Several considerations guided our data analysis. First, despite the fact that mean patient satisfaction scores were high, the staff believed that any process-improvement initiative should nonetheless directly include mean values. Alternative measures
should also be constructed, but mean values provide an intuitive (and, in the staff’s view, necessary) benchmark against which any new measures should be compared. Moreover, any additional measures used in the analysis should be small in number, simple to calculate, and easily interpreted to be clearly and concisely communicated to the institute’s staff. Otherwise, such measures become meaningless because the staff may be unwilling or unable to implement performance-improvement initiatives on the basis of those metrics (Neufelder,
Raisor, and Friesner 2002).