Quality Indicators for Head and Neck Oncologic Surgery: Academic versus Nonacademic Outcomes

to determine national benchmarks for established quality indicators in head and neck cancer (HNCA) surgery, focusing on differences between academic and nonacademic institutions.

Cross-sectional analysis of national database.

HNCA surgery admissions from the 2009-2011 Nationwide Inpatient Sample were analyzed for preoperative characteristics and postoperative outcomes. Multivariate analyses were used to identify factors influencing quality indicators after HNCA surgery. Quality metrics-including length of stay (LOS), inpatient death, return to the operating room (OR), wound infection, and transfusion-were compared for academic versus nonacademic institutions.

A total of 38,379 HNCA surgery inpatient admissions (mean age, 56.5 years; 52.4% male) were analyzed (28,288 teaching vs 10,091 nonteaching). Nationally representative quality metrics for HNCA surgery were as follows: mean LOS, 4.26 ± 0.12 days; return to OR, 3.3% ± 0.2%; inpatient mortality, 0.7% ± 0.1%; wound infection rate, 0.9% ± 0.1%; wound complication rate, 4.3% ± 0.2%; and transfusion rate, 4.3% ± 0.3%. HNCA surgery patients at teaching hospitals had a greater proportion of males, radiation history, and high-acuity procedures and greater comorbidity scores (all P < .001). Multivariate analyses adjusting for age, sex, income, payer, prior radiation, comorbidity scores, and procedural acuity demonstrated that teaching hospitals had a slightly increased LOS (+0.30 days; P = .009) and odds ratio for wound infection (1.54; 95% CI: 1.22-1.94) versus nonteaching hospitals. There were no significant differences in return to OR (P = .271), inpatient mortality (P = .686), or transfusion rate (P = .960).

Despite caring for substantially more complex HNCA surgery patients with greater comorbidities, teaching hospitals demonstrate only a marginally increased LOS and wound complication rate versus nonteaching hospitals, while other established quality metrics are similar.