AOBIM Internal Medicine common mistakes
AOBIM Internal Medicine · Understand the traps and how to avoid them.
Committing to an answer after reading only the stem's opening demographics and the first abnormal lab, without reading the full vignette or the final question line (e.g., 'which is the most likely cause' vs 'which is the next best step').
Under time pressure the brain pattern-matches the first familiar cue (age, sex, one lab value) and 'locks in' before the actual task is known; many stems deliberately include a late detail that changes the question from diagnosis to management.
What to do instead
Read the last line of the stem first, then scan the vignette once for the discriminating findings, then re-read the question line before selecting. Train by covering the options and writing the task type (diagnose/next step/most likely organism) in one word before looking at choices.
Treating a single abnormal value as diagnostic without asking whether it is isolated, transient, or explained by a competing condition (e.g., calling a mildly elevated TSH hypothyroidism in an inpatient with nonthyroidal illness, or diagnosing diabetes from one random glucose).
Board stems supply abnormal numbers that feel like the answer, and candidates underweight the stem's context (acute illness, medications, pregnancy, timing of draw).
What to do instead
For every abnormal value, state the confirmatory or contextual test that would be required in practice (repeat, fasting, free T4, HbA1c, aldosterone/renin pairing). If the stem does not give it, the value is a distractor, not the diagnosis.
Choosing the most aggressive or most definitive intervention (emergent surgery, thrombolysis, broadest antibiotic) when the stem describes a stable patient, or choosing watchful waiting when red-flag features are present.
Candidates conflate 'serious disease' with 'emergent management' and are drawn to the option that sounds most decisive; conversely, fear of over-treatment leads to under-triage of unstable vitals buried in the stem.
What to do instead
Before answering management questions, extract a stability snapshot (BP, HR, mental status, oxygenation, lactate if given) and let it gate the answer: unstable = resuscitate/stabilize first; stable = confirm and treat in stepwise fashion. Practice classifying 20 management stems as stable vs unstable before reading options.
Missing the 'next best step' distinction between diagnosis and treatment, selecting a treatment when the question asks for the diagnostic test (or vice versa).
The correct diagnosis is often obvious from the stem, so the candidate's attention jumps to therapy and they answer the question they expected rather than the one asked.
What to do instead
Adopt a two-pass habit: pass one identifies the syndrome; pass two re-reads the interrogative and eliminates any option that does not match the requested category (test vs therapy vs prognosis vs prevention).
Applying adult general-medicine defaults to special populations (pregnancy, elderly/frail, CKD, transplant, neutropenia) without adjusting drug choice, dosing, or thresholds.
Most studied material is anchored to the general adult; population-specific contraindications (ACE inhibitors in pregnancy, fluoroquinolones in pregnancy, tight glycemic targets in frail elderly) are recalled less fluently and lose to the familiar answer.
What to do instead
When the stem names a special population, mentally list three hard constraints before reading options (drugs to avoid, target ranges, altered diagnostic thresholds) and eliminate options that violate them. Build a one-page per-population constraint card and review weekly.
Anchoring on the first plausible diagnosis and failing to reopen it when a later stem detail is discordant (e.g., sticking with community-acquired pneumonia when the stem later mentions recent hospitalization and MRSA risk).
Early closure is cognitively efficient, and once a diagnosis is named internally, contradictory data are discounted rather than integrated.
What to do instead
Use a 'disconfirming detail' check: after selecting an answer, scan the stem for one finding that argues against it. If present and unexplained, restart the differential. Deliberately practice stems with a late twist.
Answering osteopathic-principle or structural questions with allopathic-only reasoning, ignoring somatic dysfunction, viscerosomatic reflexes, or OMT-relevant anatomy when the stem frames the patient structurally.
Candidates prepare primarily from allopathic question sources and treat OPP/OMT content as low-yield, so structural cues in the stem go unrecognized and the structural option is dismissed.
What to do instead
Reserve scheduled review of somatic dysfunction by region (TART findings, sympathetic levels by organ) and practice identifying when a stem's musculoskeletal or autonomic findings point to a structural explanation or an OMT indication versus a medical emergency requiring referral.
In the longitudinal OCC format, answering rapidly to 'clear the queue' without using available references or without reviewing the rationale after each item, treating it as a speed test rather than a learning assessment.
The open-resource, spaced format invites rushing and over-reliance on lookups for items that should be answered from knowledge; conversely, over-searching consumes time and erodes judgment.
What to do instead
Set a per-item time budget, answer from knowledge first, then verify only genuinely uncertain facts (dosing, guideline thresholds). Review every rationale after the item and log misses by domain to drive the next week's reading.
Confusing guideline-concordant care with the 'textbook classic' presentation, choosing an answer that reflects an outdated or superseded recommendation (e.g., old blood pressure or lipid targets, older anticoagulation thresholds).
Study habits formed years ago persist, and candidates do not systematically refresh guideline updates, so the remembered answer is the one that was correct at training.
What to do instead
Maintain a living 'changed recommendations' list (targets, first-line agents, screening ages/intervals) refreshed each cycle, and when two options both seem reasonable, favor the one matching the most recent major society guideline named in the stem's context.
Mismanaging time and item order - spending excessive minutes on a few hard items and then rushing the remainder, or changing correct answers on review without a specific reason.
Hard items create sunk-cost pressure, and answer-changing is driven by anxiety rather than new information; both degrade performance independent of knowledge.
What to do instead
Adopt a hard stop per item, flag and move on, and only change an answer when a concrete stem detail or recalled fact justifies it. Rehearse this timing under timed blocks that mirror the real exam's item count.