Table 3. Association between Clinician Type, Location, Shift and Probability of an Intercepted Drug Name Confusion Error.

Variablea	With Fluticasone		Without Fluticasone
	Odds Ratio	Confidence Interval	Odds Ratio	Confidence Interval
Clinician Type
Resident	-	-	-	-
Attending	0.35b	0.23–0.54	0.69	0.30–1.56
Nurse	0.53c	0.28–0.996	0.69	0.21–2.26
Location
Inpatient	-	-	-	-
Ambulatory	3.54b	2.51–5.00	0.51c	0.27–0.95
ED	0.27c	0.11–0.68	0.29c	0.11–0.72
OR	0.27	0.04–1.98	0.29	0.04–2.09
Shift
Day	-	-	-	-
Evening	0.80	0.55–1.16	0.47c	0.23–.95
Overnight	0.74	0.44–1.27	1.05	0.52–2.12

^aResident physician, inpatient location and day shift were used as reference categories. Testing global null hypothesis for model with fluticasone, −2 log likelihood  = 2563.8, chi-square  = 94.9, p<.0001. For model without fluticasone, −2 log likelihood  = 1245.9, chi-square  = 24.6, p<.0001.

^bp<0.001.

^cp<0.05.