Interhospital Variability in 180-day Infections Following Cardiac Surgery

Objective: This study sought to: (1) evaluate hospital-level variation in infections following cardiac surgery and (2) develop and evaluate a 180-day infection quality metric.

Methods: This study evaluated Medicare claims that were merged with institutional Society of Thoracic Surgeons Adult Cardiac Surgery Database files among patients undergoing cardiac surgery across 33 Michigan centers. The primary outcome was an infection within 180 days of surgery. Adjusted institutional infection rates were estimated using logistic regression with robust variance estimation. Terciles of observed/expected ratios were created to assess interhospital variability in infections and associated morbidity and mortality.

Results: In total, 5,466 operations were evaluated. Average (SD) age was 71.1 ± 7.8 years, 29.5% were female, and 4.8% of patients were of Black race. The infection rate was 21.2% although higher among females. Infections were associated with lower left ventricular ejection fraction, diabetes, severe chronic lung disease, cerebrovascular disease, and urgent operations (all p<0.0001). The most common infection was pneumonia (8.5%). Adjusted infection rates varied 39.5% across hospitals (7.2%-46.7%). Patients in the highest versus lowest observed/expected infection tercile hospitals had increased associated discharge to extended care/rehabilitation (27.9% versus 24.7%, p<0.0001) although equivalent stroke and mortality risk.

Conclusions: One in five Medicare beneficiaries develop a 180-day infection following cardiac surgery, with rates varying 39.5% across hospitals. Patients at higher versus lower O:E tercile hospitals were more commonly discharged to extended care/rehabilitation settings although equivalent rates of stroke and mortality. Collaborative learning interventions may be warranted to advance the observed variability in 180-day infections.