Home

Reminder

Certainty is rare in medicine.
Uncertainty is normal.

Not mathematics Medicine operates on likelihoods, incomplete information, and evolving data. A diagnosis is not always certain. It is a ranked probability.
Not weakness Uncertainty is the normal state of clinical practice. It is not a weakness.
The weakness lies in unstructured thinking in the face of uncertainty.
Not ignorance "Two diagnoses are possible and I have a plan" is not the same as "I don't know."
The first is clinical uncertainty managed with structure. The second, on its own, is not yet clinical reasoning.
The Core

A diagnosis is not a fact you retrieve. It is a probability you calculate.

Probabilities

Diagnosis Tiers

Tier 1: working diagnosis

Most Likely

Supported by the balance of evidence from history and examination. This tier represents your leading working diagnosis.

Tier 2: keep on the list

Less Likely

Partial evidence supports it. Not enough to act on first, but enough to investigate or keep under consideration. It remains on the differential until further evidence lowers its probability sufficiently or makes it less clinically relevant.

Tier 3: must not miss

Unlikely but Dangerous

Low probability, but the consequence of missing it is catastrophic. It requires deliberate consideration and, when appropriate, targeted investigation or exclusion.

These three tiers map directly onto the Differential Diagnosis note.
Every diagnosis you retain on your differential should be assigned to one of these tiers. If you cannot tier a diagnosis, that is a signal to examine your evidence base and reasoning, not necessarily to exclude the diagnosis.
The Problem

Why learners freeze, and why they should not

What learners believe

  • There must be one correct diagnosis.
  • If I cannot name a diagnosis, I have failed.
  • Being unsure means I am wrong.
  • Examiners expect certainty.
  • Ambiguity is a sign of ignorance.

What examiners actually assess

  • Can you commit to a working diagnosis?
  • Can you justify it with evidence?
  • Can you challenge your own reasoning?
  • Can you name what you do not know?
  • Can you state a plan to resolve it?
Take note

Attendings are not asking: "Do you know the answer?"
They are asking: "Can you reason like a doctor?"

You are not alone

Anxiety and shame in the face of uncertainty are common among medical students and trainees. They are understandable and often arise from a mistaken belief: that uncertainty means incompetence..
It does not.
The discomfort you feel is not evidence of incompetence. It may simply reflect how much you care about getting it right. The framework is what you do with that discomfort.

The Distinction

Ignorance versus clinical uncertainty

A critical distinction for every learner to internalize

Two very different states of not knowing

Uncertainty, unstructured
"I don't know what this is."
No differential or competing hypotheses identified
Evidence has not been organized or weighed
No clear plan to reduce the uncertainty
Uncertainty is experienced but not explicitly managed
Clinical uncertainty, structured
"These are the leading possibilities."
Differential diagnoses identified and ranked
Evidence for and against each possibility is weighed
Next steps are chosen to reduce uncertainty or manage risk
Uncertainty is explicitly recognized, monitored, and managed
Uncertainty becomes unsafe primarily when it is unrecognized or unmanaged.
On the other hand, recognized uncertainty without a resolution plan is also incomplete.
A clinician who knows they are uncertain, and has a structured plan, is far safer than one who is confidently wrong.
The Framework

Clinical Medicine = The discipline of structured uncertainty

Bedside & Oral Exam Framework

How to reason out loud when you are not certain

Apply in sequence. Each step builds on the one before. In practice, experienced clinicians may move fluidly between steps, but for learning, sequential application builds the habits that fluency later requires.
1

Commit

Say it: State your provisional diagnosis:
"My provisional diagnosis is X."
Commitment is not arrogance. It is the starting point for all reasoning that follows.

2

Justify

Defend it: List the positively relevant findings:
"because of A, B, and C."
These are the features that support this diagnosis. Be specific: name the findings rather than describing them in general terms.

3

Challenge yourself

Question it: Name the points against it:
"However, the absence of D makes it less classical." Or: "However, Y is also possible because of E."
This is not doubt; it is rigor.

4

Acknowledge what is missing

Name it: State explicitly what information would change your thinking:
"I would want to know whether…"
Naming what you do not know is not weakness. It is intellectual honesty.

5

Plan resolution

Act on it: State how you would confirm or exclude:
"I would confirm with X / I would exclude Y by doing Z."
Every uncertainty needs a next step. If there is no plan, the uncertainty is unmanaged.

This framework converts panic into structure. A learner who cannot find a diagnosis is not stuck: they are at step 1, and the subsequent steps will carry them forward.
Take note

Not all findings carry equal weight. A specific finding that strongly favors one diagnosis over another counts for more than several non-specific features that are consistent with many conditions. When justifying or challenging a diagnosis, ask not just what fits, but how strongly each finding shifts the probability.

In Practice

How to reason uncertainly

An 8-month-old infant with underlying ventricular septal defect and growth faltering presents with respiratory distress.
Do not say this
"This is heart failure."
No justification given: could be memory or luck
No differential considered
No acknowledgment of incomplete picture
No plan stated
Attending cannot assess your reasoning
Say this instead
"An infant with known ventricular septal defect and growth faltering who presents with respiratory distress is suggestive of heart failure due to a significant left-to-right shunt. However, respiratory distress could also be due to a respiratory tract infection. I would assess for fever, cough, and increased work of breathing, and examine the lungs (crepitations), heart (murmur, tachycardia), and liver (hepatomegaly). I would confirm the diagnosis with a chest X-ray and echocardiogram."
Committed to a provisional diagnosis
Justified with specific clinical findings
Challenged own reasoning: acknowledged an alternative diagnosis
Offered an explanation for the uncertainty: named what is missing
Named a clear plan to confirm
Both students reached the same diagnosis. One demonstrated recall. One demonstrated clinical reasoning. Only one is safe to practice.
The most dangerous doctor

It is not the uncertain one.
It is the one who is certain too early.

Safety-netting

It is a practical tool for managing residual uncertainty at the point of discharge or end of consultation; it does not substitute for a resolution plan, but complements it.
When you cannot be fully certain: tell yourself what to watch for, when to re-evaluate, and under what circumstances to seek urgent help from colleagues or an attending.
Seeking help is not admitting failure. It is responsible clinical practice, and it protects both the patient and the doctor.

The Metaphor

Narrowing the circle

Medicine is not about being right at the first step.
It is about reducing uncertainty safely.
At admission
○ Wide

Many diagnoses remain possible. The circle of uncertainty is large. History begins to narrow it, but you do not yet know.

After examination & initial investigations
◎ Narrowing

Each finding shifts the probabilities. Some diagnoses fall away. One or two remain. The circle shrinks with each piece of evidence.

At discharge
• Small

The diagnosis is confirmed, or confidently assigned. The circle is small. The uncertainty has been managed, not eliminated.

Your job is not to eliminate uncertainty instantly, but to shrink it logically. Each step is not a guess. It is a Bayesian update: new information modifies the probabilities. This is what clinical medicine is.
Take note

The circle does not always narrow smoothly. An unexpected investigation result, a newly revealed co-morbidity, or a clinical deterioration inconsistent with the working diagnosis can widen the circle before it narrows again. When this happens, treat it as evidence, not failure.
Reopening a differential is not a step backwards; it is the framework working as intended.

Clinical Toolkit

What to say when you are not sure

Converting panic into structure

The Uncertainty Script

These phrases do not hide uncertainty. They demonstrate that you know how to manage it. That is precisely what examiners want to see.
1
"Based on the available information …" Opens the reasoning: signals that you are drawing conclusions from evidence, not guessing.
2
"My provisional diagnosis is…" Commits. The word "provisional" signals appropriate epistemic humility; you are not overclaiming.
3
"This is supported by… / because of…" Justifies. Name specific findings, not vague generalities. Evidence, not assertion.
4
"However, I cannot exclude… / I also consider…" Challenges. Opens the differential. Shows you are not anchored to your first answer.
5
"The absence of… makes… less likely, but does not exclude it." Uses pertinent negatives actively: shows you understand that absence of a sign is evidence, not silence.
6
"To confirm, I would… / To exclude, I would…" Plans. Every uncertainty must have a next step. This is the hallmark of a safe clinician.
When applied in sequence, these six phrases constitute a structured and mature clinical response that demonstrates reasoning, humility, and a plan. They demonstrate insight, safety, and clinical maturity simultaneously.
Common Errors

The most common diagnostic error in medicine

Accepting the first reasonable diagnosis without testing it against the evidence A reasonable fit is not the same as the best fit.
Ignoring features that do not fit the working diagnosis Unexplained findings are data, not noise to be dismissed.
Stopping thinking once a diagnosis label has been assigned The diagnosis is a hypothesis, not a conclusion, until confirmed.
Overconfidence in the absence of a complete clinical picture Confidence without evidence is a patient safety risk: calibrate certainty to evidence weight, not to feeling.
In short

Diagnostic uncertainty is not a gap in knowledge to be hidden , it is the normal operating state of clinical medicine, and the discipline lies in naming it clearly, ranking what you know, identifying what remains unknown, and committing to the next step that will move you closer to resolution.

Key Takeaway
Uncertainty is normal.
Unstructured thinking about it is not.

Explaining uncertainty demonstrates insight, safety, and clinical maturity.
That is what makes a good doctor.

Commit to a working diagnosis Challenge your own reasoning Name what you do not know