D. F. Butler1, J. Anandam1, B. Williams1, S. C. Oltmann1 1University Of Texas Southwestern Medical Center,Colorectal Surgery,Dallas, TX, USA
Introduction:
The National Surgical Quality Improvement Program (NSQIP) defines a prolonged length of stay (LOS) after colectomy as greater than 6 days, and uses this as a marker for quality of care. Causes for prolonged LOS can vary from medical to social, and understandably can be influenced by patient access to resources for post-hospital care. Protocol driven care may aid in overcoming those discrepancies. The aim of this study was to compare the rate of prolonged LOS after colectomy at a county hospital (CH) to a university hospital (UH), and evaluate the underlying factors contributing to the prolonged LOS.
Methods:
NSQIP participant user files from October 2014 to December 2016 from the CH and affiliated UH were utilized to identify all patients captured by respective institutional NSQIP, who underwent colectomy. During this time period, enhanced recovery pathways (ERP) were in place and operational at both institutions. Patients were flagged as prolonged LOS as defined by NSQIP. Charts were reviewed to determine the primary cause of increased LOS on post-operation day 6, and classified as ileus, leak or intra-abdominal abscess (IAA), surgical site infection (SSI), other infection, hemorrhage, medical complication or disposition planning.
Results:
The cohort included 239 patients, 57 from CH and 182 from UH. There was no statistically significant difference between the number of patients with prolonged LOS between the university and county setting, 37% vs 33% (p=0.75). The reasons for increased LOS were equivalent at both locations. Notable differences were apparent in an increased number of Hispanic (30% vs 8%, p<0.001) and African American (42% vs 9%, p<0.001) patients at the county hospital. Other demographic variables such as BMI, gender and age were similar. Tobacco use was also increased at the community hospital (39% vs 13%, p<0.001). All other NSQIP defined comorbidities were equivalent. There was no difference in emergent case status. An ERP was in place at both institutions, with usage of 55% at UH and 69% at CH (p=0.072). Of those patients on an ERP, 3 total patients from UH (4.5%) and 6 total patients from CH (15%) had a prolonged LOS (p=0.003).
Conclusion:
Despite different practice environments, there were no statistically significant differences in the reasons for prolonged LOS between a large charity county hospital and a tertiary university hospital. In this case the treating physicians practice at both locations and practice patterns can be somewhat standardized despite differences in available resources at the two locations. This fact may account for the equivalence in both locations. Additionally, both a large county hospital and a university hospital may be subject to a disproportionately higher volume of more complex patients based on their referral and transfer patterns accounting for increased LOS. The difference in ERP usage between the two institutions is likely the result of a more inclusive program at the CH.