Background: The excess health care costs caused by obesity are a concern in many countries, yet little is known about the additional resources required to treat obese patients in hospitals.

Objective: To estimate differences in hospital resource use, measured by length of stay, between severely obese and other patients, conditioning on a range of patient and hospital characteristics.

Research Design: Administrative patient-level hospital data for 122 Australian public hospitals over the financial year 2005/06 (Victorian Admitted Episodes Data).

Subjects: Episodes (435,147) for patients above 17 years of age and with a stay of one night or more.

Measures: Quantile Regression analysis is used to generate 19 estimates of the difference between severely obese and other patients across the whole range of length of stay, from very short to very long staying patients. Separate estimates for 17 hospital specialties and for medically and surgically treated patients are generated.

Results: There are significant differences in average length of stay for almost all specialties. For some, differences are less than 1 day, but for others, severely obese patients stay up to 4 days longer. For a number of specialties, obese patients have significantly shorter length of stay. Overall, medically managed obese patients stay longer, whereas surgically treated patients stay shorter than other patients. Differences tend to increase with length of stay.

Conclusions: Differences in length of stay may arise because severely obese patients are medically more complex. The observed shorter stays for obese patients in some specialties may result from their observed greater likelihood of being transferred to another hospital.