ED Utilization of Diagnostic Tests for CAP Across 36 PHIS Hospitals

TestUnadjusted Distribution Across Hospitals (% of Patients Receiving Test Across All Hospitals)Adjusted Distribution Across Hospitalsa,b (% of Patients Receiving Test Across All Hospitals)
Min25th %ileMedian75th %ileMaxMin25th %ileMedian75th %ileMax
Blood cultures11.318.627.
Coagulation studies0.180.450.650.831.
Viral studies1.48.513.322.140.11.6914.724.243.7
Inflammatory markers12.24.313.4471.74.67.523.582.4
Chest radiograph39.173.677.282.886.34071.275.778.287.3
Chest ultrasound00.
Chest CT scan0.10.30.490.61.4NAc
  • Max, maximum; Min, minimum; NA, not available.

  • a Adjusted for age, gender, race/ethnicity, season of presentation, year of presentation, and insurance status.

  • b P < .001 for between-hospital variation for all tests. P value represents the significance of hospital variation after adjustment by using the χ2 statistic for the random intercept of the mixed-effects logistic regression model.

  • c Mixed-effects regression would not converge for ultrasound and CT due to small cell size.