P062. Assessment of Wound Infection Risks Post-Hemiarch Surgery: A Logistic Regression Approach
Presented During: POD Abstracts (Available for viewing in the exhibit hall for the duration of the meeting)
University of Colorado Anschutz
Denver, CO
United States - Contact Me
Adam Carroll is a current PGY-3 surgical resident at the University of Colorado. Adam attended medical school at the University of Colorado and has been involved in research with the cardiothoracic surgery department throughout medical school and residency. He has interest specifically in endovascular and transcatheter aortic interventions, as well as neurologic outcomes in aortic research. He is currently in the aortic surgery research labaratory in the Department of Cardiothoracic Surgery at the University of Colorado. He plans to pursue a career in cardiothoracic surgery following his general surgery residency.
Thursday, April 25, 2024: 5:38 PM - 7:00 PM
Sheraton Times Square
Room: Central Park
Description
Purpose:
Surgical site infection carries a significant risk of increased mortality and cost. In the context of aortic surgery, there is elevated risk due to the potential necessity for multiple debridements in the case of sternal wound infection and potential graft material. We utilized logistic regression algorithms in hemiarch surgery to predict patients at risk for infection and to elucidate specific risk factors.
Methods:
We identified a total of 602 adult patients who underwent hemiarch replacement between June 2009 and October 2022 from our single institution prospectively maintained database. These patients were randomly divided into training (70%) and testing sets (30%) with various logistic regression models constructed to predict post-operative infection in the cardiothoracic intensive care unit (CTICU). From the index hospitalization data, we extracted 17 demographic and pre-operative characteristics. To assess model performance, we used multiple evaluation metrics, including accuracy, Brier score, and area under the receiver-operating characteristic curve (AUC-ROC). Furthermore, we calculated odds ratios and confidence intervals derived from the logistic regression model.
Results:
Development of post-operative infection in the CTICU was noted in 40 patients (6.64%) who underwent hemiarch replacement. The final logistic regression model demonstrated a cross-validation accuracy of 94% and was well-calibrated as evidenced by the low Brier score of 0.06. The predictor also demonstrated strong performance on the testing set, achieving an accuracy of 89%. Our best performing CTICU post-operative infection prediction model achieved an AUC-ROC of 0.71. Increased infection risk was associated with most comorbidities, particularly diabetes, a prior history of CT surgery, concomitant root replacement, and urgent/emergent procedures. Protective factors included undergoing hemiarch surgery without concomitant root replacement and elective procedures.
Conclusions:
Fine-tuned logistic regression models have the potential to provide excellent prognostic accuracy for those at risk for wound infection. Specific features elucidated by the algorithm may help to better predict those at risk for infection, which may in turn affect clinical decision making.
Authors
Adam Carroll (1), Nicolas Chanes (1), Muhammad Aftab (1), T. Brett Reece (1)
Institutions
(1) University of Colorado Anschutz, Denver, CO
Surgical site infection carries a significant risk of increased mortality and cost. In the context of aortic surgery, there is elevated risk due to the potential necessity for multiple debridements in the case of sternal wound infection and potential graft material. We utilized logistic regression algorithms in hemiarch surgery to predict patients at risk for infection and to elucidate specific risk factors.
Methods:
We identified a total of 602 adult patients who underwent hemiarch replacement between June 2009 and October 2022 from our single institution prospectively maintained database. These patients were randomly divided into training (70%) and testing sets (30%) with various logistic regression models constructed to predict post-operative infection in the cardiothoracic intensive care unit (CTICU). From the index hospitalization data, we extracted 17 demographic and pre-operative characteristics. To assess model performance, we used multiple evaluation metrics, including accuracy, Brier score, and area under the receiver-operating characteristic curve (AUC-ROC). Furthermore, we calculated odds ratios and confidence intervals derived from the logistic regression model.
Results:
Development of post-operative infection in the CTICU was noted in 40 patients (6.64%) who underwent hemiarch replacement. The final logistic regression model demonstrated a cross-validation accuracy of 94% and was well-calibrated as evidenced by the low Brier score of 0.06. The predictor also demonstrated strong performance on the testing set, achieving an accuracy of 89%. Our best performing CTICU post-operative infection prediction model achieved an AUC-ROC of 0.71. Increased infection risk was associated with most comorbidities, particularly diabetes, a prior history of CT surgery, concomitant root replacement, and urgent/emergent procedures. Protective factors included undergoing hemiarch surgery without concomitant root replacement and elective procedures.
Conclusions:
Fine-tuned logistic regression models have the potential to provide excellent prognostic accuracy for those at risk for wound infection. Specific features elucidated by the algorithm may help to better predict those at risk for infection, which may in turn affect clinical decision making.
Authors
Adam Carroll (1), Nicolas Chanes (1), Muhammad Aftab (1), T. Brett Reece (1)
Institutions
(1) University of Colorado Anschutz, Denver, CO
Presentation Duration
PODS will be on display in the exhibit hall for the duration of the meeting during exhibit hall hours. PODS will also be available for viewing on the meeting website. There is no formal presentation associated with your POD, but we encourage you to visit the PODS area during breaks to connect with those viewing.