Brazilian System for Cardiac Operative Risk Evaluation CABG

Overview

Grouping relevant demographic and clinical information at specific stratified levels, and in correlation with the required resource sets, represents the possibility to adapt, improve, and innovate in cardiac surgery programs. In this path, the pursuit of implementing large databases for continuous improvement cycles becomes the foundation of the entire process. In Brazil, where structural and socioeconomic differences are significant, the standardization and sustainability of a large database become a major challenge. Projects like REPLICCAR (São Paulo Cardiovascular Surgery Registry) and BYPASS Registry, which were created to identify improvement opportunities, were temporary. The absence of a continuous multicenter registry makes it difficult to understand risk-adjusted patient outcomes and to implement cost-effective quality initiatives. For example, in 2022, the Ministry of Health launched the QualiSUS Cardio program, which began to consider hospital reimbursement not only in terms of surgical volume but also in relation to mortality, length of stay, and readmission rates after cardiac surgery. Changes like this already occur in various scenarios where payers, whether public or private, seek reimbursement through value-based models. However, for this to happen, the results need to be adjusted to patient risk. Otherwise, hospitals treating more severe patients would be at a disadvantage. This study aims to develop CABG-BraSCORE, a risk score to stratify patients referred for CABG surgery, adjusting outcomes to patient risk, and seeking continuous improvement for cardiac surgery programs in Brazil.

Study Procedure: Data collection will be conducted from patient records, including demographic characteristics, preoperative status, surgical and postoperative data of CABG surgery, and follow-up up to 30 days post-discharge. Centers with prospective data collection: Participants will be contacted 30 days after surgery and invited to participate in the study. The Free and Informed Consent Form (FICF) will be applied via telephone (Appendix 1). All FICF sessions will be recorded, downloaded, and stored according to the established study protocols. Centers with retrospective data collection: Thirty days after surgery, it should be recorded whether there was a return visit to the outpatient clinic, readmission to the institution, the reason for readmission, such as surgical complications, stroke, surgical wound infection, myocardial infarction, and mortality at 30 days. All records should be consulted in the patient's medical record. If there is no information recorded about the patient's return visit, loss to follow-up should be considered. A loss of 20% was considered for the sample composition. After the telephone FICF, a questionnaire (Appendix 2, within the Cardux platform's "Post-discharge Events" section) will be applied to gather information regarding readmissions, death, adverse events, among others, within 30 days after hospital discharge. Patient data will be entered into the Cardux platform according to study protocol variables. All inclusion steps will be guided by the coordinating center, providing clarification about the study protocol and ensuring its compliance, respecting the ethical and legal principles of the study. * Risks to Participants Participants will be exposed to minimal risk as this study will not involve interventions, only a single telephone interview. Therefore, no risks to participants are anticipated, as the data collection will be entirely retrospective and the consent forms will be stored confidentially. * Benefits to Participants: Participants will not receive any direct benefit from participating in this study. However, in the future, they might benefit from improved risk stratification for CABG surgeries, contributing to enhanced patient care and resource allocation. * Cardux Platform Cardux platform will be used for data management and analysis. This platform facilitates real-time monitoring and auditing of data, ensuring consistency and accuracy. The platform also supports predictive modeling and quality assessment. * Quality Audit of Records: To ensure data quality, a series of audits will be conducted. These audits will involve checking data completeness, consistency, and accuracy. Discrepancies will be resolved through communication with the participating centers. * Data analysis will include statistical methods to evaluate risk factors and develop the BraSCORE. The model's performance will be validated and compared with existing international risk scores (EuroSCORE III and STS) * Development of Machine Learning Algorithms Machine learning algorithms will be employed to develop predictive models for hospital mortality, readmission risk, and prolonged hospitalization.