place or reinforced to prevent recurrences. In order to have continuous quality improvement, you need to limit blame and negative consequences of mistakes so that you don't just drive them underground. But on the third hand, some individuals should not be practicing, or should be required to take time out to be retrained or solve personal problems such as addiction that are interfering with their performance. You can't make quality improvement entirely blame free.

So this is really about balance. It isn't about a dichotomy, and it isn't simple.