Clinical Decision Rules

Clinical decision rules (CDRs) are validated algorithms that combine clinical findings to guide specific decisions — typically whether to order a test, admit a patient, or initiate treatment. The Ottawa Ankle Rules, for example, specify exactly when ankle X-rays are needed after injury, reducing unnecessary imaging by 30-40% without missing fractures. The Wells Score estimates the probability of pulmonary embolism based on clinical signs. The HEART Score risk-stratifies chest pain patients. What makes CDRs powerful is their development process: they're derived from large datasets, validated in independent populations, and explicitly designed to outperform individual clinical judgment for the specific decision they address.

When to use it

When seeking to standardize common decisions to reduce variability and error, designing decision support tools for any domain, determining when algorithmic guidance should override individual judgment and vice versa, or reducing unnecessary resource use (testing, investigation) while maintaining safety.

How it can help

This model shows the power of converting expert judgment into explicit, validated algorithms while preserving the role of human override. In any domain, CDR-like tools can standardize common decisions, reduce variability, and free expert attention for genuinely complex cases. The model teaches that algorithmic decision support works best when: the decision is well-defined, there's sufficient data to derive and validate the algorithm, and there's a clear process for human override when the algorithm's assumptions don't fit. It also demonstrates that even imperfect algorithms can improve average outcomes by reducing the worst-case judgments.