Adjusted hospital sentinel lymph node positivity rates in melanoma a novel potential measure of quality

Objective: Our objectives were to examine whether hospital characteristics are associated with lower-and higher-than-expected sentinel lymph node biopsy (SLNB) positivity rates and whether hospitals with lower-or higherthan-expected SLNB positivity rates have worse patient outcomes. Background: Surgeon and pathologist SLNB technical errors may lead to incorrect melanoma staging. A hospital's SLNB positivity rate may reflect its SLNB proficiency for melanoma, but this has never been investigated. Methods: Stage IA-III melanoma patients undergoing SLNB were identified from the National Cancer Data Base (2004-2010). Hospital-level SLNB positivity rates were adjusted for patient-and tumor factors. Hospitals were divided into terciles of adjusted SLNB positivity rates. Hospital characteristics (using multinomial logistic regression) and survival (using Cox modeling) were examined across terciles. Results: Of 33,639 SLNB patients (from 646 hospitals), 2916 (8.7%) had at least 1 positive lymph node. Hospitals with lower-(low tercile) and higherthan-expected (high tercile) SLNB positivity rates were more likely to be lowvolume hospitals (low tercile: relative risk ratio (RRR)=2.57, P=0.002; high tercile: RRR=2.3, P=0.004) compared to hospitals with expected rates (middle tercile). Stage I patients treated at lower-than-expected SLNB positivity rate hospitals had worse 5-year survival than those treated at expected SLNB positivity rate hospitals (90.0% vs 91.9%, P=0.014; hazard ratio=1.28, 95% CI: 1.05-1.57); survival differences were not observed by SLNB positivity rates for stage II/III. Conclusions: Adjusted hospital SLNB positivity rates varied widely. Surgery at hospitals with lower-than-expected SLNB positivity rates was associated with decreased survival. Hospital SLNB positivity rates may be a novel measure to confidentially report to hospitals for internal quality assessment.