AHRQ Data Inform University of Michigan Research on How to Improve High-Risk Surgery Outcomes

May 2008

Using Nationwide Inpatient Sample (NIS) data from the Healthcare Cost and Utilization Project (HCUP), the Comprehensive Cancer Center at the University of Michigan in Ann Arbor conducted research showing that hospitals with varying mortality rates have important differences in structure and systems. The research provides a preliminary roadmap to aid in determining factors that are related to, and would be good measures of, safe and effective surgical outcomes, particularly when high-risk procedures are involved.

Brent Hollenbeck, MD, MS, Assistant Professor of Urology, used the 2003 NIS to identify the number of patients undergoing a cystectomy to treat bladder cancer. A cystectomy is the surgical removal of all or part of the bladder. The NIS data-together with American Hospital Association data-is helping bring attention to the ways hospitals should work to improve patient care for bladder cancer, radical cystectomy, and for all patient care, despite the volume of surgeries that are performed.

The data allowed Hollenbeck and his colleagues to attempt to measure differences in hospital capacity, staffing, and health services according to the volume of cystectomy operations. Using the 2003 NIS, the researchers identified 1,847 patients undergoing a cystectomy. To determine whether a hospital was a high- or low-volume provider, the number of procedures was categorized. Differences in hospital structure based on five characteristics- capacity, staffing, diagnostic services, interventional services, and specialty services-were a primary outcome measure. Secondary measures were death during hospitalization and prolonged length of stay.

The study found significant differences in hospital structure according to volume. The researchers determined that measurable differences in the availability and breadth of services (consultative, diagnostic, and ancillary) may partially explain the difference between procedure volume and short-term cystectomy outcomes. Adjusting for case mix, patients treated at low-volume hospitals had a mortality rate 3.2 times greater than high-volume providers.

Hollenbeck first became aware of HCUP due to his interest in measuring quality. He and his colleagues were very interested in quality improvement in cancer care. They were also concerned that, in general, younger patients with higher income and better insurance have a greater ability to seek out better care. "When we began looking at the issue of patient outcomes after radical cystectomy," Hollenbeck says, "it [NIS] was the most recent year available at the time. It is the best data source that is representative of what is happening in the U.S. across payer types."

Hollenbeck feels that future studies will focus on assessing quality of care, and a movement toward regionalizing surgery may gain momentum. HCUP databases will be essential to this effort at the state, regional, and national levels. The overall issue of whether there are better outcomes in high-volume hospitals merits further study and is an important quality issue.

In addition to NIS, Hollenbeck often uses the State Ambulatory Surgery Database and the State Inpatient Database in his research, which, he says, "...focus on variation in physician practice style. These data sources are used to understand the upstream influences and the downstream consequences of this variation."

Impact Case Study Identifier: 
AHRQ Product(s): HCUP-Nationwide Inpatient Sample, Healthcare Cost and Utilization Project (HCUP)
Topics(s): Quality, Cancer
Geographic Location: Michigan
Implementer: University of Michigan Comprehensive Cancer Center
Date: 05/01/2008

Hollenbeck BK, Daignault S, Dunn RL, et al. Getting under the hood of the volume-outcome relationship for radical cystectomy. Journal of Urology 2007; 177: 2095-2099.

Page last reviewed October 2014