Thinking About Medical Errors

Press releases following a recent publication in the British Medical Journal (BMJ) hysterically echoed the article’s headline: “Medical error—the third leading cause of death in the U.S.”1 The authors used a variety of published sources on the incidence, lethality and preventability of medical errors to produce an estimate of 251,000 deaths per year attributable to medical error, out of a total of about 2.6 million. As a cause of death this would rank behind only heart disease (611,000) and cancer (585,000). While the purpose of the authors was to advocate for improved coding of the cause of death in vital statistics, the purpose of the commentary was to alarm the public regarding the current state of healthcare. Should we panic?

I think not. Here’s why:

First, understand that I’m not a criminal or even an apologist. I hate medical errors and I have devoted my career to their eradication. We should strive every day to make healthcare as predictable and safe as humanly possible, and we should embed continuous quality improvement in everything we do. So looking at errors is important. But we have to recognize their place in any complicated human system.

It’s possible that errors are increasing because of increased reporting. This would be a good thing—we can’t fix what we don’t measure, to paraphrase Deming—so maybe the increase reflects increased recognition of something that’s been going on all along.

It’s possible that errors are increasing because we are providing more complex care, with more opportunities to make mistakes. This is an unintended consequence of medical practice, and can be viewed as a cost of doing business, with the increased rate of errors balanced by the benefits of the new procedure. We can expect the rate of unintended consequences to go up during our learning curve and then back down as we figure things out. Corneal abrasions associated with robotic prostatectomy are a good example of this phenomenon, and one that has been seen to wax and wane as these operations have gone from long and experimental to short and routine. Since robotic prostatectomies are ultimately beneficial, the increase in errors is also a good thing. It shows we are appropriately pushing the envelope.

Finally, it is possible that errors are increasing because our definition of what constitutes an error is continuing to evolve. Failure to treat h. pylori in a patient with a stomach ulcer today would be considered a significant error— whereas 30 years ago this therapy was not even considered. Ten years ago all of our patients were nothing by mouth (NPO) for eight hours, whereas today we are handing them oral electrolytes in the preoperative holding area. Today the failure to swab the IV port before injecting a medication is a mild deviation; a year or two from now it will be considered a major error.

Second, it is important to recognize that every patient we care for is going to die. Benjamin Franklin noted: “In this world nothing can be said to be certain, except death and taxes.” Jimi Hendrix put it another way: “No one here gets out alive.” Or if you’re a fan of Game of Thrones, try “Valar Morghulis” (“all men must die”). As those responsible for the battle against the inevitable, physicians have long studied the causes of death and attempted to categorize them. This activity, in turn, has led to scientific and governmental attention to the leading causes, which in turn has led to significant improvements in care. And we continue to push this envelope today: just ask any provider how the coronary artery bypass graft (CABG) patient of today compares to the CABG patient of 20 years ago. Today’s patient is vastly sicker, because all the easy patients either never get coronary disease in the first place (statins) or get fixed in the cath lab (angioplasties). So the patient that makes it through to the OR is older and sicker. Similar advances have occurred in cancer, AIDS, COPD and other leading causes of death, not to mention that many scourges of the developing world—infectious diseases—have been eradicated in the U.S.

So what do our patients die from? Getting old, mostly. As we eliminate preventable causes, our population gets older and frailer, balanced on the edge of a progressively narrower knife blade until reaching a point where the slightest gust of wind will blow them to their demise. Many of these breezes, of course, can be associated with a medical error of some kind.

Which brings me to the third and final point, that medical errors are ubiquitous in medical care. I’ve delivered tens of thousands of anesthetics in my career and I can honestly say that I have never done a perfect case. In retrospect there is always something I would have done differently—one mg less of some medication, increasing or decreasing the volatile agent one minute sooner, giving just a little more or little less fluid, etc. Healthcare is complicated, and the odds of delivering the perfect anesthetic are far, far lower than the odds of filling out a perfect NCAA bracket. Don’t believe me? Open the medical record of the next inpatient you take care of and look to see if they’ve gotten every prescribed medication at the prescribed time. Invariably there will be both omissions and delays. Depending on how you’re keeping score, every one of these events would count as a medical error. Errors occur in the care of every patient!

Now put these thoughts together: death is inevitable, medical error is inevitable, and thus death due to medical error is an inescapable conclusion. When we have eliminated every named disease, accidental death will be all we have left. So maybe what the BMJ article is really documenting is an improvement, not a cause for alarm. My glass is half full!