Today's Washington Post headline warns:

Researchers: Medical errors now third leading cause of death in United States

The article goes on to say that researchers at Johns Hopkins estimate that 251,454 deaths per year are caused by medical errors, following death from heart disease and cancer, and exceeding death from respiratory disease, accidents, etc.

They go on to quote one of the researchers:

"We all know how common it is," [Martin Makary, a professor of surgery at the Johns Hopkins University School] said. "We also know how infrequently it’s openly discussed."

As a layman, I might well believe this article. But I'm not a layman. I'm an insider, until recently a busy Emergency Department physician, who worked with a lot of other doctors simultaneously and over long periods of time, who knows a lot about doctors in the community, who saw a lot of different patients and a lot of outcomes. I'm not blind to errors (they are very powerful teachers) - we all make them, and many of us do talk about them in hushed, confidential tones. We all know that sometimes medications are ordered in the wrong dosages, or meds are given to the wrong patient, that critical diagnoses are sometimes missed, and even much worse*. But with all the patients I've seen and cared for - and I have seen a lot of death both in the ED and on the 'floors' (other parts of the hospital) - they are not dropping like flies around me because of bad medical care, at least not that I can see.

Is it likely that these numbers are accurate, or more likely that minor errors are weighed in a manner which exaggerates the results? Or, is this article and its claims true?

I did see this question, the answer to which uses some of the same numbers, and of which my question might be considered a duplicate.

*I saw my first death from medical error as a fourth-year medical student during an ED rotation in the busiest ED in the country at the time. Someone had fallen from a building and arrived unconscious. Even as a medical student, I had heard the rule, never place an ng (nasogastric) tube in a patient with a severe head injury until a basilar skull fracture had been ruled out. One of the residents had placed an ng tube in the patient, and I will never forget the xray of the entire ng tube coiled in the patient's skull. As I said, mistakes are very powerful teachers, even the mistakes of others, and are not easily forgotten.

  • 5
    Makary and Daniel's paper was published in a peer-reviewed journal, the BMJ: bmj.com/content/353/bmj.i2139. Typically on this site, we take peer-reviewed publications as our "gold standard". This is not to say they are infallible, but it means their claims have been evaluated by people much more qualified in their fields than the typical user of this site. Adding our own opinions of such a paper wouldn't add much. So I don't think an answer on this site could really go much beyond "you have correctly stated the claims made in the paper". May 4, 2016 at 12:30
  • 6
    It's worth adding that, as medical science gets better at fixing the common causes of death, one would expect death due to error to rise, since in general it's difficult to reduce error below a certain point.
    – Ethan
    May 4, 2016 at 18:01
  • 4
    Something to keep in mind: Many of these medical "errors" mean nothing in the big picture--they're the guy died today instead of next week type things. May 4, 2016 at 19:48
  • 4
    If you agree with all the stats, but disagree with the conclusion perhaps you could pinpoint the point in their argument that you think is unjustified. Then we could maybe look for evidence to check whether that specific step they take is supported or not.
    – user30557
    May 4, 2016 at 20:42
  • 4
    @NateEldredge, we may take peer-reviewed journals as the gold standard, but there are often nuances that get lost when the papers are summarized in popular media. For example, "cause of death" has varying definitions: a coroner might classify the cause of a death as "blood loss", while the popular media would describe it as "he was shot a dozen times".
    – Mark
    May 5, 2016 at 0:41

2 Answers 2


The paper the newspaper cites is in a peer reviewed journal and extracts its information from other peer reviewed journals. On that basis I have no reason to doubt the raw data they cite. There is, however, a strong possibility that your disbelief stems from a category error.

The paper defines medical error as:

Medical error has been defined as an unintended act (either of omission or commission) or one that does not achieve its intended outcome, the failure of a planned action to be completed as intended (an error of execution), the use of a wrong plan to achieve an aim (an error of planning), or a deviation from the process of care that may or may not cause harm to the patient. Patient harm from medical error can occur at the individual or system level.

The anecdote you give is an example of "a deviation from the process of care" - the accepted process of care being do not stick a NG tube into someone's brain. While these are the most obvious and graphic sources of "medical error", the definition used in the paper is much wider than that.

Without being able to access the source documents I cannot say exactly what makes up the "medical errors" being included in the statistics. From the description it would include:

  • misdiagnosis
  • late diagnosis
  • incorrect medication
  • incorrect dosage
  • incorrect timing of medication
  • delays in treatment
  • errors in triage
  • ordering of inappropriate tests
  • not ordering appropriate tests
  • etc.

Did you consider all of these in your idea of "medical error"?

As an aside, if the data is correct, then many of the causes of death currently classified as heart disease or cancer should be reclassified as medical error (if for example the incorrect dosage of a drug caused a cancer patient to die six months earlier then they otherwise would) making those figures smaller.

  • 8
    "Medical error has been defined as an ...act... that does not achieve its intended outcome" So if the only treatment for a fatal condition has a 10% chance of succeeding, is using it a medical error 90% of the time?
    – DJohnM
    May 5, 2016 at 4:09
  • @DJohnM That sort of thing could be an error under "one that does not achieve its intended outcome" - the fact is without seeing the original papers (which are behind a paywall) we don't know if that is or is not counted as an error.
    – Dale M
    May 5, 2016 at 5:20
  • 3
    Do you know if this definition of medical error includes a patient not getting appropriate or standard treatment because they can't afford it, or because it's in some way excluded in their insurance plan? May 5, 2016 at 9:21
  • Yes, they are medical errors, with a few caveats. E.g., delay in diagnosis because the MD failed to consider the correct diagnosis in the differential is a medical error. If the insurance didn't pay for a diagnostic test resulting in a delay, that is a health care system failure, not a medical error. How ordering of inappropriate tests is an error leading to death is also subject to some scrutiny. Unnecessary tests can cause harm and cost lives. But the vast majority of unnecessary tests do not result in physical harm (not that I approve of them), or we really would all be dropping like flies. May 5, 2016 at 14:16
  • Regarding your final comment, I do agree that any error leading to death - even a fairly imminent death - is an error. However, you seem to imply that most heart disease and cancer should be curable, therefore should be added to the numbers of deaths due to medical error. Please believe me that doctors can't cure many things; often the best we can do is try hard to stave off the inevitable. If it's not our best effort by definition (i.e. deviates from accepted/established standard), it is an error. May 5, 2016 at 14:27

As an medical insider from the ICU nursing perspective. I will agree that doctors do their best to prevent medical errors. The point of "To Err is Human" is that our best isn't good enough. Humans at their 'best' have a 3% error rate and the IOM report of 1999 is responsible for making hospitals accept that the problem exists and resulted in changes to improve 'error trapping' such as two identifiers before providing a therapy, having coworkers sign your med sheets when giving high risk medications and instituting 'time out' before beginning invasive procedures. https://www.ncbi.nlm.nih.gov/books/NBK2673/

What the ER doctor failed to mention is this study: two recent studies of malpractice claims revealed that diagnosis errors far outnumber medication errors as a cause of claims lodged (26 percent versus 12 percent in one study) https://www.ncbi.nlm.nih.gov/books/NBK20492/

and these findings were reiterated by a separate study that concluded clinical diagnosis is verified with autopsy in 49.1% of the cases. Doctors are worried about the assigning of blame for obvious reasons, fear of litigation. He isn't aware of patients 'dropping like flies' simply because lots of patients die in hospitals and fewer than 10% ever receive autopsies. http://emedicine.medscape.com/article/1705948-overview


Failure to recognize a problem exists is very common in the medical community. One paper attempted to minimize the Johns Hopkins open letter by suggesting the IOM was the actual standard. Any attempting to inflate these numbers were merely alternative medicine quacks seeking to make a quick buck by frightening the public about the dangers of medical care. https://sciencebasedmedicine.org/are-medical-errors-really-the-third-most-common-cause-of-death-in-the-u-s/

Very sad indeed, that anyone attempting to improve patient safety is labeled as a crazed quack. The medical system needlessly kills people. This is why we have medical attorneys. And before the reader suggests most law suits are frivolous and are done solely for profit, consider the Harvard malpractice study that found 47 malpractice claims occurred among 30,195 patients after initial hospital visits; of 280 patients identified with adverse events caused by medical negligence (defined per study protocol), only eight filed claims (<2%). That means fewer than 2% of all injured by the medical system actually file suit and less than 1% actually receive awards (that are generally capped at $250,000, darn few get millions) http://www.nejm.org/doi/full/10.1056/NEJM199107253250405

The numbers do not reflect deaths from adverse drug reactions that the FDA concedes results in a considerable number of annual deaths: http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Surveillance/AdverseDrugEffects/ucm070461.htm

The FDA also concedes that physicians grossly under report these serious and potentially fatal drug interactions. The median under-reporting rate was ...high at 85%. This systematic review provides evidence of significant and widespread under-reporting of ADRs to spontaneous reporting systems including serious or severe ADRs. https://www.ncbi.nlm.nih.gov/pubmed/16689555

Add to these the unconscionable number of amenable deaths (those due to untimely access to medical care but otherwise preventable). One study concluded: In 2004–2005, amenable mortality in the United States ranged from a low of 63.9 deaths per 100,000 people under age 75 in Minnesota to highs of 142.0 in Mississippi and 158.3 in the District of Columbia. I'll spare you the math, the average is around 1 out of 1000 deaths in the US could be prevented if they received appropriate care in a timely fashion.

I am a former Traveling ICU nurse and worked in 26 different hospitals nationwide. I am also a former Molecular Biologist who worked for a major drug company. I am a former Koolade drinker but no more. I have nothing to sell except to tell the reader to do as the Public Health Service suggests and choose your parents wisely.

  • 1
    I don't usually tell people (especially doctors) this, but I have done a fair amount of medical malpractice work, and started as a plaintiff's expert, because someone has to stand up for those injured by bad medical practices. I ended up doing a lot of defendant work as well, because of the balance that I saw was needed. I think we can definitely cut down on the number of medical errors we commit. I just don't believe - from my decades of experience - that medical error is the third leading cause of death. Feb 1, 2017 at 17:36
  • depends on how you define 'error'. In my world they include any potentially avoidable event that cause harm in anyway (including requiring additional hospital stays or testing that incurs financial loss). My list includes, med errors (preventable), adverse drug reactions (unpreventable side effects). iatrogensis, nosocomial infections, and amenable deaths resulting from untimely access to reasonable standard of care. And these only include hospital related deaths, add to these deaths in nursing homes, out patient clinics and doctors offices and you see my concern. I am provider of this care. Feb 1, 2017 at 17:59
  • Take fatal adverse drug reactions. They are not always medical errors. If a doc prescribes ACE inhibitors to someone who needs them, with no history of allergies or family/personal history of angioneurotic edema, discusses the side effects, etc., and the patient has a fatal episode of angioneurotic edema, that is not an error. If the ACE inhibitor was the best drug for the job, that's an inherent risk. All drugs carry inherent risks; if the benefit far outweighs the risk, the drug is given. Bad reactions are not errors; there is no way to predict who will have the bad reaction. Feb 1, 2017 at 20:07
  • 1
    If the drug was not necessary, that is an error. If it was the wrong drug, that is an error. But if someone gets MDR C. diff after presenting in sepsis, that's not an error; it's an inherent risk. I will not argue that errors occur; I never have. They do, and they are not infrequent. But the 3rd leading cause of death? I simply don't see that many deaths caused by medical errors. God knows I've had lots of opportunity to do so. Feb 1, 2017 at 20:11
  • 1
    fed law states hospitals assume the cost for harm they cause that extends patient's care. It's hell in the ER, so I understand, ICU wasn't much better, as a former scientist I learned never put a drug into the body unless the doc knows the risks, God I still think of that poor MI that got a thrombolytic after he had a heart attack on the toilet, No one did a head CT, he had fallen, I got report and he was already well on his way to brain dead, my hat is off to you for your service is why I only write now and do home health amazon.com/Bottom-Line-Laymans-Guide-Medicine/dp/0875864554 Feb 1, 2017 at 20:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .