Good at One or Good at All? Variability of Coronary and Valve Operation Outcomes Within Centers

Abstract

The technical expertise required for treatment of coronary and structural heart valve disease differs. Correlation between center-specific mortality rates after coronary artery bypass grafting (CABG) and valve operations has not been demonstrated. This study tested the hypothesis that risk-adjusted outcomes between coronary and valve procedures do not correlate within centers. Records of patients undergoing isolated CABG, isolated aortic valve replacement (AVR), or isolated mitral valve replacement (MVR) procedures from 2008 to 2015 in a multi-institutional Society of Thoracic Surgeons (STS) database were used to generate observed-to-expected (O/E) ratios for morbidity and death. Ratios were based on the STS predicted risks of morbidity and death and were calculated by procedure for each institution. Linear regression models evaluated the relationship between institutional performance in CABG and valve operations. A total of 22,258 records from 18 institutions were analyzed: 17,026 CABG, 3,238 isolated AVR, and 1,994 MVR procedures. With respect to deaths, the correlation coefficients were weak; for AVR and CABG, it was 0.22 and was 0.26 for MVR and CABG. With respect to morbidity, a strong relationship was seen between the morbidity O/E ratios, with coefficients of 1.03 for AVR and 0.97 for MVR, suggesting a nearly 1:1 relationship between morbidities observed in an institution's CABG and valve operations. Sites that perform CABG with low mortality rates may not have similarly low mortality rates with valve operations. Most striking, however, is the nearly identical O/E ratio for morbidity for CABG and valve operations at each center. These findings suggest postoperative care as a major determinant for morbidity after cardiac operation.