Life at ZS

ZSers shine in FDA V-CHAMPS Challenge

By Shriram Tarawade

March 7, 2024 | Article | 3-minute read

ZSers shine in FDA V-CHAMPS Challenge

Preventing and lowering the risk of heart disease, a leading cause of death globally, requires advanced forecasting capabilities to predict cardiovascular-related health outcomes. To drive advancements in this space, the U.S. Food and Drug Administration (FDA), Veterans Health Administration (VHA) and collaborating partners created a competition to develop and evaluate AI and ML models to improve care for U.S. veterans. More than 300 participants from 16 countries registered for the Veterans Cardiac Health and AI Model Predictions (V-CHAMPS) Challenge, including ZS. Our predictive AI model was a finalist in Phase 1 and exceeded expectations when validated on synthetic and real-world data (RWD) in Phase 2, earning us second place overall. The ZS team used its diverse experiences and skill sets to develop a high-performing, transparent and explainable AI model.

We met with our talented team to learn more about the innovative solution and success in the competition. “We feel proud of ourselves for having contributed to the advancement of science in chronic heart failure. We are excited that our work will help improve patient outcomes,” said Prerna Goel, a ZS data scientist and V-CHAMPS team member.

Where passion changes lives

The team featured these talented ZSers from our global offices: Pranava Goundan, Qin Ye, Abhinav Bansal, Sagar Madgi, Prakash Prakash, Subhrajit Samanta, Sayan Chakraborty, Hitashu Kanjani, Prerna Goel, Shubhankar Thakar, Vijay Boda and Xiaoying Shi. Each team member has a passion for improving healthcare outcomes for patients using AI and RWD. This includes developing algorithms to identify undiagnosed patients, expediting clinical trials, predicting disease progression and understanding barriers to treatment initiation. “Sayan is a tech and AI geek. Prerna has deep RWD, ML and healthcare domain expertise. Vijay and Hitashu are adept ML engineers. I’m proficient at designing end-to-end AI and ML systems and putting the pieces of puzzles together. Our skill sets added up to form a natural coalition,” said Shubhankar Thakar, ZS data scientist.

The team’s approach leveraged intelligent AI models to create impactful and clinically meaningful solutions. The ZS model captured the recency, frequency, occurrence and variability of critical events in patients' medical histories. Noteworthy aspects of our methodology included:

  • Intelligent feature discovery: ZS’s automated feature discovery tool uses a cloud-based genetic algorithm to create cross-sectional features that assess event aggregation and recency. 
  • Purposeful algorithmic experimentation and optimization: The team followed good ML practices to build transparent, credible, interpretable and generalizable algorithms.
  • Enhanced interpretability: Methods such as the Shapley additive explanations approach enhanced the understanding of the magnitude of feature importance and its directional impact on outcomes. Focused data wrangling and desk research across a vast array of literature helped establish a baseline understanding of key and novel risk factors, similar model design architectures, benchmark model performance KPIs and patient cohort definitions.

Throughout the challenge, what fueled the team’s motivation was each member’s drive to make a meaningful impact in healthcare. “Our passion for continuous learning and pushing boundaries in predictive AI kept us engaged and inspired. We were determined to leverage our skills and contribute positively to the health of veterans. This shared zeal for making a difference, coupled with our technical expertise, played a significant role in securing the second position,” Prerna said.

ZS congratulates team members for their remarkable achievement—a model that aims to enhance the health of veterans in the U.S. We’re excited to continue working with the FDA, VHA and others to advance research in this area by sharing new findings and methodologies from this challenge and envisioning scalable ways to use similar algorithms. The ZS team will explore ways to improve these solutions by ensuring wider patient representation and exploring more risk factors. We also plan to align multidisciplinary perspectives from pharma, payers, providers and patients toward solution development, validation and scalable implementation.

To stay tuned on how ZSers worldwide are driving a positive impact in healthcare and beyond, follow us on social media.

Life at ZS

Learn what it’s like to be a ZSer: LinkedInInstagram and Facebook 


View ZS’s open roles.

Add insights to your inbox

We’ll send you content you’ll want to read – and put to use.