National study of the effect of patient and hospital characteristics on bariatric surgery outcomes.

The influence of patient and hospital demographics on gastric bypass (GB) outcomes is unknown. We analyzed year 2000 data from the Nationwide Inpatient Sample database for all GB patients. In 2000, 5876 GB were performed in the 137 sample hospitals (M:F, 14%:86%). Length of stay (LOS, days), charges, comorbidities, and morbidity were higher for those aged >60 years compared to < 40 years. LOS, charges, comorbidities, morbidity, and mortality were highest in males. LOS was longest in African Americans compared to Caucasians and Hispanics. Charges and comorbidities were greatest in African Americans and Hispanics compared to Caucasians. Medicare and Medicaid-insured patients have higher LOS, charges, comorbidities, morbidity, and mortality compared to privately insured and self-pay patients. Lower income patients have higher LOS and total charges. Nonteaching hospitals have an increased LOS and charges and treat patients with more comorbidities compared to teaching hospitals. LOS, charges, and morbidity are directly proportional to hospital size. Urban hospitals have lower LOS and higher charges compared to rural hospitals. As hospital GB volume increases, LOS, charges, and morbidity decrease with no mortality effect. After controlling for all other covariates, male gender, increased age, and large hospital size were predictors of increased morbidity. Having had a complication predicted increased mortality, while female gender had a protective effect. Patient income, insurance status, and race did not play a role in morbidity or mortality. Neither academic, teaching status of the hospital or hospital gastric bypass volume influenced patient outcomes. Patient and hospital demographics do affect the outcomes of patients undergoing GB. Increasing age, male gender, and surgery performed in large hospitals are predictors of morbidity. Male gender and postoperative complications predict increased mortality. Neither comorbidities, race, payer, income, hospital academic status, location, nor hospital volume affect the outcome after GB.