90th Annual Meeting Abstracts

---

Can The American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) Be Used to Improve “As Expected” Outcomes? The Partners Healthcare Surgical Consortium Experience
Stanley Ashley, MD¹, *Matthew Hutter, MD², Pardon Kenney, MD³, Frederick Millham, MD⁴, Selwyn Rogers, MD¹, *Marc Rubin, MD⁵, David Shahian, MD², Andrew Warshaw, MD², Michael Zinner, MD¹, *Jennifer Daley, MD^6
1Brigham and Women's Hospital, Boston, MA;²Massachusetts General Hospital, Boston, MA;³Faulkner Hospital, Boston, MA;⁴Newton-Wellesley Hospital, Newton, MA;⁵North Shore Medical Center, Salem, MA;⁶Partners Healthcare, Inc, Boston, MA

Objective: NSQIP has been employed to improve surgical outcomes by identifying and targeting outliers with higher-than-expected risk-adjusted morbidity and mortality. The NSQIP impact on “As Expected” outcomes hospitals is less well established. Five Partners hospitals developed a commercial payor pay-for-performance contract (P4P) in which a hospital-dollar withhold was returned based on achievement of two of three procedure-specific targets reducing NSQIP-identified postoperative occurrences by 1% compared to the previous year.
Design/Setting/Patients/Measures: In addition to a Partners-wide goal designed to encourage collaborative efforts, each hospital chose two unique NSQIP targets.
Interventions: After a series of consortium meetings, the hospital surgical representatives selected, as the 2008 Partners-wide measure, a 1% reduction in colorectal surgery postoperative occurrences from a 2007 (12 months ending June 2007) overall rate of 29.1%. A variety of approaches were employed to address this target, including group brainstorming, education, and dissemination of process recommendations based on available evidence in the literature to address infectious complications, the major cause of post-op occurrences in partial colectomy.
Results: Final data (1/1/08-12/31/08) are as follows:

Site	2007 Actual	2008 Target	2008 Actual
Hospital A	33.3%		20.0%
Hospital B	11.7%		4.6%
Hospital C	42.5%		34.2%
Hospital D	21.4%		25.3%
Hospital E	37.1%		19%
System-Wide	29.1%	28.1%	22.4%

Employing similar approaches, each of the hospitals achieved at least one their two hospital-specific targets.
Conclusions: Although these results may at least in part reflect a Hawthorne effect more than specific process improvements, they do suggest that ACS-NSQIP can be employed to improve specific “As Expected” outcomes.

---