Fifty-Percent Reduction in CABG Mortality During a 3-Year Multi-Institutional Quality Improvement Project

Neil Worrall, Providence Sacred Heart Medical Center and Children's Hospital, Spokane, WA
James Brevig, Providence Regional Medical Center, Everett, WA
Ruyun Jin, Center for Cardiovascular Analytics, Research, and Data Science; Providence St Joseph Heart Institute, Portland, OR
Tyler J Gluckman, Providence Heart and Vascular Institute, Portland, OR, USA.
Robert Hunter, Providence Sacred Heart Medical Center and Children's Hospital, Spokane, WA
Matthew Ducsik, Providence St Joseph Heart Institute, Renton, WA
Braden Batkoff, Providence Sacred Heart Medical Center and Children's Hospital, Spokane, WA
Gary Grunkemeier, CARDS, Providence St. Joseph Health


Objective: Risk-adjusted operative (30 day or before discharge) mortality is a key quality measure for isolated coronary artery bypass graft (CABG) surgery. Through a multi-faceted quality improvement (QI) initiative, we sought to aggressively improve this measure at 14 surgical programs within a large and geographically dispersed health care system.

Methods: Observed mortality rates (O) for isolated CABG surgery were collected from January 2012 through June 2017. Expected mortality rates (E) during this same period were determined using the STS risk model. Comparisons were made between O/E ratios during the baseline (January 2012 through December 2014) and intervention periods (January 2015 through June 2017).
The QI intervention was multifaceted, consisting of 1) regular sharing of un-blinded institutional- and surgeon-level data, 2) a standardized cardiac surgery QI process at each institution, with designated QI physician leads, regular multi-disciplinary morbidity/mortality meetings, review of all surgical deaths using the Phase of Care Mortality Analysis (POCMA) format, and implementation of a prospective high-risk case review process for patients with a predicted mortality >2% or combined mortality/morbidity >12.5%, 3) regular system-wide QI teleconferences and in-person meetings, 4) annual O/E mortality targets, 5) identification of underperforming institutions and creation of a corresponding non-punitive QI action plan, and 6) implementation of checklists to drive care standardization throughout the perioperative period.

Results: 11,201 isolated CABG procedures were performed at 14 institutions across 5 states during the 5.5 year period (annual institutional case volumes of 22 to 356). The baseline (3-year) O/E ratio was 1.05, rising to 1.19 during the 12 months prior to the intervention period (Table). During the intervention period, the O/E ratio decreased significantly to 0.59 for July 2016-June 2017 (p=.007 vs. baseline period) (Figure). Stratification of expected mortality into risk bands of <0.56%, <1.01%, <2.05%, < 10%, and >10% did not demonstrate a shift in expected mortality. Additional analyses did not identify a shift of high-risk patients to percutaneous coronary intervention (data not shown), nor a change in surgical case volume (Table). One underperforming program was closed with consolidation into a neighboring program; another was mentored by a high-performing program.

Conclusions: This study demonstrates a highly significant and clinically meaningful reduction in the O/E ratio of mortality for isolated CABG surgery following the implementation of a multi-faceted QI initiative across a large multi-institutional health care system. Over a 2.5 year intervention period, the O/E ratio of mortality was reduced by 50% compared to the prior year and 44% compared to the prior three years, without changes in institutional/aggregate case volume or expected mortality. Keys to success included regular, un-blinded sharing of data, development of a standardized QI process, improvement and standardization of care delivery, setting of QI targets, and a shared vision for improved patient outcomes.