Join EVERSANA’s Ramaa Nathan, PhD, VP, Data Science & RWE Insights at the 7th Annual Real World Evidence and Life Science Analytics Conference.
October 17 -18, 2024, Boston, MA
Ramaa Nathan will be presenting on Thursday the 17th at at 5:05pm EST on, “Machine Learning Techniques for Identification of Risk Factors and Guided Personalized Therapies in Outcome Based Models”. 
Session Summary:
Observational real-world evidence (RWE) studies are a valuable and cost-effective alternative to lengthy prospective randomized clinical trials. They provide insights into the effectiveness of treatment pathways outside experimental controlled settings, allowing to identify best therapies that achieve cure or to reduce disease progression in everyday clinical practice.
Machine learning-based models, rather than traditional rule-based models, add predictive value to RWE studies by uncovering hidden risk factors, and by leveraging diverse large amount of diverse data, facilitate the design of personalized treatment strategies. This will be demonstrated through two case studies:
- Case Study #1: Traditional outcome-based models rely on pre-determined, hypothesis-driven risk factors. We introduce ML-Based Outcome Models, a two-step approach that first employs tree-based machine learning models using administrative claims data to identify and interpret risk factors, which are then integrated into classic statistical outcome models.
- Case Study #2: Multiple Sclerosis (MS) is a chronic, progressively debilitating disease. Predictive models can significantly impact patient care by identifying individual risk factors associated with progression from one stage to another. Providers can use this data to customize treatment plans to slow disease progression. We will present a multi-predictive model approach that identifies risk factors associated with disease progression and that identify the optimal treatment sequence to mitigate progression in MS patients.