86: Identifying Characteristics of High-Performing Hospital Networks for Major Cancer Surgery

Saturday, March 25, 2023

10:16 AM – 10:26 AM

Location: Ballroom B

Abstract Presenter(s)

Sara L. Schaefer, MD

Surgical Resident
University of Michigan
Ann Arbor, Michigan, United States

Disclosure(s): No financial relationships to disclose

First Author(s)

Sara L. Schaefer, MD

Surgical Resident
University of Michigan
Ann Arbor, Michigan, United States

Disclosure(s): No financial relationships to disclose

Author(s)

Adrian Diaz, MD (he/him/his)

Clinical Housestaff
The Ohio State University
Columbus, Ohio, United States

Disclosure(s): No financial relationships to disclose
UN

Usha Nuliyalu, MPH

Senior Statistician
University of Michigan, United States

Disclosure information not submitted.
AI

Andrew Ibrahim, MD MPH

Professor
University of Michigan, United States

Disclosure information not submitted.
Hari Nathan, MD, PhD, FACS, FSSO (he/him/his)

Associate Professor of Surgery and Chief, Division of Hepato-Pancreato-Biliary Surgery
University of Michigan
Ann Arbor, Michigan, United States

Disclosure(s): No financial relationships to disclose

Disclosure(s):

Sara L. Schaefer, MD: No financial relationships to disclose

Adrian Diaz, MD: No financial relationships to disclose

Hari Nathan, MD, PhD, FACS, FSSO: No financial relationships to disclose

Introduction: A growing body of literature has described variation in surgical quality within and between hospital networks. However, our understanding of the relationship between quality and structural characteristics of hospital networks remains poor. In this study we sought to assess the relationship between network structural characteristics and outcomes of major cancer surgery.

Methods: We evaluated outcomes for Medicare beneficiaries undergoing colectomy, rectal resection, lung resection, or pancreatectomy for cancer from 2016 to 2020 in hospitals belonging to networks. Hospitals were assigned to the corresponding network using the American Hospital Association Annual Survey. Networks were classified into quintiles of quality with an equivalent number of patients based on the risk- and reliability-adjusted composite rate of 30-day serious complications and mortality. We assessed the relationship between risk- and reliability-adjusted surgical outcomes and network characteristics such as network size, procedure concentration, and annual volume.

Results: The analytic cohort included 123,622 patients (59,054 undergoing colectomy, 13,077 rectal resection, 43,646 lung resection, and 7,845 pancreatectomy). The adjusted composite rate of 30-day serious complications and mortality in high-quality vs low-quality networks was 5.4% vs 8.7% for colectomy, 8.8% vs 10.1% for rectal resection, 4.4% vs 5.9% for lung resection, and 12.1% vs 14.2% for pancreatectomy (all p< 0.001). Networks with the best outcomes had fewer hospitals per network performing the procedure, concentrated procedures to fewer sites and had fewer patients treated at for-profit hospitals. For example, networks with the best colectomy outcomes had fewer hospitals per network (8 vs 16, p< 0.001), performed more procedures per hospital (93 vs 52, p< 0.001), and fewer patients were treated at for-profit hospitals (6.2% vs 44.2%, p< 0.001) compared with networks with the worst outcomes. In higher quality networks, more hospitals met Leapfrog volume standards for pancreatectomy (40.4% vs 17.4%, p< 0.001) and for lung resection (31.7% vs 13.9%, p< 0.001). There was less variation in quality between hospitals belonging to high-quality networks (Figure 1).

Conclusions: For these four major cancer operations, networks with higher annual volumes concentrated in fewer hospital sites achieved superior clinical outcomes. Understanding the characteristics of high- and low-quality networks may guide strategies to improve the quality of surgical care.

Learning Objectives:

Upon completion, participant will be able to describe structural characteristics associated with high- and low-quality hospital networks for cancer surgery.
Upon completion, participant will be able to recommend procedure-specific strategies (i.e. concentrate volume or standardize care) to reduce variation and improve overall quality of care across networks.
Upon completion, participant will be able to identify how specific network characteristics may alter the approach to improving network-wide hospital quality.