Physician Decision-Making in Acute Ischemic Stroke with and without a Guideline-Encoded Clinical Decision Support System

LUCID access in acute ischemic stroke vignettes produced near-uniform full guideline concordance—99.3% (150/151) versus 71.0% (142/200) without it.

• Why it matters: Applying extensive, frequently revised stroke guidelines at the point of care is difficult, and it is unclear whether guideline-encoded decision support improves concordance and reduces errors.
• What they did: In a pilot vignette study, 15 clinicians answered 8 stroke vignettes (24 questions) either alone or with LUCID, yielding 29 decision points per question bundle and cluster-bootstrap risk differences for concordance.
• The result: Concordance was 100% versus 67.9% for Class 1 and 0/66 versus 10/64 commission errors for Class 3: Harm (P<0.001), with workload tending lower in the LUCID arm.