Can We Perform Fewer Head CTs in the Emergency Department?

Perhaps, but crafting a good decision rule for computed tomography isn't easy.

The Centers for Medicare and Medicaid Services recently issued a utilization measure for cranial computed tomography (CT) scanning for nontrauma indications in the emergency department (ED), but the measure has met with strong criticism (JW Emerg Med Sep 21 2012).

To derive a decision rule for emergent head CT, researchers in Hamilton, Ontario, Canada, retrospectively examined records of 4000 adults who were scanned for nontrauma indications at three hospital EDs during 3 years. Patients with previously known intracranial pathology were excluded. An abnormality was defined as hemorrhage, acute or subacute infarction, mass lesion, or other findings that required intervention or follow-up.

In a 2000-patient derivation cohort, the prevalence of abnormal head CT was 14% (275 cases). Multivariable analysis yielded six independent predictors of abnormality: older age, nausea or vomiting, altered mental status, focal neurological deficit, cancer history, and derangement in coagulation. Findings were as follows:

Performing CT only in patients with any of the five non-age predictors would have detected 94% of abnormalities and lowered CT scanning by 30%.

Including all older patients (age, >70), plus younger patients with any of the other five predictors, would have increased sensitivity to 96%, and 21% of scans would have been avoided.

Although seizure was not an independent predictor, adding seizure as an indication to scan would have increased sensitivity to 99%, and 14% of scans would have been avoided.

With the last paradigm, the four "missed" cases would not have been clinically important. Findings were similar in a 2000-patient validation cohort.

CITATION(S):

Wang X and You JJ. Head CT for nontrauma patients in the emergency department: Clinical predictors of abnormal findings. Radiology 2013 Mar; 266:783.