Obesity and morbid obesity represent one of the major public health problems in the United States (U.S.) that affects nearly one-third of the adult American population. Gastric bypass (GB) is a complex operation, performed in a high-risk morbidly obese population, requiring well-trained surgeons and well-equipped hospital facilities to ensure optimal surgical outcomes. The volume-outcomes relationship is well-established for providers (both surgeons and hospitals) performing GB procedures. However, the findings of improved outcomes after GB for high volume providers have been attributed only to the high volume of GB and not the volume of other non-gastric bypass (non-GB) procedures. The studies in this dissertation were undertaken to examine the effect of provider's (general surgeon and hospital) non-GB complex (non-GBC) and non-complex (non-GBNC) volume on in-hospital complications and length of stay (LOS) for patients undergoing GB. The population-based studies used a combination of various existing retrospective data to address the research objectives. The datasets used include: a two-year (2003-2004) Florida hospital inpatient discharge data as the main analytic dataset, the 2003-2005 work Relative Value Units (RVU) data (available from the Physician Fee Schedule from the Centers of Medicare and Medicaid, to segment the provider's non-GB case load into non-GBC and non-GBNC procedures performed by a provider per year), 2005 Florida hospital characteristics file, 2005 Florida surgeon characteristics file, and 2004 Area Resource File data. Separate generalized estimating equation (GEE) regression models, adjusting standard errors for the non-nested surgeon and hospital cluster effect, were constructed for each outcome: composite complications (one or more complications), technical complications (including unexpected reoperations, splenic injury, hemorrhage, anastomotic leaks, small bowel obstructions, and wound), systemic complications (including pulmonary, cardiac, thromboembolic, genitourinary tract, and postoperative shock), and LOS. Covariates included were patient characteristics, year, surgeon GB volume, and hospital characteristics.In adjusted analyses, the gastric bypass patients operated by general surgeons with high non-GBNC volume (>142 procedures/year) had 70% and 88% higher likelihood of composite and systemic complications, respectively. In contrast, those operated at hospitals with high non-GBNC volume (>6,478 procedures/year) had 49% and 40% lower likelihood of composite and technical complications, respectively. There was no clear association between providers' high non-GBC volume and adverse outcomes. Furthermore, patients operated by general surgeons with high GB volume (>50 GBs/year) had 27% and 41% lower likelihood of composite and systemic complications, respectively. However, those operated at hospital's with high GB volume (>125 GBs/year) had 30% lower likelihood of technical complications. The study findings suggest that while provider GB volume matters for in-hospital complications, the complexity of overall surgical load also matters for general surgeons but the overall scale matters for hospitals to deliver better in-hospital outcomes for GB. In particular, the outcomes may improve if GB patients avoided general surgeons with a high volume of non-complex procedures and if GB patients avoided hospitals with low total volume.
