Elsie Item-Writing Academy
About 45 minutes · Academy module: Flaw Clinic: Step 2 CK — Flaws That Teach the Wrong Medicine
Unofficial faculty-development resource. Not affiliated with, endorsed by, or representing the AAMC, NBME, USMLE, LCME, or NRMP. Supports faculty-development documentation.
Step 2 CK items test clinical decision-making — next steps, management priorities, and diagnostic reasoning under time pressure. The flaw taxonomy is the same as at Step 1, but the stakes of each flaw change: a flawed management item can teach the wrong clinical lesson, because students study the key as a practice guideline. Restraint and precision matter more here than anywhere.
Window dressing. The CK vignette tempts authors to write a "realistic" patient — job, family, hobbies, a full review of systems. Realism is not the goal; discrimination is. Every detail should move the needle between at least two options. A blood pressure of 186/104 matters if the decision is thrombolysis; the patient's occupation does not. Long vignettes also create an inadvertent cueing channel: fatigued readers skim, and skimmers anchor on the most vivid (not the most diagnostic) detail.
Clueing between stem and options. At the CK level, clueing often takes the form of management buzzwords echoed from stem to key — the stem says "reassuring tracing" and the key says "continue expectant management with reassuring monitoring." Students learn to match tone rather than weigh risk. Keep the stem's language neutral and the options' language parallel.
Implausible distractors. Management items are especially vulnerable: options like "discharge home" for a clearly ill patient, or "diagnostic laparoscopy" for a medical emergency, are dead on arrival. They do more than waste space — they signal to the student which options the author considered serious, narrowing the real choice set. Every distractor should represent a management path a real clinician might at least consider.
Two defensible answers. Management guidelines have gray zones, and gray zones breed disputed keys. If two options both reflect defensible practice (e.g., two reasonable next steps), the item is unscoreable no matter how the key is set. Resolve it by tightening the clinical picture — a vital sign, a gestational age, a lab threshold — until guidelines point to one option. Both independent cold solves must converge on the key with no defensible alternative.
Unfocused stems. "What is the most appropriate management?" after an undifferentiated vignette is only focused if the vignette converges on one decision. Ask the single decision the vignette earns: "What is the most appropriate next step?" is fine when the stem presents one fork in the road — not three.
Absolute terms, all/none of the above, negative stems, grammatical cueing. The standing rules apply unchanged: no absolutes, no "all of the above"/"none of the above," emphasized negation only (EXCEPT/NOT in capitals) when a negative stem is unavoidable, and every option read back into the stem to catch agreement giveaways.
---
Flawed item (intended key: A)
Mr. H is a 68-year-old retired high-school history teacher who volunteers at the public library twice a week and enjoys gardening with his wife of 40 years. He has hypertension and hyperlipidemia. This morning, while reading the newspaper at the kitchen table, he suddenly developed right-sided weakness and difficulty speaking; his wife noticed the symptoms began about 90 minutes ago and called emergency services. In the emergency department, noncontrast CT of the head shows no hemorrhage. His blood pressure is 186/104 mm Hg. He is alert. What is the most appropriate next step in management?
A) Administer intravenous alteplase B) Lower blood pressure aggressively before any reperfusion therapy C) Obtain MRI of the brain before initiating therapy D) Start aspirin 325 mg immediately E) Admit for observation and repeat examination
Flaw taxonomy label: Window dressing.
Correction: The teaching career, the library volunteering, the gardening, the 40-year marriage, the newspaper — none of it changes the decision. What matters is all present in three facts: focal deficit within the thrombolysis window, no hemorrhage on CT, blood pressure below the treatment threshold. The decorative biography inflates reading time and gives skimmers vivid but non-diagnostic details to anchor on. The rewrite keeps every discriminating fact and cuts everything else.
Model rewrite — cold-solve verified: 2/2 (solver 1: A, solver 2: A)
A 68-year-old man with hypertension and hyperlipidemia presents with sudden right-sided weakness and aphasia that began 90 minutes ago. Noncontrast CT of the head shows no hemorrhage. Blood pressure is 186/104 mm Hg. What is the most appropriate next step in management?
A) Administer intravenous alteplase B) Lower blood pressure aggressively before any reperfusion therapy C) Obtain MRI of the brain before initiating therapy D) Start aspirin 325 mg immediately E) Admit for observation and repeat examination
Key: A
Flawed item (intended key: D)
A 45-year-old man reports fatigue, arthralgias, and new-onset diabetes. On examination he has bronze discoloration of the skin. Laboratory studies show elevated ferritin and transferrin saturation of 78%. Liver biopsy demonstrates iron deposition in hepatocytes. What is the most likely diagnosis?
A) Alcoholic liver disease B) Wilson disease C) Autoimmune hepatitis D) Hereditary hemochromatosis ("bronze diabetes") E) Alpha-1 antitrypsin deficiency
Flaw taxonomy label: Clueing between stem and options.
Correction: The stem hands the solver the word "bronze," and only one option echoes it back ("bronze diabetes"). A student can match the vocabulary without knowing anything about iron studies — the item becomes a word search. The clinical content (transferrin saturation 78%, hepatocellular iron, the diabetes–arthralgia–fatigue triad) is sufficient on its own; the rewrite removes the echo from both sides so the diagnosis must be reasoned from the iron studies, not matched from the adjectives.
Model rewrite — cold-solve verified: 2/2 (solver 1: D, solver 2: D)
A 45-year-old man reports fatigue, arthralgias, and new-onset diabetes. Laboratory studies show elevated ferritin and transferrin saturation of 78%. Liver biopsy demonstrates iron deposition in hepatocytes. What is the most likely diagnosis?
A) Alcoholic liver disease B) Wilson disease C) Autoimmune hepatitis D) Hereditary hemochromatosis E) Alpha-1 antitrypsin deficiency
Key: D
Flawed item (intended key: B)
A 32-year-old woman at 28 weeks gestation presents with severe headache, visual changes, and blood pressure 168/112 mm Hg. Urinalysis shows 2+ protein. What is the most appropriate management?
A) Oral labetalol with discharge home B) Delivery of the fetus C) Reassurance only, with routine follow-up D) Home blood pressure monitoring E) Bed rest with daily nonstress testing
Flaw taxonomy label: Implausible distractors.
Correction: Options C and D are not management plans a thinking clinician would consider for severe preeclampsia — they are straw men. Their presence does active harm: it signals which options the author took seriously, quietly narrowing the real choice to A, B, and E. The rewrite replaces the straw men with distractors that represent genuine clinical errors — magnesium as definitive therapy (it is seizure prophylaxis, not treatment), expectant management to 37 weeks (wrong for severe features), outpatient-style temporizing — so every option must be evaluated on content. The added laboratory findings (thrombocytopenia, elevated transaminases) further anchor the severity and remove the gray zone that made the original key debatable.
Model rewrite — cold-solve verified: 2/2 (solver 1: B, solver 2: B)
A 32-year-old woman at 28 weeks gestation presents with severe headache, visual changes, and blood pressure 168/112 mm Hg. Urinalysis shows 2+ protein. Laboratory studies show platelets 78,000 per microliter, AST 210 U/L, and LDH 640 U/L. Fetal heart tracing is reassuring. What is the most appropriate management?
A) Oral labetalol with discharge home B) Delivery of the fetus C) Magnesium sulfate alone as definitive therapy D) Expectant management until 37 weeks E) Bed rest with daily nonstress testing
Key: B
---
Recording it adds the module to your Academy completion record on this device — module, track, level, and date, ready to download from your account page for faculty-development documentation.