The same question can have different diagnostic yields depending on how you ask.
Clinician: "How has the fever been?"
Patient: "It's been up and down, mostly in the evenings. Some days better, some days worse."
Clinician thinks: Informative, but this does not separate malaria from a bacterial infection from an autoimmune fever. The story is received; the diagnosis is no closer.
"Does the fever come with shaking chills before it rises?"
"Do you sweat heavily once it breaks?"
"Are there days where the temperature is completely normal?"
"Have you traveled outside the city in the past month?"
A chronological, emotionally framed account of events as the patient perceived them. Rich in context, sequence, and subjective meaning, but low in diagnostic discrimination. Essential for building rapport and understanding the patient's perspective. Use open questions to gather narrative.
Specific, characterized features that are present in some diagnoses and absent in others. Discriminating power is the measure of value. Diagnostic data is often brief, precise, and comparative, not a full story. Use targeted questions to elicit diagnostic data once a hypothesis exists.
Allow the patient to respond in their own words without
constraint.
Purpose: narrative, rapport, and unexpected
disclosure.
They yield breadth, not depth, and should be used early and
returned to whenever the patient signals they have more to say.
Semi-open questions that direct attention without dictating the
answer.
Purpose: steer toward clinically relevant dimensions
(timing, severity, pattern) while preserving the patient's
voice.
Yes/no or short-answer questions testing a specific feature.
Purpose: directly measure discriminating attributes
linked to competing diagnoses. Every targeted question should be
traceable to a diagnosis on your differential.
Questions that explicitly test for the absence of features
expected in competing diagnoses.
A deliberate negative is evidence, not absence of
information.
These should appear in both the history and the presented
narrative.
Question types describe the form and purpose of a question: they are
tools you use throughout the interview. The funnel below explains when
each type becomes most useful.
They are not meant to be followed in a strict order.
A good clinical interview moves back and forth between question types
depending on the patient’s responses. Whenever a patient signals they
have more to say, return to open-ended questioning; a new detail or
unexpected disclosure can completely shift the differential
diagnosis.
In other words, your questions should follow the patient’s story and
your evolving clinical reasoning, not a rigid checklist.
Tests a specific hypothesis without suggesting the answer. The phrasing is neutral; any response is equally welcomed and clinically informative.
Implies or suggests the expected answer. The phrasing itself biases the patient toward a particular response, distorting the history and introducing anchoring error.
Begin with an open question to let the patient speak freely. This establishes rapport, surfaces unexpected disclosures, and gives the patient's own framing of the illness. This is the foundation, not the complete history.
Once the patient has spoken, steer them toward clinically relevant territory without fully closing the answer space. You are narrowing without restricting. Focused questions are the bridge between open invitation and targeted testing.
Now ask closed questions directed at the specific features that discriminate between the diagnoses on your working differential. Each question should correspond to a diagnosis you are considering. If you cannot name the diagnosis the question is testing, it is not a targeted question; it is a random one.
Once you have a leading hypothesis, draw on both closed questions and deliberate pertinent negatives, asking for features that would be highly specific or sensitive for that diagnosis, and explicitly testing for the absence of features expected in diagnoses you need to rule out. This stage does not introduce a new question form; it combines the tools you already have in service of committed clinical reasoning.
The funnel describes when to shift your questioning: it is a temporal sequence that follows the arc of the consultation. Each stage builds on the last: open questioning establishes the territory, focused questioning narrows it, closed questioning measures discriminating features, and the final stage combines closed and pertinent negative questions in service of a committed working diagnosis. The sequence is not arbitrary; it reflects the cognitive order in which clinical reasoning must proceed.
Four types of questions, each with a specific clinical job
: describes form, what each type does.
Four stages (the funnel) from invitation to investigation
: describes sequence, when to use each mode.
Both are necessary to master questioning strategy. The types are the
tools; the funnel is the map for when to reach for each one.
Effective clinical questioning is not a matter of question type but of sequence and cognitive intent: open questions establish the patient's narrative, focused and targeted questions test specific hypotheses generated from that narrative, and the transition between them, made deliberately and at the right moment, is where diagnostic reasoning becomes visible.