How do you decide whether to empirically cover Pseudomonas for pneumonia in hospitalized patients?