ABSTRACT
RATIONALE, AIMS, AND OBJECTIVE: Bariatric surgery is an effective procedure for morbidly obese patients when all else fails. The purpose of this study was to compare the hospital length of stay (LOS) for two surgical procedures, laparoscopic adjustable gastric banding (LAGB) and laparoscopic sleeve gastrectomy (LSG). METHODS: This study was a retrospective cross-sectional analysis of the Nationwide Inpatient Sample (NIS) from 2009 to 2014. Patients who received bariatric surgery as indicated by International Classification of Diseases, Ninth Revision (ICD-9) procedure codes were selected (N = 4001). Cases were limited to uncomplicated diabetic patients. Differences in the odds of long vs short (2< and ≥2) stay for a patient receiving LSG were compared with LAGB while adjusting for hospital volume, hospital size, patient age, gender, ethnicity, season, and year using logistic regression analysis. RESULTS: The odds for LSG (odds ratio [OR] = 0.100, 0.066-0.150, P < 0.001) patients for long LOS are lower when compared with LAGB. In the stratified logistic regression model, both male (OR = 0.157, 0.074-0.333, P < 0.001) and female (OR = 0.077, 0.046-0.127, P < 0.001) had reduced odds of extended LOS for LSG. Discharged patients in the year 2012 (OR = 0.660, 0.536-0.813, P < 0.001) had decreased odds of having a longer LOS when compared with the year 2014. Both government, nonfederal (OR = 0.452, 0.251-0.816, P = 0.008), and private investor-owned (OR = 0.421, 0.244-0.726, P < 0.001) patients had similar odds for long duration of stay when compared with government or private. Urban non-teaching (OR = 1.954, 1.653-2.310, P < 0.001) patients had higher odds for long LOS in comparison with urban teaching. New England patients' (OR = 0.365, 0.232-0.576, P < 0.001) odds for extended LOS were lower when compared with pacific. Both patients who received care in low (OR = 1.330, 1.109-1.595, P = 0.002) and medium (OR = 1.639, 1.130-2.377, P = 0.009) volume hospital had increased odds for long duration of stay. Female patients in the stratified logistic regression model with high (OR = 1.330, 1.109-1.595, P < 0.002) volume had elevated odds of extended LOS when compared with very low volume hospital. CONCLUSION: Among the uncomplicated diabetic patients, LSG provides a substantially low odds of extended LOS after adjusting for covariates when compared with LAGB. The finding of the relative reduction in LOS for LSG suggests opportunities for improvement both for cost reduction for third party insurance payers and greater efficacy and outcomes for patients.
