Abstract 194: Real-time Outcomes Analysis Dashboard

    loading  Checking for direct PDF access through Ovid


Background: Surveillance of risk-adjusted pediatric cardiac surgery outcomes has moved beyond mortality to complications. We report a novel dashboard that provides for real-time graphic presentation of risk-adjusted complications to ensure that program performance is maintained and to identify areas for improvement.

Methods and Materials: This was a single center study, performed at a medium volume program after IRB approval. Data were collected on consecutive patients over a one-year period (Oct 4, 2016- Oct 3, 2017) coincident with the restart of a pediatric cardiac surgical program. Mortality data are plotted using Variable Life Adjusted Display (VLAD). Complications are collected for each patient prospectively using STS standard definitions and CUSUM and O-E plots generated. Complication burden (accounting for both frequency and severity of complications) was calculated for each STAT category and compared to historic data.

Results: Mortality rate for the cohort was 3.6% (6/164, Fig 1, panel A). In 94 patients, there were no complications (57.3%). Thirty-five patients had 1 complication and 35 had <2 complications. Complications resulting in mortality (failure-to-rescue rate) is 8.7%. Most common complications were pulmonary (40, 23.4%) arrhythmia (27, 15.8%), and neurologic (28, 16.4%). High risk complications occurred infrequently including acute renal failure (n=14), bleeding requiring reoperation (n=4), unexpected cardiac arrest (n=3), low cardiac output (n=11), mechanical circulatory support (n=7), neurologic deficit (n=9), and sepsis (n=1) (Fig 1, panel B). Four outlier patients account for spikes in the O-E CUSUM plot (Fig 1, panel C). Complication severity and frequency escalate with increasing STAT category (Fig 1, panel D).

Conclusion: A continuously updated dashboard permits real-time adverse outcome reporting, potentially allowing earliest identification of areas of concern or focus. Proof of the value of this approach will require replication in additional centers of various size.

Related Topics

    loading  Loading Related Articles