DIAGNOSTIC REASONING UNDER UNCERTAINTY
ALL YEARS | PAGE 78A - PROBLEM REPRESENTATION -> PRIOR PROBABILITY -> DIFFERENTIAL -> DISCRIMINATORS
A. DIAGNOSTIC REASONING ARCHITECTURE
| Step | Reasoning job | Output / checkpoint |
|---|---|---|
| 1. Define the clinical | Setting, acuity, time course, dominant syndrome and immediate instability. | One sentence describing what kind of problem this is before naming a disease. |
| frame 2. Build the | Compress age/context + tempo + key positives/negatives + decisive modifiers. | A short statement that preserves the features that actually separate diagnoses. |
| problem repr esentation | ||
| 3. Set prior probability | Use prevalence, exposures, comorbidities and context as the starting point. | A ranked starting distribution - not a verdict. |
| 4. Build a ranked | Keep three lanes: most likely, dangerous alternative, plausible competing explanation. | A short differential organized by probability and consequence. |
| differential 5. Seek disc | Ask which history, exam, test or trajectory feature would move the leading | Decision-changing evidence rather than indiscriminate data collection. |
| riminators 6. Choose | options apart. Act when the expected benefit, risk and setting justify it; test only when the result | A defensible next step tied to the current uncertainty. |
| the next action | can change the plan. | |
| 7. Reassess | Update after new data, treatment response, time, consultation or follow-up. | Revised probability + revised plan; diagnostic reasoning remains open until the case is stable. |
TOP 10 DIAGNOSTIC-REASONING CONCEPT - 1: PROBLEM REPRESENTATION COMPRESSES WITHOUT ERASING DANGER
| DEEP EXPLANATION | WHY STUDENTS MISS IT | EASIEST CLUE |
|---|---|---|
| A strong representation removes trivia while preserving tempo, severity, context and the few clues that change probability. It is | Students either copy the entire stem or compress so aggressively that the dangerous modifier disappears. | If two very different diagnoses would fit your one-line summary equally well, the representation is too vague. |
| not merely a shorter history. | ||
| TOP 10 DIAGNOSTIC-REASONING CONCEPT - 2: BASE RATES SET THE STARTING LINE - THEY DO NOT CANCEL RED FLAGS | ||
| DEEP EXPLANATION | WHY STUDENTS MISS IT | EASIEST CLUE |
| Prevalence and context influence the prior probability. New evidence then moves that probability. Common diseases | Learners either ignore prevalence completely or use “common things are common” as permission to dismiss discordant | Ask: What did I believe before this clue, and how much should this clue move me? |
| remain common, but strong contradictory evidence should force updating. | findings. | |
| TOP 10 DIAGNOSTIC-REASONING CONCEPT - 3: MOST LIKELY AND MUST-NOT-MISS ARE DIFFERENT LANES | ||
| DEEP EXPLANATION | WHY STUDENTS MISS IT | EASIEST CLUE |
| The most probable diagnosis and the diagnosis with the highest consequence of delay may be different. Safe reasoning | Single-best-answer training encourages one winner, so learners may stop considering dangerous alternatives too early. | Before closing the case, name one diagnosis that is less likely but unsafe to ignore. |
| keeps both visible until evidence or risk allows one to be closed. 2026/27 diagnostic-reasoning expanded edition - educational use; follow school requirements, supervising clinicians and current clinical guidance for patient care |