Integration of diagnostic information

Differential Diagnosis - Building The List

Building a comprehensive differential diagnosis (DDx) is a core clinical skill. Systematically consider possibilities to avoid premature closure and cognitive errors.

• Core Approaches:

  • Anatomic: What structures are in the area of the chief complaint? (e.g., chest pain → heart, lungs, esophagus, chest wall).
  • Physiologic: What is the underlying pathophysiology? (e.g., vascular, infectious, neoplastic, inflammatory).

• Frameworks for Breadth:

  • Use mnemonics to ensure a wide net is cast, especially when stuck.
  • 📌 VINDICATE:
    • Vascular
    • Infectious/Inflammatory
    • Neoplastic
    • Degenerative
    • Iatrogenic/Idiopathic
    • Congenital
    • Autoimmune
    • Traumatic
    • Endocrine/Metabolic

⭐ Always consider the most life-threatening conditions first (e.g., MI, PE, aortic dissection for chest pain). This "worst-first" approach is critical in emergency settings.

Diagnostic Info - Refining The List

• Use new clinical data (history, exam, labs) to dynamically update and narrow your differential diagnosis (DDx).
• The goal is to move a diagnosis across a treatment threshold (probability high enough to treat) or test threshold (probability low enough to stop testing).
• 📌 SpIN & SnOUT mnemonic:
  • A highly Specific test, when Positive, helps rule IN a disease.
  • A highly Snensitive test, when Negative, helps rule OUT a disease.
• Likelihood Ratios (LRs) quantify the power of a test to change probability.
  • LR+ > 10 is strong evidence to rule IN.
  • LR- < 0.1 is strong evidence to rule OUT.

⭐ Likelihood ratios are superior to sensitivity and specificity as they can be applied to an individual patient's pre-test probability, providing a direct estimate of post-test probability.

![Image showing Fagan's nomogram for calculating post-test probability from pre-test probability and likelihood ratio]

Integration - The Final Picture

• Clinical Synthesis: The art of weaving together patient history (Hx), physical exam (Px) findings, and all diagnostic data. The goal is to move from a broad differential (DDx) to a refined, ranked list or a single final diagnosis.
• Bayesian Inference in Practice:
  • Pre-test Probability: The initial clinical suspicion for a disease before new information is known.
  • Post-test Probability: The revised probability of disease after a diagnostic test result is integrated.
  • 📌 Fagan's Nomogram visually connects pre-test probability, likelihood ratio, and post-test probability.
  • Post-test odds = Pre-test odds × Likelihood Ratio
• Likelihood Ratios (LR):
  • LR+ > 10: Large increase in disease probability.
  • LR- < 0.1: Large decrease in disease probability.

⭐ A test with an LR+ of 10 increases the probability of disease by ~45% if the pre-test probability was 50%. Conversely, an LR- of 0.1 would decrease it to ~10%. This demonstrates the power of LRs in clinical decision-making.

High‑Yield Points - ⚡ Biggest Takeaways

• Bayesian inference is key: continuously update the probability of disease using new clinical data and test results.
• Always establish a pre-test probability from history, exam findings, and disease prevalence before ordering tests.
• Use Likelihood Ratios (LRs) to interpret tests. An LR+ >10 strongly rules in disease; an LR- <0.1 strongly rules it out.
• Beware of cognitive biases like anchoring on initial findings or premature closure, which cause diagnostic errors.
• A negative test does not exclude disease if pre-test probability was high.