What is a diagnostic error?
A diagnostic error is a diagnosis that is missed, wrong or delayed, where the information available at the time should reasonably have led to a different or earlier conclusion. Not every bad outcome is an error. Some conditions present in ways no reasonable clinician would recognise. The question a reviewer, coroner or regulator asks is whether the assessment and decision were reasonable given what was known, or could reasonably have been found out, at the time.
Concerns about assessment and diagnosis are among the most common clinical concerns raised about doctors, and failures to assess and recognise deterioration are frequently raised about nurses and midwives. In many cases the clinician knew the relevant medicine. What failed was the thinking process, the system around them, or communication with the patient.
Patterns that recur in missed diagnoses
- Sepsis treated as a simple viral illness, especially in children, older people and people with learning disabilities.
- Cancer symptoms put down to a benign cause over several consultations without review.
- Cardiac chest pain or aortic dissection labelled as musculoskeletal pain or anxiety.
- Stroke or subarachnoid haemorrhage attributed to migraine or intoxication.
- Ectopic pregnancy diagnosed as gastroenteritis or a urinary infection.
- A new problem in an inpatient blamed on the existing diagnosis.
How do clinicians make diagnostic decisions?
A useful simplification is the dual process model. Type 1 thinking is fast pattern recognition built from experience. Type 2 thinking is slow and deliberate, working through possibilities step by step. Experienced clinicians rely on Type 1 most of the time, and rightly so. Errors happen when a presentation looks familiar but is not, and nothing prompts a switch to Type 2 to check.
| Type 1: intuitive | Type 2: analytical | |
|---|---|---|
| How it works | Fast, automatic pattern recognition | Slow, deliberate reasoning |
| Best for | Common, typical presentations | Unusual, complex or high-risk presentations |
| Main weakness | Vulnerable to bias, fatigue and atypical presentations | Time-consuming and easily skipped when busy |
Fatigue, hunger, interruptions, high workload and emotional pressure all make it more likely that fast thinking runs unchecked. Bias is also more likely when someone has already given you a diagnosis, when you find a patient difficult, or when the presentation is atypical, for example in older people, people with diabetes, pregnant women, young children or people who find it hard to describe symptoms.
Which cognitive biases cause missed diagnoses?
Cognitive biases are predictable errors in thinking that affect everyone. Knowing their names does not protect you, but recognising them in real cases is where insight starts.
| Bias | What happens | Clinical example |
|---|---|---|
| Anchoring | Sticking to a first impression despite new information | Deciding chest pain is musculoskeletal and dismissing a later finding of unequal arm blood pressures |
| Premature closure | Accepting a diagnosis before it is confirmed and stopping the search | Diagnosing a urinary infection from a dipstick in a confused older person and looking no further |
| Diagnostic momentum | A label gains certainty each time it is handed on | 'Known anxiety' at triage leads every clinician to assume breathlessness is panic |
| Confirmation bias | Noticing only information that fits | Focusing on a normal first ECG and overlooking that a repeat troponin is still due |
| Availability bias | Judging likelihood by what comes to mind easily | After several viral illnesses, assuming the next feverish child has the same |
| Search satisficing | Stopping once one abnormality is found | Treating a wrist fracture after a fall without asking why the person fell |
| Attribution and stigma | Blaming symptoms on a person's characteristics or history | Assuming someone who drinks alcohol is intoxicated when they have a head injury or low blood glucose |
How can you reduce diagnostic error?
No single strategy prevents every error, but some habits reliably lower the risk. They work best used routinely, because the cases where you feel most certain are often the ones where bias is at work.
- Take a diagnostic timeout: Before you finalise a diagnosis or discharge, ask: does everything fit, is there anything I cannot explain, and what else could this be?
- Ask the worst-case question: Name the most dangerous diagnosis that fits and decide whether you have reasonably excluded it. Hold at least three possibilities: most likely, most dangerous and one alternative.
- Look at the observations and trends: Review the full set and, for adults, the NEWS2 score where appropriate. A rising respiratory rate is an early warning that is easily missed.
- Treat labels as hypotheses: When you inherit a diagnosis from triage or handover, take the key history again yourself.
- Listen to patients and families: When someone says 'this is not like him', take it seriously. Families often notice changes first.
- Get a second opinion: Talk uncertain cases through with a senior or colleague. Explaining your reasoning aloud often exposes gaps.
- Close the loop on tests: Be clear who will review each result you request and what happens if it is abnormal. Unread results are a common cause of delay.
A 58-year-old man with a history of back strain comes in with sudden severe back pain spreading to his abdomen and asks for stronger painkillers. The likely answer is mechanical back pain. The worst case that fits is a leaking abdominal aortic aneurysm. Asking that question leads the clinician to check his blood pressure, examine his abdomen, consider his age and smoking history, and discuss with a senior or arrange imaging if there is any doubt. It takes a minute, and it is the moment slow thinking checks fast thinking.
What does good safety-netting look like?
Safety-netting manages uncertainty. It accepts that some diagnoses cannot be made at first contact and puts a plan in place to catch them if things change. Good safety-netting tells the patient what you think is happening and that you are not completely sure, which specific symptoms should make them seek help, how quickly and where, when they should expect to improve, and what follow-up is arranged. Check understanding, for example by asking them to repeat it back, and give written information if you can.
Safety-netting given. Return if worse.
Likely viral chest infection. Explained I cannot fully exclude early pneumonia. Advised to call 999 or go to the emergency department if breathless at rest, chest pain, coughing blood, or confused or drowsy. Contact the practice if fever lasts beyond three days or not improving within a week. Daughter present; both repeated the advice back correctly. Written information given.
The weak entry is common and nearly impossible to rely on, because no one can tell what the patient was told. The strong one shows reasoning, admits uncertainty, names specific symptoms and timescales and confirms understanding. Recording relevant negatives, such as 'no neck stiffness, no rash', also shows that you looked for serious causes.
What should you do after a missed diagnosis?
Finding out you missed a diagnosis is distressing. Guilt, anxiety and loss of confidence are normal reactions. Speak to your supervisor or a trusted colleague, and if it is affecting your wellbeing, your GP, occupational health or NHS Practitioner Health can help.
- Act for the patient first: if they can still be helped, make sure the right assessment and treatment happen urgently.
- Be open with the patient or their family under the professional duty of candour. Explain, apologise and say what will happen next.
- Report the incident and take part honestly in any learning review. In England many organisations use the Patient Safety Incident Response Framework.
- If there is a complaint, inquest or regulator referral, contact your defence organisation or trade union before responding.
How do you show insight after a missed diagnosis?
Regulators want to know whether you understand specifically what went wrong and whether the risk of it happening again has been reduced. 'It was an unusual presentation' is not insight on its own. Naming the exact point where your reasoning went wrong is.
With hindsight it was an atypical presentation and I could not have known. The department was very busy. I have read about chest pain since.
I accepted the triage label of anxiety and did not take the history again myself. I anchored on that label and did not repeat the observations, which would have shown a rising respiratory rate. I should have asked what else could explain her breathlessness. I now take a diagnostic timeout before every discharge, I note a worst-case differential, and my supervisor has reviewed twenty of my discharge records over three months.
| Remediation action | Evidence |
|---|---|
| Learning on clinical reasoning and the relevant condition | Certificate and a reflection applying it to your practice |
| Case-based discussion with a supervisor | Signed records of discussions over several months |
| Audit of your notes for differentials and safety-netting | Audit results reviewed by a senior colleague |
| Sharing the learning | Record of a teaching session or case presentation |
Taking this further
This guide is educational and does not replace national guidance, local pathways or senior advice. Our course Clinical Decision-Making and Missed Diagnosis works through clinical reasoning, cognitive bias in real cases, debiasing strategies, safety-netting and documentation, and ends with a structured reflection and remediation plan. The related courses on medication errors and record keeping can support a wider learning plan.
Frequently asked questions
What is anchoring bias in medicine?
Anchoring is when a clinician locks on to an early impression, such as a triage label or their first idea, and fails to adjust when new information appears. For example, deciding chest pain is muscular and then dismissing a later abnormal finding. A diagnostic timeout and a worst-case question help counter it.
What is premature closure?
Premature closure means accepting a diagnosis before it has been properly confirmed and then stopping the search for other explanations. It is one of the most common reasoning errors behind missed diagnoses. Asking whether everything fits, and what you cannot yet explain, helps prevent it.
What is the most common cause of diagnostic error?
Reviews of diagnostic error generally find that it usually results from a combination of thinking errors, such as anchoring and premature closure, and system factors, such as workload, interruptions, handover and test results not being followed up. A lack of medical knowledge is less often the main cause.
How do you document safety-netting?
Record what you think is happening and any uncertainty, the specific symptoms that should prompt the patient to seek help, how quickly and where to go, the expected time to improve, any follow-up arranged, and that the patient understood. 'Return if worse' on its own is not enough.
Will I be referred to the GMC or NMC for a missed diagnosis?
Most missed diagnoses are handled locally through incident review and learning. A referral is more likely with serious harm combined with a pattern of concerns, a lack of insight or dishonesty, such as altering records. If a referral is possible, contact your defence organisation or union before responding.
What is diagnostic overshadowing?
Diagnostic overshadowing is when physical symptoms are wrongly attributed to an existing mental health condition, learning disability or autism. It has been identified repeatedly in reviews of deaths of people with learning disabilities. Always consider pain, infection, constipation and other physical causes for a change in behaviour.
Related courses
Clinical Decision-Making and Missed Diagnosis
This course explains how clinicians make diagnostic decisions, why skilled people miss or delay diagnoses, and the practical strategies that reduce…
Medication Errors: Reflection and Remediation
This course explains why medication errors happen, what you must do when one occurs, and how to show a regulator or employer that you have understood…
Record Keeping and Documentation: Remediation Course
This course explains what good clinical documentation looks like, why record-keeping concerns arise, and why altering or falsifying a record is…
Sources and further reading
- Good Medical Practice (2024). General Medical Council.
- The Code: Professional standards of practice and behaviour for nurses, midwives and nursing associates. Nursing and Midwifery Council.
- Suspected cancer: recognition and referral (NG12). National Institute for Health and Care Excellence.
- Suspected sepsis in people aged 16 or over: recognition, assessment and early management (NG253). National Institute for Health and Care Excellence.
- National Early Warning Score (NEWS) 2. Royal College of Physicians.
- Patient Safety Incident Response Framework. NHS England.
Practitioner Insight is an independent education provider. It is not affiliated with or endorsed by the GMC, MPTS, NMC, HCPC, GPhC or any NHS organisation, and this information is not legal advice. If you have been referred to a regulator, speak to your defence organisation, trade union or legal representative.
More guides
When to Escalate a Deteriorating Patient: Observations, Assessment and Escalation for Nurses
When to escalate a deteriorating patient: NEWS2 triggers, ABCDE assessment, SBAR calls and what to do if no one comes. A practical guide for UK nurses.
Human Factors in Healthcare: The Swiss Cheese Model With Examples
The Swiss cheese model in healthcare explained with clinical examples: active failures, latent conditions, human factors and how to use it in a reflection.
PSIRF Explained: What the Patient Safety Incident Response Framework Means for Staff
PSIRF explained for NHS staff: what the Patient Safety Incident Response Framework is, its four aims, learning responses, and what to do if you are involved.