Abstract
STUDY DESIGN: Retrospective Cohort Study. BACKGROUND: Elective surgical management of refractory degenerative cervical disease with anterior cervical discectomy and fusion (ACDF) is considered the standard of care but may lead to prolonged postoperative length of stay (LOS), increasing patient burden and healthcare costs. While national databases such as the National Surgical Quality Improvement Program (NSQIP) have identified risk factors for extended LOS, their applicability to individual institutions remains uncertain. We sought to evaluate the consistency of prolonged LOS risk factors between NSQIP and a single institutional database. MATERIALS AND METHODS: A retrospective observational study was conducted on patients undergoing elective ACDF from 2008 to 2019 using NSQIP and single institutional databases. Inclusion criteria were age >18 years and current procedural terminology codes 22551, 22552, and 22554. Independent variables, including comorbidities, demographics, and perioperative data, were assessed as predictors of prolonged LOS, defined as >2 days. Patients were matched 1:1 using logistic regression and stratified into short- and long-stay cohorts. RESULTS: 1353 NSQIP patients were matched to 1353 institutional patients. In the NSQIP cohort, anemia (P = 0.01), age (P = 0.01), length of surgery (P < 0.001), and female sex (P = 0.02) were associated with extended LOS. In the institutional cohort, significant predictors included age (P = 0.001), American Society of Anesthesiologists >2 (P = 0.001), length of surgery (P < 0.001), diabetes (P = 0.02), and female sex (P = 0.002). CONCLUSION: Risk factors for prolonged LOS after ACDF differed between NSQIP and the single-institution cohort, suggesting that national databases may lack the granularity needed for institution-specific decision-making. Institutional analyses may better support targeted quality improvement for elective ACDF.