Failure-to-rescue rate as a measure of quality of care in a cardiac surgery recovery unit: a five-year study.

Abstract

Failure to rescue, which is defined as the probability of death after a complication that was not present on admission, was introduced as a quality measure in the 1990s, to complement mortality and morbidity outcomes. The objective of this study was to evaluate possible incremental benefits of measuring failure to rescue after cardiac surgery, to facilitate quality improvement efforts. Data were collected prospectively on 4,978 consecutive patients who underwent cardiac operations during a 5-year period. Institutional logistic regression models were used to generate predicted rates of mortality and major complications. Frequency distributions of morbidities were determined, and failure to rescue was calculated. The annual failure-to-rescue rates were contrasted using χ(2) tests and compared with morbidity and mortality measures. The overall mortality rate was 3.6%, the total complication rate was 16.8%, and the failure-to-rescue rate was 19.8% (95% confidence interval, 17.1% to 22.7%). The predicted risk of mortality and of major complications increased during the last 2 years of the study, whereas the observed complication rate decreased. Failure to rescue for new renal failure was the highest of all complications (48.4%), followed by septicemia (42.6%). Despite the decreased complication rate toward the end of the study, the failure-to-rescue rate did not change significantly (p = 0.28). Failure to rescue should be monitored as a quality-of-care metric, in addition to mortality and complication rates. Postoperative renal failure and septicemia still have a high failure-to-rescue rate and should be targeted by quality improvement efforts.