Predictors of prolonged length of stay after major elective head and neck surgery

Objective/Hypothesis: Longer length of stay (LOS) after elective surgery is associated with an increased use of health care resources and higher costs. The objectives of this study were to determine the perioperative factors that predict a prolonged LOS after elective major head and neck operations and to test the hypothesis that factors related to process of care (intra- and postoperative) independently predict prolonged LOS after adjustment for preoperative patient characteristics. Study Design: Prospective hospital-based cohort study. Methods: The National VA Surgical Quality Improvement Program data were accessed for seven head and neck operations: radical neck dissection (RND) (n = 398), modified RND (n = 891), total laryngectomy (n = 431), total laryngectomy with RND (n = 747), hemiglossectomy with unilateral RND (n = 201), composite resection (n = 105), and composite resection with RND (n 312). Prolonged LOS was defined as exceeding the 75th percentile for the LOS distribution of each operation. Multivariable logistic regression analysis was performed to identify factors that predicted prolonged LOS. Results: Sixty-eight variables were analyzed among 3,050 patients who qualified for inclusion. Preoperative patient characteristics that predicted prolonged LOS were older age, poorer functional status, consumption of more than two drinks of alcohol per day, history of chronic obstructive pulmonary disease, and diabetes mellitus. Intraoperative processes that predicted prolonged LOS were a longer operative time and transfusion of erythrocytes. The postoperative variables that predicted a prolonged LOS were a return to the operating room within 30 days of the index operation and the occurrence of two or more operative complications. Conclusion: Several intraoperative processes and postoperative adverse events contributed additional predictive information for prolonged LOS, after consideration of preoperative patient characteristics.