Are medical errors the third leading cause of death? What the evidence actually shows

Short answer: one widely cited study says yes, and that single claim now anchors most of what people believe about how dangerous American healthcare is. The number behind it is real. So is the scientific dispute over it. The gap between those two facts is where most readers get misled, so it is worth walking through carefully — what the research establishes, and what it does not.

Where the 'third leading cause' claim comes from

In May 2016, The BMJ published a short analysis with a blunt title: 'Medical error — the third leading cause of death in the US.' Two Johns Hopkins researchers, surgeon Martin Makary and Michael Daniel, estimated that more than 251,000 deaths a year in the United States trace back to medical error. That figure would place medical error third on the CDC's list of leading causes of death — behind heart disease and cancer, and ahead of chronic respiratory disease, accidents, and stroke.

The framing stuck. It has been repeated in congressional testimony, patient-safety campaigns, and law-firm advertising ever since. But a number quoted this often deserves to be examined as carefully as the harm it describes.

The first thing to understand is that the 251,000 figure was not produced by counting deaths. It could not be. The US system for coding cause of death relies on ICD codes, and those codes have no field for 'a clinician's mistake.' Medical error is structurally invisible in official mortality statistics, which is exactly why every estimate in this area is an estimate rather than a tally.

Why the number is contested

Because the data cannot be counted directly, Makary and Daniel extrapolated. They took preventable-death rates from a small set of earlier studies and applied those rates to the roughly 35 million hospital admissions the US sees each year. That method is precisely what other researchers challenged.

• The underlying studies were small and not nationally representative — some looked only at Medicare patients aged 65 and over — yet the rates were applied to all admissions, including low-risk ones such as childbirth.
• No formal meta-analysis was performed. The 251,000 figure is a weighted extrapolation, not a pooled statistical estimate.
• The source studies measured harm associated with care, not deaths caused by error. Establishing that an error actually caused a death is a far higher bar than recording that harm occurred.

Critics were direct about it. The McGill University Office for Science and Society concluded flatly that medical error is not the third leading cause of death — not because errors do not kill people, but because the methodology does not support that specific ranking. The CDC has never adopted the figure into its official cause-of-death tables.

A range, not a single number

Set the headline aside and you are left with something more honest and, frankly, more useful: a range of serious estimates, produced by different methods, all pointing at a large problem.

• The Institute of Medicine's 1999 report 'To Err Is Human' estimated 44,000 to 98,000 deaths a year from medical errors in hospitals, drawn from adverse-event studies in New York, Utah, and Colorado.
• A 2013 analysis by John James in the Journal of Patient Safety, using a record-review method called the Global Trigger Tool, estimated 210,000 deaths a year at the low end and more than 400,000 at the upper bound.
• The 2016 BMJ analysis by Makary and Daniel landed at roughly 251,000, by extrapolating from prior studies.

The low and high ends differ by nearly a factor of ten, and that spread is itself the story: this is poorly measured, the estimates are sensitive to method, and careful experts land in very different places. What none of them say is zero. Even the most conservative figure — the 44,000 in the 1999 IOM report — exceeded US deaths from motor vehicle accidents that year, a comparison the report drew itself.

'Medical error' is not 'malpractice'

This distinction matters, and it is the single most common mistake in coverage of the topic. Medical error is broad: a missed lab value, a system that allowed two drugs to interact, a handoff where information was dropped. The central finding of the 1999 IOM report was that most errors come from bad systems, not bad people.

Medical malpractice is a narrow legal standard: care that fell below the accepted standard of care and that caused harm. Most medical errors never meet that bar, and the deaths in these estimates were never adjudicated as malpractice. The research is about error and preventable harm; malpractice is a legal conclusion that requires its own proof.

The takeaway in one line: 'third leading cause of death' is a contested estimate from a single 2016 paper, not a settled fact — and 'medical error' is a far broader category than 'malpractice.' Both distinctions matter the moment you move from a headline to a specific case.

Why the deaths are so hard to count

The reason all of these figures are estimates instead of counts comes back to one thing: the evidence lives in the medical record, and the record is rarely read forensically. A death certificate captures the proximate physiological cause — sepsis, cardiac arrest, respiratory failure. It does not capture whether a critical result sat unacknowledged for nine hours, whether an order was entered and then quietly changed, or whether a note was authored after the patient had already coded.

Those facts do exist. They live in the EMR's audit trail — the time-stamped, action-level layer that records what was created, edited, and deleted, and when. It is the layer that the national statistics are trying to estimate from the outside, and the layer that turns 'something felt wrong' into a documented timeline in any individual case.

What to take from this

• 'Third leading cause of death' is a contested claim from a single 2016 analysis, not a settled fact. Use it with that caveat.
• The underlying problem is real and large across every credible estimate, from 44,000 to more than 400,000 deaths a year.
• Error is not malpractice. Preventable harm is common; legally actionable malpractice is a smaller, harder-to-prove subset.
• The truth of any single case lives in the record — specifically the audit trail — not in the death certificate or the discharge summary.

If you are trying to understand what happened in a specific case, the national statistics are context, not evidence. The evidence is in the chart's metadata — the audit trail that shows what was changed, when, and by whom. That is the layer emrcheck examines, and it is where a general statistic about medical error becomes a specific, documented account of one.

References

• Makary MA, Daniel M. 'Medical error — the third leading cause of death in the US.' The BMJ, 2016;353:i2139. bmj.com/content/353/bmj.i2139
• Institute of Medicine. 'To Err Is Human: Building a Safer Health System.' National Academies Press, 1999. nap.nationalacademies.org
• James JT. 'A New, Evidence-Based Estimate of Patient Harms Associated with Hospital Care.' Journal of Patient Safety, 2013;9(3):122–128. pubmed.ncbi.nlm.nih.gov/23860193
• Jarry J. 'Medical Error Is Not the Third Leading Cause of Death.' McGill Office for Science and Society. mcgill.ca/oss

This article is general information about patient-safety research, not medical or legal advice. Whether any particular outcome reflects an error, preventable harm, or actionable negligence depends on the facts of the case and the applicable standard of care — questions for qualified clinicians and counsel, not for a statistic.