Building a Diagnosis From Physiological Evidence

Let me walk you through how a clinician actually thinks — because diagnosis is not a guess. It is an argument built from evidence. When a patient arrives, you collect four categories of data. First, symptoms: what the patient reports, such as chest pain, fatigue, or shortness of breath. Symptoms are subjective — they depend on the patient's perception — but they tell you where to look. Second, vital signs: objective, measurable signals produced by a monitoring device, like heart rate, blood pressure, respiratory rate, temperature, and oxygen saturation (SpO2, a measure of blood oxygen saturation). These tell you how the body's control systems are currently performing. Third, physical examination signs: findings the clinician directly observes or measures during the exam, such as skin that appears pale and clammy, swelling of a limb, or abnormal breath sounds heard through a stethoscope. These are objective but detected by the clinician's senses rather than a device readout. Fourth, diagnostic tests: blood panels, imaging, urinalysis, ECGs. These drill into specific tissues or biochemical pathways that symptoms, vital signs, and examination alone cannot fully reveal. The clinician's job is to hold all four categories at once and ask: what single explanation would produce this exact pattern? Here is the key move — the differential diagnosis. You list every condition that could explain the findings, then use additional data to rule each one out or in. Suppose a patient has a heart rate of 118 beats per minute (tachycardia), a respiratory rate of 24 breaths per minute, a temperature of 37.4 °C (essentially normal), and reports sharp chest pain that worsens when breathing in (called pleuritic chest pain). Each abnormality alone has many causes. Tachycardia can come from pain, dehydration, anemia, infection, or cardiac arrhythmia. But when you combine elevated heart rate, elevated respiratory rate, normal temperature, and pleuritic chest pain, the pattern points strongly toward pulmonary embolism (a blood clot in the lungs) or a non-infectious lung problem — fever's absence makes an infectious cause less likely. Next you order targeted tests. A chest X-ray, a D-dimer blood test, and an SpO2 reading give you the next layer of evidence. D-dimer is a fibrin degradation product released when blood clots break down, but it is non-specific — it is also elevated in infection, cancer, pregnancy, surgery, and trauma. That means a positive D-dimer alone does not confirm a clot; it only tells you a clot is possible. In this patient, however, the pre-test probability of pulmonary embolism is already moderately high based on the clinical picture (tachycardia, pleuritic pain, normal temperature), so an elevated D-dimer meaningfully raises that probability further. If SpO2 is also 91% (low) and the chest X-ray is clear, the pattern now strongly favors pulmonary embolism, because pneumonia typically shows a visible infiltrate on X-ray and would more often cause fever. This is Bayesian reasoning in medicine: each new piece of evidence updates your probability estimate. You are not looking for certainty — you are building the most defensible explanation from available data, then acting on it while remaining open to revision. A critical habit: never anchor too early. Anchoring bias happens when a clinician fixates on the first plausible diagnosis and stops weighing new contradicting evidence. Good clinical reasoning keeps the differential open until the evidence truly converges.