How common is misdiagnosis? What the data on diagnostic errors actually shows

Short answer: more common than almost anyone assumes, and far more consequential than the word 'misdiagnosis' suggests. Diagnostic error is now the largest single category of serious harm in American medicine — and unlike a lot of patient-safety statistics, the recent numbers are concentrated, specific, and built on careful disease-by-disease work rather than a single broad extrapolation.

How common is a diagnostic error?

In 2015 the National Academy of Medicine, in its report 'Improving Diagnosis in Health Care,' reached a conclusion that still reframes the question: most people will experience at least one diagnostic error in their lifetime, sometimes with devastating consequences. Diagnosis, in other words, is not a solved problem running quietly in the background. It is a recurring point of failure.

On the frequency, the most-cited estimate comes from Hardeep Singh and colleagues, writing in BMJ Quality & Safety in 2014. Pooling three large observational studies, they estimated that about 5 percent of US adults — more than 12 million people a year — experience a diagnostic error in outpatient care alone. Most of those errors are caught or cause no lasting harm. A meaningful share are not, and do.

The serious-harm number: about 795,000 a year

The sharpest recent estimate of the damage comes from David Newman-Toker and colleagues at Johns Hopkins, published in BMJ Quality & Safety in 2024. They put the number of Americans who die or are permanently disabled by diagnostic error at roughly 795,000 a year, within a modeled range of about 549,000 to 795,000.

What makes this estimate more credible than the headline figures that came before it is the method. Earlier attempts to size diagnostic-error harm ranged absurdly widely — from about 40,000 to 4 million a year — because they extrapolated from thin data. The Hopkins team instead built the total up from disease-specific error rates and disease-specific harm rates, condition by condition. It is still an estimate, and the same caution applies here as to any patient-safety number: these harms are not counted directly, they are modeled. But it is modeling anchored to specific diseases rather than a single broad multiplier.

It is concentrated: the 'Big Three'

The most useful finding is not the total — it is where the harm clusters. Three disease categories, which the researchers call the Big Three, account for roughly three-quarters of all serious misdiagnosis-related harm: vascular events, infections, and cancers. This pattern held up across both their malpractice-claims analysis and the national incidence estimate.

Drill down further and the concentration is even sharper. Just five conditions account for close to 39 percent of the most serious harms:

• Stroke — frequently mistaken early for a benign cause of dizziness or headache.
• Sepsis — where hours of delay change outcomes.
• Pneumonia.
• Venous thromboembolism, including pulmonary embolism.
• Lung cancer — often a finding on imaging that is not flagged or followed up.

There is a counterintuitive twist in the data worth knowing. For the common dangerous diseases, error rates are relatively low — myocardial infarction and stroke are missed a small fraction of the time. The highest error rates show up in the rarer, harder-to-spot conditions: spinal abscess, aortic dissection, meningitis and encephalitis. The danger is a product of how often a disease appears and how often it is missed, which is why a short, specific list of conditions drives so much of the total harm. The researchers estimated that cutting diagnostic errors in half for just the top five could prevent on the order of 150,000 deaths and permanent disabilities a year.

Why diagnostic error is so often invisible

Diagnostic error rarely looks like an error at the moment it happens. The Hopkins work attributes the large majority of serious misdiagnoses to failures of clinical judgment — not broken equipment, not a missing test, but the wrong conclusion drawn from the information available. Nothing in the chart announces it. The harm surfaces later: the stroke sent home from the emergency department as a migraine, the pulmonary embolism read as anxiety, the lung nodule on a scan that no one circled back to.

That is exactly why these cases turn on the record rather than on any single note. The questions that decide a missed-diagnosis case are questions of sequence and timing: when symptoms were documented, when a critical result became available versus when anyone acted on it, whether a recommended follow-up was ordered, and whether a note describing the clinical picture was written contemporaneously or filled in after the outcome was known. Those facts live in the EMR's audit trail — the time-stamped record of what was created, viewed, and changed — not in the printed chart.

The shape of diagnostic error: it is common, it is concentrated in a short list of conditions, and it is largely a timeline problem — a result available but not acted on, a follow-up not closed. And the timeline lives in the record's metadata, not in the discharge summary.

What to take from this

• Diagnostic error is the largest category of serious harm in US medicine — an estimated 795,000 deaths or permanent disabilities a year.
• It is not random. Roughly three-quarters of the harm comes from vascular events, infections, and cancers, and about 39 percent from just five conditions.
• Most serious misdiagnoses are failures of clinical judgment, which means they leave no obvious mark in the chart at the time.
• Whether a specific diagnosis was reasonably missable turns on the record's timeline — what was known, when, and what was done about it.

National statistics tell you that misdiagnosis is common and predictable. They cannot tell you what happened in one patient's care. That answer lives in the sequence the record actually documents — when the result came back, who saw it, and what happened next. Reconstructing that timeline from the audit trail is what turns a general statistic about diagnostic error into a specific, documented account of a single case.

References

• Newman-Toker DE, et al. 'Burden of serious harms from diagnostic error in the USA.' BMJ Quality & Safety, 2024;33(2):109–120. qualitysafety.bmj.com
• National Academy of Medicine. 'Improving Diagnosis in Health Care.' National Academies Press, 2015. nap.nationalacademies.org/catalog/21794
• Singh H, Meyer AND, Thomas EJ. 'The frequency of diagnostic errors in outpatient care: estimations from three large observational studies involving US adult populations.' BMJ Quality & Safety, 2014;23(9):727–731. pubmed.ncbi.nlm.nih.gov/24742777
• Newman-Toker DE, et al. 'Serious misdiagnosis-related harms in malpractice claims: The Big Three — vascular events, infections, and cancers.' Diagnosis, 2019;6(3):227–240. pubmed.ncbi.nlm.nih.gov/31535832

This article is general information about patient-safety research, not medical or legal advice. Whether any particular diagnosis fell below the standard of care depends on the facts of the case and qualified clinical and legal judgment — not on a statistic.