Over the last several years, interest in benchmarking health services' quality—particularly patient satisfaction (PS)—across organizations has increased. Comparing patient experiences of care across hospitals requires risk adjustment to control for important differences in patient case‐mix and provider characteristics. This study investigates the individual‐level and organizational‐level determinants of PS with public hospitals by applying hierarchical models. The analysis focuses on the effect of hospital characteristics, such as self‐discharges, on overall evaluations and on across hospital variation in scores. Sociodemographics, admission mode, place of residence, hospitalization ward and continuity of care were statistically significant predictors of inpatient satisfaction. Interestingly, it was observed that hospitals with a higher percentage of Patients Leaving Against Medical Advice (PLAMA) received lower scores. The latter result suggests that the percentage of PLAMA may provide a useful measure of a hospital's inability to meet patient needs and a proxy indicator of PS with hospital care. © 2013 The Authors. International Journal of Health Planning and Management published by John Wiley & Sons, Ltd.