ID 2322: Create & evaluate a predictive model for engagement with digital health applications

Master thesis

Digital health applications have been on the rise in recent years, particularly in the management of chronic disease, as these patient-centric applications can extend care into patients` homes and provide self-management assistance crucial to improving patient outcomes. A major challenge in the success of these digital health applications, however, is the high rate of attrition. An observational trial of a large real-world cohort revealed only a 2% sustained, continuous use of the application that would actually improve clinical outcomes.
This master thesis will be conducted together with a startup, Veta Health. This is a digital health company that delivers remote care solutions that help patients with chronic conditions achieve better health outcomes from the comfort of their homes. Veta Health maximizes patient engagement and participation in its programs through the delivery of a personalized experience. In addition to the software solution (Prosper), Veta Health deploys care teams to connect with the patient directly when an intervention is required.

In order to optimize the timing on engagement-based interventions, this thesis will evaluate the engagement of patients with the application in order to find further implications for an individual’s patient journey. Further aspects can include the quantification of risk of attrition or the evaluation of health outcomes.

Details

• Together with a Startup for remote patient monitoring, the aim of this thesis is to develop predictive models for engagement in order to understand the treatment
• In accordance with Veta Health, the models will be developed based on a real-world dataset

Tasks

• Literature research: identify relevant work resulting in a comprehensive overview about existing studies on user engagement with digital health applications
• Development of different ML or AI models to classify engagement
• Evaluation of the implemented models regarding their accuracy on the collected data and evaluate the models with a real-world dataset

Requirements

• Good Knowledge of Python / Deep Learning / PyTorch
• Experience with Git (Commits, Versioning, Project Management, …)
• Strong interest in Digital Health

Supervisors

Anastasiya Zakreuskaya, M. Sc.

Researcher & PhD Candidate