Diagnosis
Core content
Diagnosis questions reward structured clinical reasoning. Build a prioritized differential from the history and exam, then test hypotheses. Know test characteristics: sensitivity is the property of ruling out (a highly sensitive test that is negative makes disease unlikely), specificity rules in, and positive predictive value depends on pretest probability. Likelihood ratios combine both and shift probability. Order tests only when the result will change management.
Know common diagnostic thresholds: diabetes with A1c ≥6.5%, fasting glucose ≥126 mg/dL, 2-hour OGTT ≥200 mg/dL, or random glucose ≥200 mg/dL with symptoms; hypertension at ≥130/80 mmHg confirmed with proper technique on repeat occasions; CKD as eGFR <60 mL/min/1.73 m² or albuminuria for at least three months. Distinguish screening tests from confirmatory tests, and interpret labs in context - anemia, hemoglobinopathies, pregnancy, and advanced kidney disease all make A1c less reliable.
Traps
Anchoring on an initial impression, premature closure, and confirmation bias are the named reasoning errors exams test. Other traps: ordering imaging before estimating pretest probability, treating a positive screening result as a diagnosis without confirmation, and ignoring age-related change in test performance. Spurious results matter too - hemolyzed samples causing false hyperkalemia, biotin interference with immunoassays.
Rule of thumb
Choose the test that best separates your leading diagnosis from the most dangerous alternative. A high pretest probability warrants the definitive test directly; a low pretest probability warrants the sensitive exclusion test.
Worked example
A 58-year-old with unilateral leg swelling has a Wells score in the high range. The correct next step is duplex ultrasound, not D-dimer. D-dimer is useful only when pretest probability is low or intermediate, because a positive result there still requires imaging to confirm.