Early prediction of critical events for infants with single-ventricle physiology in critical care using routinely collected data

Abstract

Intensive care units (ICUs) provide care for critically-ill patients who require constant monitoring and the availability of specialized equipment and personnel. In this environment, a high volume of information and a high degree of uncertainty present a burden to clinicians. In specialized cohorts, such as pediatric patients with congenital heart defects (CHDs), this burden is exacerbated by increased complexity, the inadequacy of existing decision support aids, and the limited and decreasing availability of highly-specialized clinicians. Among CHD patients, infants with single ventricle (SV) physiology are one of the most complex and severely-ill sub-populations. While SV mortality rates have dropped, patient deterioration may happen unexpectedly in the period before patients undergo stage-2 palliative surgery. Even in expert hands, critical and potentially catastrophic events (CEs), such as cardiopulmonary resuscitation (CPR), emergent endotracheal intubation (EEI), or extracorporeal membrane oxygenation (ECMO) are common in SV patients, and may negatively impact morbidity, mortality, and hospital length of stay. There is a clinical need of predictive tools that help intensivists assess and forecast the advent of CEs in SV infants. Although ubiquitous, widely adopted ICU severity-of-illness scores or early warning systems (EWS), e.g., PRISM and PIM, have not met this need. They are often v developed for general ICU use and do not generalize well to specialized populations. Furthermore, most EWS are developed for prediction of patient mortality. Among SV patients, however, death is semi-elective. On the other hand, prediction of CEs may help clinicians improve patient care by anticipating the advent of patient deterioration. In this dissertation, we aimed to develop and validate predictive models that achieve early and accurate prediction of CEs in infants with SV physiology. Such models may provide early and actionable information to clinicians and may be used to perform clinical interventions aimed at preventing CEs, and to reducing morbidity, mortality, and healthcare costs. We assert that our work is significant in that it addresses an unmet clinical need by achieving state-of-the-art, early prediction of patient deterioration in a challenging and vulnerable population. vi TABLE OF CONTENTS

Lessons Learned From the First Pilot Study of the Compensatory Reserve Index After Congenital Heart Surgery Requiring Cardiopulmonary Bypass

“…[5][6][7] Dense physiologic data may provide further insights about critically ill pediatric patients, especially when analyzed with ML. 26 Rusin et al 9 and Ruiz et al 10 have successfully used dense physiologic data and ML to predict deterioration in patients with single ventricle physiology. However, existing predictive algorithms have important limitations.…”

“…Models published to date rely on a relatively sophisticated informatics infrastructure to collate, clean, and process physiologic data not accessible to all hospitals. [9][10][11] Data fed into existing models are often invasively obtained or inherently flawed (ie, traditional vital signs), [9][10][11] which may contribute to known issues with missingness and low positive predictive values. 8 Only one algorithm to date (IDO 2 index, Etiometry, Inc) integrates with a visualization platform in near real time at the bedside, 11 which is a prerequisite to making timely inferences and decisions about critically ill patients.…”

“…7 Providers caring for critically ill patients have access to dense physiologic data, which can be better understood using advanced analytic techniques. 8 Predictive algorithms based on physiologic data have been developed for critically ill children designed to herald clinical decompensation 9,10 or inadequate oxygen delivery (IDO 2 index, Etiometry, Inc). 11 The Compensatory Reserve Index (CRI) is a noninvasive Food and Drug Administration (FDA)-cleared monitoring technology that predicts hemodynamic decompensation from hemorrhagic shock in adults using machine learning (ML).…”

Commentary: I am not throwing away my shot…to predict when your patient will decompensate

“…[5][6][7] Model scheme is essential to unlocking the full potential of predictive modeling and the ability to adapt to different populations, institutions, and clinical settings. Ruiz and colleagues 1 begin the transition from physician-based decision making to models reliant on expert knowledge to guide machine learning. A structured approach is intuitive, but it is a form of brute force logic that has limitations.…”

Contribuciones de las técnicas machine learning a la cardiología. Predicción de reestenosis tras implante de stent coronario

“…Prediction of One-Year Transplant-Free Survival after Norwood Procedure Based on the Pre-Operative Data DL Predicción de mala evolución en niños con ventrículo único hipoplásico sometidos a cirugía de Norwood Lu et al (2018) 703 A Novel Deep Learning based Neural Network for Heartbeat Detection in Ballistocardiograph DL Detección automática de latidos cardiacos durante la balistocardiografía (registro de los movimientos producidos por el impacto de la sangre en el corazón y grandes vasos) Chen et al (2018) 704 Artificial Neural Network: A Method for Prediction of Surgery-Related Pressure Injury in Cardiovascular Surgical Patients Machine Learning Approach for Predicting Wall Shear Distribution for Abdominal Aortic Aneurysm and Carotid Bifurcation Models ML, DL Algoritmo de cálculo de la distribución de la tensión de cizallamiento de la pared arterial, como mecanismo esencial en el desarrollo de aterosclerosis Diaz et al (2018) 710 Modeling the control of the central nervous system over the cardiovascular system using support vector machines Predictive performance of six mortality risk scores and the development of a novel model in a prospective cohort of patients undergoing valve surgery secondary to rheumatic fever ML Predicción de riesgo tras cirugía valvular en pacientes con cardiopatía reumática y comparación con scores de riesgo clásicos Marateb et al (2018) 719 Prediction of dyslipidemia using gene mutations, family history of diseases and anthropometric indicators in children and adolescents: The CASPIAN-III study ML Diagnóstico de dislipemia usando datos genéticos, familiares y antropométricos en niños y adolescentes Jalali et al (2018) 720 Prediction of Periventricular Leukomalacia in Neonates after Cardiac Surgery Using Machine Learning Algorithms Prognostic value of CT myocardial perfusion imaging and CTderived fractional flow reserve for major adverse cardiac events in patients with coronary artery disease ML Evaluación de valores pronósticos de imagen de perfusión miocárdica por tomografía y reserva fraccional de flujo a partir de tomografía para la predicción de eventos cardiacos mayores en pacientes con enfermedad coronaria Ruiz et al (2019)735 Early prediction of critical events for infants with single-ventricle physiology in critical care using routinely collected data…”