Use of patient and hospital variables in interpreting patient satisfaction data for performance improvement purposes.

Satisfaction scores of 349 patients being discharged from a state psychiatric hospital were examined in relation to available norms for the instrument used and selected patient and hospital variables. Mean item scores fell within the less-than-satisfied category on both total and factor scores. Regression analyses indicated minimal effects of patient attributes. Two hospital factors (restraint rate on patient's unit and accessibility of psychosocial groups) significantly predicted satisfaction, with the former having an unexpected positive relationship to satisfaction. Clinicians were able to use the survey data to improve care, but patients' tendency toward undifferentiated positive or negative responding hindered the prioritizing of change efforts.