Clinical Decision Rules are Ruining Medicine

Clinical decision rules often look more objective than they are, and many were adopted before they proved they improved real-world care. For pulmonary embolism, pediatric trauma, chest pain, and C-spine clearance, the key question is simple: does the rule outperform usual practice and clinical judgment alone?

Why Decision Rules Mislead

• Four-step evidence pathway: A rule is not ready for routine use after derivation alone; it should survive validation, external validation, and an implementation study showing better care, not just cleaner statistics.
• Implementation study standard: The most important test is whether clinicians using the rule improve outcomes or resource use versus clinicians not using it, a step very few decision rules ever reach.
• False objectivity problem: Scores can look precise while depending on subjective inputs, so two clinicians may generate different numbers and still have the result treated as if it were hard fact.
• Zero-miss culture pressure: Many rules are built around near-100% sensitivity, a design goal that predictably drives overtesting and can turn rare misses into the dominant standard of care.
• Shared decision-making loss: Binary positive-or-negative rules often replace bedside judgment with cookbook medicine and make nuanced risk conversations harder. We get into that cultural shift in the episode.

Rules Worth Trusting More

• PERC and YEARS for PE: For pulmonary embolism, PERC and the YEARS algorithm are among the few tools backed by controlled implementation data showing they can improve practice rather than just classify risk.
• Canadian C-spine advantage: Canadian C-spine has both external validation and implementation evidence, with 99.4% sensitivity and a 9% absolute reduction in imaging in a cluster trial.
• Population matters most: Even strong rules can fail when moved into lower-risk settings, so trauma-room performance should not be assumed to hold in fast track or community populations.
• Learning from likelihood ratios: A practical use for rules is educational: inspect which variables carry the strongest likelihood shifts and compare them against your own gestalt. We walk through that approach in the chapter.

Common Rules That Underperform

• Ottawa ankle limits: Ottawa Ankle was validated, but outside Ottawa it increased x-ray use in a multicenter study, and a narrow rule-based exam can miss injuries elsewhere in the ankle.
• PECARN head tradeoff: PECARN head injury has near-100% sensitivity, but compared with clinical judgment it has much lower specificity, meaning more children get scanned while the same injuries are missed.
• PECARN abdominal uncertainty: The pediatric blunt abdominal injury rule has external validation but no implementation study, and its low specificity raises concern for more CT use in community settings.
• HEART score mismatch: HEART was built for simplicity rather than calibrated risk, creating odd equivalences such as age 50 carrying the same points as a positive troponin.
• Judgment versus checklist: The recurring test is whether a rule beats what experienced clinicians already do at the bedside; for several popular tools, the answer appears to be no.

References: 

1. Morgenstern, J. Clinical decision rules are ruining medicine. First10EM. February 2, 2023.