In the first episode of the television show The Resident, a nurse tells the young protagonist that medical error is the third leading cause of death in the United States after cancer and heart disease. “They don’t want us talking about that,” she adds.
This shocking and unforgettable line did not begin life with The Resident. Since 2016, it has “earwormed” its way into the public discourse. A recent email I received linked to this myth and asked me to have a look at it before blindly trusting the “official narrative” in medicine. The implication was that medicine kills and I should be more open-minded to the alternatives.
Is medical error really the third leading cause of death in the United States? Investigating a claim like this invites accusations of insensitivity, so allow me to state a few important things. Medical errors are real. Some people have died or been permanently injured because of errors fostered by a healthcare system that needs to be improved. Errors in medicine include wrong diagnoses, drug dosage miscalculations, and treatment delays. These errors are likely to be underestimated because studies tend to focus exclusively on hospitals and not on the rest of the healthcare system; because some errors may only have debilitating effects years down the road for a patient and are thus harder to trace; and because reporting these errors may not be encouraged by the medical culture. The patient safety movement is important because errors that can be prevented should be prevented. I have personally been on the receiving end of a minor medical error, in which a clear laboratory report was misread by my doctor and, had my condition deteriorated, I presumably would not have been given antibiotics because my doctor thought the report said my infection was viral in nature. I have, in this small way, experienced part of this problem and am sensitive to it.
But as has been written on the topic, “there are no useful fictions in medicine.” The idea that medical error is the third leading cause of death in the U.S. is indeed a fiction, an overestimation that has negative consequences.
Turning apples into oranges
This whole story has its prelude in a 2000 report called To Err Is Human: Building a Safer Health System by the Institute of Medicine. The report took two studies, one done in Colorado and Utah and the other in New York, and extrapolated their results to all hospital admissions in the United States, concluding that between 44,000 and 98,000 Americans must be dying each year as a result of medical errors. The lower estimate exceeded the eighth leading cause of death and trumped fatalities from motor vehicle accidents.
In 2016, the British Medical Journal (BMJ) published an “analysis” by a research fellow, Michael Daniel, and a professor who had developed the operating room checklist, Martin A. Makary, both from the Department of Surgery at Johns Hopkins University. To call it a study would be inaccurate. It was a call for better reporting of medical errors, motivated by a lack of funding available to support quality and safety research and propped up by a back-of-the-envelope calculation. The authors looked at the few studies that had been published on the problem since the Institute of Medicine report. They took the mean death rate from medical error from those studies and extrapolated them to the total number of U.S. hospital admissions in 2013. After adding that this extrapolation was surely an underestimation of the actual problem, they concluded that this would mean medical error would rank third in the Centers for Disease Control’s list of causes of death in the U.S. This became the title of their published analysis, which has been cited in at least 1,265 papers according to Scopus, and this memorable idea spread to news articles, television shows, and alternative medicine circles.
Critics of this analysis have pointed out many flaws. It is based on studies whose data was never meant to be generalized to the entire U.S. hospitalized population. For example, one of these studies, by the Office of the Inspector General of the U.S. Department of Health and Human Services, was conducted in beneficiaries of Medicare, who are aged 65 or older, have disabilities or have end-stage renal disease which requires dialysis or transplant. The study authors counted the number of deaths in their sample to which they believed medical errors had contributed, and this number was then used in the BMJ analysis to extrapolate to all U.S. hospitalizations. However, this makes the mistake of extrapolating an observation found in one sample to a different type of population. Case in point: if we look at everyone hospitalized in the United States, one patient out of ten is there to deliver a baby. Taking death statistics from a sample of Medicare patients and extrapolating it to all hospitalized patients is like turning apples into oranges, to adapt a popular saying to the current situation.
Moreover, the studies whose results were averaged for the BMJ analysis were never about uncovering preventable deaths; rather, their objective was to round up numbers on harm from medical care. Harm can lead to death, but this causal link needs to be properly evaluated, and it wasn’t in those studies. Dr. Kaveh G. Shojania and Pr. Mary Dixon-Woods, who wrote a sharp commentary of the BMJ back-of-the-envelope calculation, give an example of how easy it can be to mistakenly draw the causation arrow from medical error to death. Imagine a patient who enters the intensive care unit with multi-system organ failure due to their body’s extreme response to an infection. Doctors mistakenly give the patient an antibiotic to which they have had an allergic reaction in the past, and the patient develops a rash from the antibiotic. The antibiotic is changed, but a week later, the patient dies as their organs stop working. Yes, the authors argue, a medical error was committed, but it probably did not cause the patient’s death. Using studies that identify medical errors that were followed by death to declare that these medical errors necessarily caused these deaths is not fair. What these studies do not take into account is how long these patients would have lived had they received optimal medical care. Since it is not considered, it can skew the impact of medical errors.
Another problem arises when we look at how many deaths were reported in the studies combined into the BMJ analysis. The Office of the Inspector General study mentioned above reported 12 deaths associated with medical errors. Two more studies used in the analysis listed nine and 14 deaths. The remaining one claimed nearly 400,000 deaths. Generalizing from so few deaths (with the exception of this last study) to all U.S. hospitalizations, as Shojania and Dixon-Woods put it, “surely warrants substantial skepticism.”
What we end up with, when we look beyond the scary headline of medical errors as the third leading cause of death, is an analysis of studies that were never meant to look at deaths caused by medical errors, often reporting a very small number of deaths from populations that are not generalizable to the whole of the United States, and being combined in a crude way. The BMJ’s higher estimate of preventable deaths due to medical error—440,000 patients a year—translates to 62% of all hospital deaths, as was pointed out by Drs. Benjamin L. Mazer and Chadi Nabhan. That nearly two thirds of all deaths occurring in hospitals would be due to medical error strains credulity. Indeed, more recent studies have looked at the phenomenon and the numbers that have emerged are a far cry from 62%. A study from the UK reports that 3.6% of hospital deaths were due to preventable medical error; a similar study out of Norway reports 4.2%; and a meta-analysis of the problem published in the BMJ in 2019 concludes that at least one in 20 patients are affected by preventable patient harm, with 12% of this group suffering from permanent disability or dying because of this harm.
The authors of this recent meta-analysis are quick to point out that the numbers reported by the studies they looked at vary considerably. It is not easy to determine if a particular case of patient harm was preventable or not. In fact, a study that specifically tested for this reported that the doctors who look at medical files to make this assessment often disagree. In their study, if one reviewer decided that a death in hospital was definitely or probably preventable, there was only a 16% chance that a second reviewer would agree with them, and there was a nearly identical chance that a second reviewer would clearly disagree. This problem of medical errors is like an iceberg. Everyone can agree on its visible tip, but when we try to assess the much larger size of the phenomenon by squinting through the waters, disagreements abound. The “third leading cause of death” then becomes a useful shorthand, an urgent rallying cry we are not supposed to question because the preventable harm is real and desperately needs to be addressed. But relying on this crude overestimation is not harmless.
Jumbo jets and magic carpets
The consequences of exaggerating the scope of this very real problem should not be dismissed. In 2019, a video released by the National Rifle Association used this myth to claim that medical malpractice was deadlier than guns, specifically that deaths from medical errors were 500 times higher than deaths from accidental gun incidents. Sure, it’s a simple bit of whataboutism, but it provides ammunition to irresponsible gun owners, allowing them to casually deflect criticism. More worryingly, the claim has been weaponized by believers in alternative medicine to paint conventional medicine as dangerous—practically the equivalent of playing Russian roulette—while touting the alleged safety of their favourite pseudomedical practices. Indeed, if you constantly read that “more Americans are killed in U.S. hospitals every six months than died in the entire Vietnam War,” that medical errors kill the equivalent of “three fully loaded jumbo jets crashing every other day,” and that these errors and injuries are “epidemics” borne of a “cult of denial and complacency,” as popular medical papers and reports tell us, you may wonder if homeopathy would be a more reasonable alternative.
Not only are these scary comparisons derived from dodgy numbers, as demonstrated earlier, but to compare the harms of medicine to the harms of alternative medicine without looking at their respective benefits isn’t fair. The health benefits of acupuncture, homeopathy, chiropractic and herbalism are few and far in between. (For an in-depth review of the evidence, I would strongly recommend Simon Singh and Edzard Ernst’s book, Trick or Treatment? Alternative medicine on trial.) Meanwhile, medicine is about balancing risks and benefits. It’s an imperfect system, one that requires active campaigning for improvements, but as the saying goes, problems in aircraft design should not encourage us to see if carpets can fly.
It has been said, with regards to medical errors, that you can’t manage what you can’t measure. But using incredible numbers borne out of unreliable calculations cannot be the solution.
Take-home message:
-A popular claim that medical error is the third leading cause of death in the United States originated in a 2016 back-of-the-envelope analysis published in the British Medical Journal
-This ranking is an exaggeration that was arrived at by combining a small number of studies done in populations that were not meant to be representative of the entire U.S. population and that were not designed to prove a link between a medical error and death
-The claim is often used by proponents of alternative medicine to scare people away from medical care.
Leave a comment!