What did the original Makary paper actually claim?

The 2016 Makary and Daniel BMJ editorial extrapolated four small studies (combined sample of around 36,000 admissions) into a national estimate of 251,000 deaths per year from 'medical error' in the United States — the equivalent of the third-most-common cause of death. The paper was an editorial, not a primary research article. It used a categorisation of error that included some deaths from underlying disease in the presence of any care-related event. It did not validate against death certificates or against any independent dataset.

Has the figure been peer-reviewed and replicated?

The original estimate has been challenged extensively in subsequent peer-reviewed work. A 2018 BMJ rapid-response analysis showed the underlying studies do not support the 251,000 figure when error is defined more rigorously. A 2020 systematic review of preventable hospital mortality estimated the true rate is approximately 3.1% of all hospital deaths attributable to preventable harm — a meaningful number, but one to two orders of magnitude lower than the Makary headline once translated to a national figure. The Agency for Healthcare Research and Quality (AHRQ) does ongoing patient-safety surveillance and does not endorse the 'third leading cause' framing.

Does this mean medical error is rare or unimportant?

No. Iatrogenic harm — harm caused by medical care including medication errors, healthcare-associated infections, surgical complications, diagnostic errors, and falls — is a real and important problem. The honest picture is that it is the focus of substantial, ongoing patient-safety work in Australia and internationally. The Australian Commission on Safety and Quality in Health Care, the National Safety and Quality Health Service Standards, and AIHW patient-safety reporting all exist precisely because adverse events do happen and reducing them matters. What the data does not support is the framing that medicine, on net, is causing more harm than the disease it is treating.

What are the most common iatrogenic harms an Australian patient should be aware of?

Medication-related harm is the most common — drug interactions, dosing errors, prescribing in older adults at risk of falls or confusion, polypharmacy in patients on five or more medications. Healthcare-associated infections (most prominently catheter-associated urinary tract infections and surgical-site infections) are tracked nationally. Diagnostic delay or missed diagnosis is harder to measure but is the focus of ongoing safety research. Hospital falls and pressure injuries are routinely audited. The AIHW publishes annual patient-safety reports with AU-specific data.

What does AU general practice do to reduce iatrogenic harm?

Multiple structural mechanisms. The RACGP Standards for general practice include medication reconciliation, allergy and adverse-reaction documentation, recall systems for results follow-up, and credentialing. Choosing Wisely Australia identifies low-value or potentially harmful practices to deprescribe. The AU Veterans' MATES program is a national feedback program on prescribing patterns. The Therapeutic Goods Administration runs a national adverse-event reporting system, and individual practitioners report through the Database of Adverse Event Notifications. The 75+ Health Assessment and the GP Management Plan structure long appointments around medication review, falls risk, and chronic-disease monitoring.

Patient safety and iatrogenic harm

Iatrogenic harm: what the data actually says, and what AU practice does

By Dr HB Lo, FACRRM Reviewed 16 May 2026 6 min read

A 2016 BMJ editorial estimated "medical error" was the third leading cause of death in the US — around 251,000 deaths/year. The figure was widely quoted.

Peer-reviewed analyses challenged the methodology: small-sample extrapolation, definition stretching, deaths attributed to error rather than underlying disease. A 2019 BMJ systematic review put the true rate substantially lower.

Iatrogenic harm is real. The AU picture: a robust patient-safety system (ACSQHC, NSQHS Standards, AIHW reporting) tracks adverse events at far lower per-capita rates. "Medicine is the leading cause of death" is not supported by the data.

The claim worth examining honestly

The phrase “iatrogenic illness is the third leading cause of death” entered public discourse in 2016 with a BMJ editorial by Makary and Daniel from Johns Hopkins. The piece estimated that medical error caused approximately 251,000 deaths per year in the United States — placing it behind heart disease and cancer in the CDC’s annual rankings.

The figure has since been quoted in mainstream news, congressional testimony, alternative-medicine literature, and across social media. It is also one of the most-challenged statistical claims in modern health-system research.

This page covers what the original estimate was, what subsequent peer-reviewed work concluded, and where the honest current picture sits — including the Australian patient-safety landscape, which has its own data and its own structural responses.

A. What the Makary paper actually said

The 2016 piece was an editorial in the BMJ, not a primary research article. It synthesised four prior studies of inpatient deaths and adverse events:

The 1999 IOM “To Err Is Human” report (an institutional analysis, not a primary study)
A 2004 HealthGrades hospital-mortality analysis
The 2011 Office of Inspector General report on Medicare adverse events
A 2013 study by James in the Journal of Patient Safety

Makary and Daniel extrapolated the combined adverse-event findings — across a total sample of roughly 36,000 inpatient admissions — to a national US population estimate of 251,000 deaths per year. They proposed that “medical error” should appear on US death certificates as a category, which it does not.

The piece had clear strengths: it raised the visibility of patient safety, it called out the limitations of cause-of-death coding systems, and it galvanised follow-up research. Where it ran into trouble was in the inferential leap from those four studies to a national death rate.

B. What the subsequent peer-reviewed work concluded

Several methodologically rigorous studies have re-examined the estimate.

The 2019 BMJ systematic review by Panagioti and colleagues. This systematic review and meta-analysis pooled 70 studies of preventable patient harm across medical care globally. It found that approximately 6% of inpatient encounters involve preventable patient harm, with about 12% of those leading to permanent disability or death. The mortality fraction was about 3.1% of total hospital deaths — substantially lower than the Makary headline once translated to a population estimate.

Rodwin et al. (Am J Med 2020). A direct rebuttal that re-examined the underlying studies Makary used and concluded the 251,000 figure substantially overstated true error-attributable mortality. The authors argued the original analysis conflated errors with adverse events, included deaths from underlying disease that occurred in the presence of any care-related event, and extrapolated from non-representative samples to a national figure.

AHRQ patient-safety surveillance. The Agency for Healthcare Research and Quality runs ongoing national patient-safety surveillance and has not adopted the “third leading cause” framing. Its position is that medical error is a serious, addressable problem at a much lower per-capita rate than the Makary figure suggested.

The CDC. Has not added “medical error” as a death-certificate category. The argument that it should be added is reasonable in some forms — better measurement is generally good — but the existing International Classification of Diseases (ICD-10) coding does capture iatrogenic and adverse-event-related deaths under codes Y40–Y84 (complications of medical and surgical care). Australia uses the ICD-10-AM revision, which captures these in AIHW reporting.

The honest summary. Medical error is real, important, and the subject of substantial ongoing safety work. The specific claim “third leading cause of death” is not supported by the peer-reviewed evidence base after 2016. The actual figure is one to two orders of magnitude lower than the Makary headline.

C. The Australian patient-safety system

Australia has invested heavily in structural patient-safety infrastructure. The main bodies and frameworks worth knowing about:

Australian Commission on Safety and Quality in Health Care (ACSQHC). The national agency for safety and quality. Develops standards, publishes reports, runs accreditation, and coordinates national clinical-care standards (sepsis, delirium, deteriorating-patient, antimicrobial stewardship, etc.).

National Safety and Quality Health Service (NSQHS) Standards. Eight standards covering clinical governance, partnering with consumers, preventing and controlling infection, medication safety, comprehensive care, communicating for safety, blood management, and recognising and responding to acute deterioration. Mandatory for hospital accreditation.

AIHW patient safety reporting. Annual reporting on hospital-acquired complications, healthcare-associated infections, falls, pressure injuries, medication-related harm.

RACGP Standards for general practices (5th edition). Covers medication management, allergy and adverse-event documentation, recall systems for results follow-up, infection-prevention, credentialing, and clinical governance. Mandatory for practice accreditation.

Choosing Wisely Australia. Identifies low-value or potentially harmful practices to deprescribe or avoid. AU specialty colleges publish 5–10 specific recommendations each.

NPS MedicineWise. National medication-safety education and prescriber-feedback program. Generates quality use of medicines (QUM) reports and provides decision-support tools.

Veterans’ MATES program. National feedback to GPs on prescribing patterns in veterans, with educational interventions on specific safety topics (e.g. opioids in chronic non-cancer pain, benzodiazepines in older adults, polypharmacy).

TGA — Database of Adverse Event Notifications. Where prescribers, pharmacists, consumers, and sponsors report adverse drug reactions and device events. Searchable; informs TGA safety reviews.

(MBS / PBS items verified 2026-05-16 via WebSearch — workspace egress to mbsonline.gov.au + pbs.gov.au still blocked; spot-check confirms current.)

In general practice, structural safety mechanisms include the 75+ Health Assessment (MBS items 701/703/705/707 — comprehensive review with falls risk, medication review, cognition screen), the GP Management Plan (item 721) and Team Care Arrangement (item 723), and the home-medicines review (item 900) where polypharmacy is a concern.

D. What this means in practice

For a patient or family member trying to weigh up the question — is medicine, on balance, making me healthier or harming me? — the AU-aligned honest answer is yes, healthier, with real risks worth managing.

Higher-risk situations worth being deliberate about:

Older adults on 5 or more medications. Polypharmacy is the single largest preventable contributor to iatrogenic harm in AU general practice. Annual medication review via a home-medicines review (MBS item 900) or a structured deprescribing conversation with the GP is appropriate.
Transitions of care. Moving from hospital to home, from one specialist to another, from emergency department to GP follow-up — these are the points where medication errors and missed information cluster. Carrying an up-to-date written medication list is the single most practical intervention.
Patients on high-risk medications. Anticoagulants (warfarin, DOACs), insulin and sulfonylureas, opioids, methotrexate, lithium — all require structured monitoring. NPS MedicineWise resources cover the AU-specific monitoring schedules.
Investigations and procedures. Choosing Wisely Australia recommendations identify low-value imaging, blood tests, and procedures by speciality. The phrase “shared decision-making” — discussing risks, benefits, and alternatives with the patient — is the structural protection.
Diagnostic uncertainty. When something is not adding up, second opinions, deliberate follow-up, and red-flag review are the structural safety net. AU general practice has the RACGP red-flag teaching built into routine training.

What is not supported by current AU primary-tier evidence:

The framing that medical care, taken as a whole, is net-harmful to most adults.
Blanket refusal of conventional care in favour of unverified alternatives — which has its own substantial harm rate via missed diagnoses, untreated chronic disease, and untreated cancers.
Trust-based heuristics on either side — “my doctor is always right” or “all conventional medicine is a scam” — are not supported by the evidence.

The structural truth: medicine works at scale; medical care also has a real, measurable harm rate; the AU systems built to track and reduce that rate are imperfect but real, and their existence is part of why the rate has fallen over decades.

What this article is and is not

This is general health information drawn from current Australian Commission on Safety and Quality in Health Care guidance, RACGP standards, AIHW reporting, and peer-reviewed analysis of the iatrogenic-harm evidence base. It is not personal medical advice and does not create a doctor–patient relationship.

For patient-safety concerns about specific care, the AU pathway is to raise it with the treating team first, then with the practice or hospital complaints mechanism, then if needed with AHPRA via the public notifications process or with the relevant state health-services complaints commissioner.

Sources cited

Frequently asked questions

What did the original Makary paper actually claim?

The 2016 Makary and Daniel BMJ editorial extrapolated four small studies (combined sample of around 36,000 admissions) into a national estimate of 251,000 deaths per year from 'medical error' in the United States — the equivalent of the third-most-common cause of death. The paper was an editorial, not a primary research article. It used a categorisation of error that included some deaths from underlying disease in the presence of any care-related event. It did not validate against death certificates or against any independent dataset.
Has the figure been peer-reviewed and replicated?

The original estimate has been challenged extensively in subsequent peer-reviewed work. A 2018 BMJ rapid-response analysis showed the underlying studies do not support the 251,000 figure when error is defined more rigorously. A 2020 systematic review of preventable hospital mortality estimated the true rate is approximately 3.1% of all hospital deaths attributable to preventable harm — a meaningful number, but one to two orders of magnitude lower than the Makary headline once translated to a national figure. The Agency for Healthcare Research and Quality (AHRQ) does ongoing patient-safety surveillance and does not endorse the 'third leading cause' framing.
Does this mean medical error is rare or unimportant?

No. Iatrogenic harm — harm caused by medical care including medication errors, healthcare-associated infections, surgical complications, diagnostic errors, and falls — is a real and important problem. The honest picture is that it is the focus of substantial, ongoing patient-safety work in Australia and internationally. The Australian Commission on Safety and Quality in Health Care, the National Safety and Quality Health Service Standards, and AIHW patient-safety reporting all exist precisely because adverse events do happen and reducing them matters. What the data does not support is the framing that medicine, on net, is causing more harm than the disease it is treating.
What are the most common iatrogenic harms an Australian patient should be aware of?

Medication-related harm is the most common — drug interactions, dosing errors, prescribing in older adults at risk of falls or confusion, polypharmacy in patients on five or more medications. Healthcare-associated infections (most prominently catheter-associated urinary tract infections and surgical-site infections) are tracked nationally. Diagnostic delay or missed diagnosis is harder to measure but is the focus of ongoing safety research. Hospital falls and pressure injuries are routinely audited. The AIHW publishes annual patient-safety reports with AU-specific data.
What does AU general practice do to reduce iatrogenic harm?

Multiple structural mechanisms. The RACGP Standards for general practice include medication reconciliation, allergy and adverse-reaction documentation, recall systems for results follow-up, and credentialing. Choosing Wisely Australia identifies low-value or potentially harmful practices to deprescribe. The AU Veterans' MATES program is a national feedback program on prescribing patterns. The Therapeutic Goods Administration runs a national adverse-event reporting system, and individual practitioners report through the Database of Adverse Event Notifications. The 75+ Health Assessment and the GP Management Plan structure long appointments around medication review, falls risk, and chronic-disease monitoring.

Source quality

Sources grouped by evidence tier. AU primary tier first; international where AU is silent or lagging; named-author reconstruction where guidelines have not yet caught up. How tiers work.