Original Article
Spontaneous regionalization of esophageal cancer surgery: an analysis of the National Cancer Database
Abstract
Background: Esophagectomy patients are up to three times more likely to die after surgery when cared for at low-volume hospitals (LVHs). Increased awareness by patients and clinicians of the hazards of esophagectomy at LVHs, may inspire a “spontaneous regionalization” away from LVHs, yet the extent to which this has taken place is unclear.
Methods: Retrospective analysis of patients undergoing esophagectomy for esophageal cancer in the National Cancer Database (NCDB) across two eras: 2004–2006 (Era 1) and 2010–2012 (Era 2). Primary outcomes included the proportion of patients at high-volume hospitals (HVHs) (≥13/year per Leapfrog Group), adjusted, and unadjusted 90-day mortality.
Results: The NCDB captured 5,968 esophagectomy patients in Era 1 and 5,580 in Era 2, a 6.5% decrease (P<0.001). Fewer hospitals performed esophagectomies in Era 2 (756 vs. 663, P=0.014), yet the proportion of patients treated at LVHs declined slightly between eras (73% vs. 70%, P<0.001). Patients with high-risk attributes (e.g., advanced age, multiple comorbidities, etc.) were disproportionately treated at LVHs in both eras (77% Era 1, P<0.001, 73% Era 2, P=0.017). However, the 90-day mortality rate for patients with highrisk attributes decreased considerably between Eras at LVHs (19.3% to 12.3%, P<0.001).
Conclusions: Spontaneous regionalization of esophageal cancer surgery has not occurred on a large scale, yet for high-risk patients, the hazards of being cared for at LVHs have dissipated. Further study is needed to optimize alignment of esophagectomy patients and hospitals.
Methods: Retrospective analysis of patients undergoing esophagectomy for esophageal cancer in the National Cancer Database (NCDB) across two eras: 2004–2006 (Era 1) and 2010–2012 (Era 2). Primary outcomes included the proportion of patients at high-volume hospitals (HVHs) (≥13/year per Leapfrog Group), adjusted, and unadjusted 90-day mortality.
Results: The NCDB captured 5,968 esophagectomy patients in Era 1 and 5,580 in Era 2, a 6.5% decrease (P<0.001). Fewer hospitals performed esophagectomies in Era 2 (756 vs. 663, P=0.014), yet the proportion of patients treated at LVHs declined slightly between eras (73% vs. 70%, P<0.001). Patients with high-risk attributes (e.g., advanced age, multiple comorbidities, etc.) were disproportionately treated at LVHs in both eras (77% Era 1, P<0.001, 73% Era 2, P=0.017). However, the 90-day mortality rate for patients with highrisk attributes decreased considerably between Eras at LVHs (19.3% to 12.3%, P<0.001).
Conclusions: Spontaneous regionalization of esophageal cancer surgery has not occurred on a large scale, yet for high-risk patients, the hazards of being cared for at LVHs have dissipated. Further study is needed to optimize alignment of esophagectomy patients and hospitals.
