3 Developing a Differential Diagnosis

What a differential diagnosis does

A differential diagnosis is not simply a long list of diseases. It is a prioritized set of plausible explanations for a patient’s findings, with each candidate assessed for how well it fits, whether its mechanism is plausible, and how harmful it would be to miss.

Diagnosis is iterative: history, examination, and testing refine a working explanation over time. The ranking should remain open to revision as evidence changes.

Define the presentation first

Begin with the patient’s main concern and describe it precisely. Relevant details include onset, time course, location, quality, severity, associated symptoms, exposures, medical history, medications, and objective findings. Use these details to form a concise problem representation that preserves the features most useful for distinguishing causes.

For example, “shortness of breath” is broad. “Sudden breathlessness after recent surgery, with low oxygen saturation and no fever” is more discriminating. Age, comorbidities, exposures, and the clinical setting can all change which explanations are plausible.

Generate explanations from patterns and mechanisms

Pattern recognition compares the presentation with familiar illness scripts: organized knowledge that links predisposing factors, disease mechanism, expected findings, and course. This can efficiently generate an initial hypothesis, but relying on a familiar match without checking alternatives can cause premature closure.

Mechanism-based thinking broadens the search by asking what could produce the observed problem. For breathlessness, possible mechanisms include impaired airflow, infection or inflammation, fluid in the lungs, reduced blood flow through the lungs, or impaired oxygen-carrying capacity. Considering mechanisms can bring forward diagnoses that do not immediately match a classic pattern.

Rank candidates by likelihood and urgency

Keep probability separate from consequence. A common condition that fits well may be the most likely explanation, while a less likely but dangerous condition may still need prompt exclusion. Consider prevalence in the patient’s context, risk factors, how well the findings fit, and the consequences of delay. Urgent red flags may require action before diagnostic certainty is achieved.

A practical differential can distinguish among:

• Most likely: the candidate with the best overall fit to the available evidence.

• Important alternatives: plausible conditions that would change evaluation or care.

• Must-not-miss diagnoses: time-sensitive or high-harm possibilities, even if less likely.

Seek evidence that separates candidates

For each leading candidate, ask: “If this explanation were true, what else would I expect to find?” Identify evidence that would raise or lower its probability. Prefer questions, examination maneuvers, or tests that distinguish competing explanations over information gathering without a clear purpose.

A finding is most useful when it differs meaningfully across candidates. No single absent symptom or normal test necessarily excludes a diagnosis.

In acute breathlessness, recent surgery and sudden onset may increase concern for pulmonary embolism; fever and focal lung findings may support infection; wheezing may support airflow obstruction; and orthopnea with peripheral edema may support heart failure. These clues guide further assessment, but they must be interpreted together with the patient’s overall risk and the limits of each finding.

Update the differential with new evidence

As information arrives, compare each candidate’s expected pattern with the actual findings. Promote explanations that account for the evidence, lower those that fit poorly, and add alternatives if the presentation changes.

If a test result conflicts with the working diagnosis, reconsider the diagnosis, the test’s limitations, and whether more than one process may be present. Structured analysis is especially valuable in complex cases or when a pattern match could obscure a mimic.