Pediatric Cardiac Catheterization: Predicting Who Will Require High Level Care
Faculty Mentor: Kimberlee Gauvreau
Graduate Student Mentor: Octavious Talbot
Program Participants: Caroline Echeandia-Francis, Andrea Rivera, Monique Sparkman
Our team aims to help predict which patients undergoing cardiac catheterization will need a higher level of post-procedure care. While cardiac catheterization is a less invasive treatment method for congenital heart defects when compared to surgery, approximately 15% of patients require a higher level of care or monitoring in the intensive care unit (ICU). It is important to be able to predict which patients will likely need a higher level of care since ICU beds are limited. An effective predictive model will help physicians schedule procedures to avoid ICU overcrowding. To accomplish this, we analyzed data from children and adults who were treated for a congenital heart defect via cardiac catheterization at Boston Children’s Hospital (BCH) between August 2017 and December 2019. Factors such as the age of the patient and whether or not the patient had a systemic illness informed our predictive model for assessing high-risk and low-risk ICU patients.
Visualizing the Effects of Climate and GDP on COVID-19 Transmission
Faculty Mentor: Rafael Irizarry
Graduate Student Mentors: Isabella Grabski
Program Participants: Runa (Yan) Cheng, Erick Ivanovich Méndez, Gabriela M. Lozano Pérez, Addison McGhee
COVID-19 (Coronavirus Disease – 2019) is a severe respiratory syndrome that quickly spread across the globe. After taking into account population size and testing capacity, our group investigated how climate (temperature/humidity) and economic indicators (GDP) relate to the number of COVID-19 cases and deaths within the United States and around the world.
Classification/Cluster-Based ML Approaches to Investigate Groundwater Contamination at Coal Ash Dumps
Faculty Mentor: Rachel Nethery
Graduate Student Mentor: Luli Zou
Program Participants: Antonella Basso, Jose Lopez, Tony Ni
Power companies and coal-fired plants across the US have dumped coal ash into landfills and ponds without regard to toxic contaminants that leak into groundwater for much of the last century, posing health risks like cancer, neurological impairments to children, and human reproductive defects. In this research project, we will investigate the prominence of contamination amongst upgradient wells through exploratory statistical analysis and classification/cluster-based machine learning approaches.
Food for Thought: An Exploration of Demographic Factors Related to Dietary Behaviors and Cardiovascular Health in the US
Faculty Mentor: Briana Stephenson
Graduate Student Mentor: Jeanette Varela
Program Participants: Daniel Chan, Tamantha Pizarro, Courtney Rabb, Austin Zane
Using data from the National Health and Nutrition Examination Survey (NHANES), we examined the relationship of population demographics on dietary behaviors, nutritional intake, and cardiovascular disease risk factors. We performed exploratory data analysis and utilized Bayesian and frequentist inference to model associations and characterize the NHANES data. We hope that our findings can increase our understanding of disparities present in nutritional health and cardiovascular health within the United States.
Association of Race and Ethnicity with Medical Outcomes in Pediatric Intensive Care
Faculty Mentor: David Wypij
Graduate Student Mentor: Christina Howe
Program Participants: Sakina Ali, Vincent Buckman, Prashit Parikh
Previous research suggests strong, but inconsistent, evidence of racial/ethnic disparities in outcomes for clinical care in both adult and pediatric populations. In this study, we conducted secondary analyses on data obtained from a cluster-randomized clinical trial of pediatric intensive care units: Randomized Evaluation of Sedation Titration for Respiratory Failure (RESTORE). Our goal was to analyze significant differences among participants identified as non-Hispanic Black, non-Hispanic White and, Hispanic of any race in the control and intervention arms of the study. We used a variety of modeling techniques, including proportional hazards regression and generalized estimating equations to identify racial/ethnic disparities within our study population.