Multidisciplinary hepatocellular carcinoma (HCC) treatment is associated with optimal outcomes. There are few data analyzing the impact of treating hospitals' therapeutic offerings on survival. We performed a retrospective cohort study of patients aged 18-70 years with HCC in the National Cancer Database (2004-2012). Hospitals were categorized based on the level of treatment offered (Type I-nonsurgical; Type II-ablation; Type III-resection; Type IV-transplant). Associations between overall risk of death and hospital type were evaluated with multivariable Cox shared frailty modeling. Among 50,381 patients, 65% received care in Type IV hospitals, 26% in Type III, 3% in Type II, and 6% in Type I. Overall 5-year survival across modalities was highest at Type IV hospitals (untreated: Type IV-13.1% versus Type I-5.7%, Type II-7.0%, Type III-7.4% [log-rank, P < 0.001]; chemotherapy and/or radiation: Type IV-18.1% versus Type I-3.6%, Type II-4.6%, Type III-7.7% [log-rank, P < 0.001]; ablation: Type IV-33.3% versus Type II-13.6%, Type III-23.6% [log-rank, P < 0.001]; resection: Type IV-48.4% versus Type III-39.1% [log-rank, P < 0.001]). Risk of death demonstrated a dose-response relationship with the hospital type-Type I (ref); Type II (hazard ratio [HR] 0.81, 95% confidence interval [0.73-0.90]); Type III (HR 0.67 [0.62-0.72]); Type IV hospitals (HR 0.43 [0.39-0.47]). Conclusion: Although care at hospitals offering the full complement of HCC treatments is associated with decreased risk of death, one third of patients are not treated at these hospitals. These data can inform the value of health policy initiatives regarding regionalization of HCC care.
© 2018 by the American Association for the Study of Liver Diseases.