Digital health care decisions made by families of individuals with Down syndrome – Dr. Heidi Berger
This project focuses on analyzing digital health care decisions made by families of individuals with Down syndrome. Currently, there are 58 Down Syndrome specialty clinics across the country in 32 different states. It is estimated that 4-5% of eligible patients are enrolled in specialty clinics. To enable better access to specialty clinics, the Down Syndrome Program and Lab of Computer Science at Massachusetts General Hospital launched a virtual clinic called Down Syndrome Clinic to You (DSC2U). The goal of this project is to better understand how DSC2U fits caregiver needs for individuals with Down syndrome.
This project builds on previous work done in understand access to care for individuals with Down syndrome. In Summer 2016, Bryan Summer research students approached this question nationwide, identifying large-scale regions of inaccessibility and proposing new clinic locations. In Summers 2018 and 2019, Bryan Summer research students administered surveys that asked families questions about their experiences in a Down syndrome specialty clinic with the goal of better understanding the value-added by attending these clinics.
A new survey is being administered nationally and we will clean these data, conduct exploratory data analyses, and hopefully create predictive models from these data. This work will be done collaboratively with Brian Skotko, the Director of the Down Syndrome Program at Massachusetts General Hospital, and Jill Wittmer, a Market Research Consultant at Mizzouri.
Data Augmentation – Dr. Marilyn Vazquez Landrove
Data augmentation is the task of increasing the amount of data available without collecting new data. This is important to applications where collecting data can be costly (e.g., labor intensive, time consuming, or requiring highly specialized machinery), but more data are needed to construct accurate machine learning models. Another very important application is in patient data privacy. Being able to produce a data set with similar characteristics to one collected from human subjects would enable analyses of the data with less risk of infringing upon patient privacy. This project builds upon results from the 2020 summer research program at the Mathematical Bioscience Institute. In this project, we will explore new methods for data augmentation and will apply several methods to both real and simulated data. Student participating in this project will gain experience coding in Python, but no prior experience in this area is required.