Knowing When to Say I Don’t Know

Three words can feel unusually difficult in medical training.

“I don’t know.”

Students often experience those words as exposure.

You worry that an attending will think you did not prepare. You worry that a patient will lose confidence. You worry that the answer was something everyone else already knew.

So the temptation is to fill the silence.

Guess.

Talk around the question.

Offer a partially remembered fact with more confidence than it deserves.

That response may protect appearance for a few seconds. It can also make learning and communication less reliable.

The more useful skill is knowing how to acknowledge uncertainty and then move the problem forward.

Clinical medicine contains unavoidable unknowns

No clinician remembers everything.

More importantly, many clinical questions do not have an immediate complete answer.

The diagnosis may still be evolving.

The literature may be mixed.

A patient may have several reasonable options.

A medication detail may need verification.

A test result may not yet be available.

A consultant may be needed.

The professional task is not to eliminate every unknown before speaking.

The task is to recognize the unknown accurately enough to decide what happens next.

Separate a knowledge gap from a reasoning gap

“I don’t know” can mean several things.

You may not remember a fact.

You may understand the facts but not know how to weigh them.

You may be uncertain between two plausible diagnoses.

You may know the diagnosis but not the recommended management sequence.

You may know what you would usually do but be unsure how a comorbidity changes the decision.

These are different problems.

Name the type of uncertainty when you can.

That makes the follow-up more efficient.

Show the part you do know

A useful answer often starts with the boundary of your knowledge.

For example:

“I do not remember the exact adverse-effect frequency, but the mechanism makes me concerned about renal function, so I would verify the dosing guidance before recommending it.”

Or:

“I am not sure which diagnosis is more likely yet. The fever and focal findings support infection, while the time course and exposure history keep the alternative active.”

That response gives your supervisor a view of your reasoning.

You are not hiding the gap.

You are also not throwing away the knowledge you do have.

Pair uncertainty with a next step

The original DDQX article emphasized the phrase:

“I don’t know, but I’m going to find out.”

The second half is what makes the response clinically useful.

Your next step may be:

• Check an authoritative reference
• Review the imaging
• Ask the pharmacist
• Repeat part of the examination
• Clarify the history
• Ask a supervisor
• Follow up with the patient later
• Read the relevant guideline after rounds

The follow-up should match the stakes.

A low-stakes teaching question can wait.

A patient-safety question may require immediate verification.

Know when guessing becomes unsafe

There is a difference between reasoning under uncertainty and inventing information.

Clinical reasoning often requires you to make a provisional judgment.

That judgment should be based on available evidence and communicated as provisional.

A factual question about a medication dose, contraindication, allergy, or other safety-sensitive detail should be verified rather than guessed when you are unsure.

The same applies to patient communication.

If you do not know when the specialist will arrive, do not create a time.

If you do not know what a test means, say that you want to review it before interpreting it.

Accuracy deserves priority over performance.

Use calibrated language with patients

Patients do not require false certainty.

They do require honesty and follow-through.

A useful structure is:

“I do not have that answer yet. I want to make sure I give you accurate information. I am going to check with the team and come back.”

Then come back.

The credibility comes from the combination of transparency and reliability.

If you repeatedly promise follow-up and fail to provide it, the words lose value.

Use uncertainty to invite team intelligence

Medicine is collaborative because no single clinician holds every relevant perspective.

A well-placed “I am not sure” can open the door to another discipline.

Nursing may have observed the trajectory over several hours.

Pharmacy may identify a medication issue.

Respiratory therapy may notice a change in work of breathing.

A consultant may see a pattern you have not encountered.

The goal is not to outsource your thinking.

It is to recognize when another person’s expertise should enter the model.

Saying I don’t know on rounds

When an attending asks a teaching question, the best response depends on what you know.

If you can reason toward the answer, do that.

If you know part of the concept, say the part you know.

If you truly do not know, say so directly.

Then follow through.

For example:

“I do not remember the receptor responsible for that effect. I will look it up and include it in tomorrow’s presentation.”

That is more useful than producing three unsupported possibilities and hoping one lands.

Use the moment diagnostically

A question you could not answer is data.

Ask why.

Was the fact never learned?

Was it learned but not retrievable?

Did you fail to connect the mechanism to the clinical pattern?

Was the question outside the level you should reasonably know right now?

Did anxiety block recall?

The response determines what you do after rounds.

Not every unknown deserves an hour of review.

The unknown should be triaged like any other learning need.

The 5P Approach™ can expose the boundary

The 5P Approach™ to Clinical Reasoning helps you locate uncertainty.

Prioritize may reveal that you do not know which clue matters most.

Paraphrase may reveal that the case is still poorly represented.

Prognose may reveal that you cannot predict what should happen next.

Pick may reveal that two options remain indistinguishable.

Post-Mortem reveals what needs to be learned after the case.

This is more useful than treating every moment of uncertainty as the same failure.

Intellectual humility is active

Humility in medicine does not mean shrinking from decisions.

It means maintaining an accurate relationship with the limits of your current knowledge.

You can be decisive when the evidence supports a decision.

You can revise when new information appears.

You can ask for help before uncertainty creates risk.

You can admit when you were wrong.

Those behaviors support learning because they keep the reasoning process open to correction.

Avoid using I don’t know as an endpoint

There is also a way to use uncertainty passively.

“I don’t know” can become a shield against thinking.

Before stopping, ask what you can infer.

What mechanism is relevant?

What would you predict?

Which answer can you rule out?

What information would separate the remaining options?

A good learner knows when to admit a gap and when to keep reasoning through the gap.

Both skills matter.

Follow up visibly

If you said you would find out, close the loop.

Tell the resident what you learned.

Return to the patient with the answer.

Update the presentation.

Add the corrected mechanism to your notes.

This transforms the moment from missing knowledge into demonstrated learning behavior.

The follow-up is often more memorable than the original gap.

The standard

Knowing when to say “I don’t know” is part of clinical judgment.

The phrase should be accurate.

The reasoning around it should remain visible.

The next step should be clear.

And the follow-through should happen.

That combination supports patient safety, learning, and professional credibility far better than bluffing.

Next step: The next time you reach the edge of your knowledge, name what you do know, identify the exact gap, and close the loop after you verify the answer.