(0705) Predictors for Discharge Locations Following Head and Neck Cancer Surgery

Introduction: Appropriate discharge location for postoperative head and neck patients has played a critical role in short-term clinical outcomes. However, while it is known that socioeconomic characteristics influence outcomes post-discharge, how these factors interact with discharge location remain poorly studied and characterized. This study aims to delineate the effect of demographic, clinical and social predictive factors to optimally guide postoperative discharge destination.

Methods: A retrospective chart review was conducted of patients surgically treated by head and neck otolaryngologists at Shands Hospital of the University of Florida between January 1, 2014 and January 1, 2023. Only patients who had inpatient head and neck surgery were included in the study. Charts were queried for demographics, comorbidities, preoperative functional status, occupational status, distance from hospital, and other social determinants of health. Analysis was performed to identify statistically significant factors influencing discharge location.

Results: Among the 1,906 patients included, 628 (32.9%) were female, 194 (10.2%) malnourished, and 1,772 (93%) patients with full preoperative function. The average age was 63.0 (SD 13.7). The majority (57.5%) of patients were discharged to home with a further 24.6% requiring home health, and 17.9% were discharged to a post-acute care facility. Significant predictors of discharge to facility or to home with healthcare were male gender (p < 0.001), advanced age (p < 0.001), malnutrition (p < 0.001), extensive smoking history (p < 0.001), advanced functional status (p < 0.001), and increased distance from home to the hospital (p < 0.001).

Conclusions: Preliminary analysis has identified the demographics of gender and age, nutritional status, preoperative functional status, and location of home as factors influencing short-term clinical outcomes. Utilizing this data, we further aim to create a machine learning model to yield a more comprehensive understanding of the magnitude of impact created by these social factors.