Table 3.

Multivariable Logistic Regression of Burnout.

Variable	Odds ratio	95% confidence interval
Age	0.96	0.95-0.98
Inadequate administrative time	1.9	1.4-2.6
Inadequate teaching time	1.8	1.3-2.5
Uncertainty of health care reform	2.0	1.4-3.1
Difficult pharmacist colleagues	2.1	1.5-2.9
Contributions underappreciated	2.2	1.6-3.1
Too many nonclinical duties	2.3	1.7-3.2

Note. All other intangible perception variables with P < .2 from univariate analysis did not predict outcome.

Hosmer-Lemeshow goodness of fit = 0.1; area under the receiver operator characteristic curve (95% confidence interval) = 0.79 (0.76-0.82).