Client success story
Drug Development Simulation
Business Value Delivered: Trial Modeling Gets Robust Statistical Basis
Summary: Statistical Model Selection For Trial Modeling
bPrescient worked with a company that developed AI tools for modeling virtual drug pipeline phases and financial impacts. bPrescient reviewed, selected and incorporated 3rd party tools for developing statistical models and performing simulations.
Challenge: Picking Right Statistical Tool From Available Options
The company wanted to incorporate public data from clinical trials and drug company finances, along with proprietary data from each customer, and then run simulations to apply that knowledge to a proposed drug development plan. A Monte Carlo simulation tool was needed to support the analysis but a large number of potential approaches existed, and picking the right one was a critical first step in developing the application.
Solution: Expert Review Of Tools And Requirements
bPrescient reviewed the client’s requirements for Monte Carlo simulation and reviewed existing publicly-available libraries, eventually electing to improve the client’s initial C# implementation. Additional exception handling and testing was implemented to improve robustness of the code, and changes were made to allow it to be run in the customer’s production pipelines, including the ability to run in a parallel context. Mathematical improvements were made, including improvements in the usage of a variety of statistical distributions. Performance was assessed to ensure the library would perform acceptably, and the foundation for an advanced use case was laid.
Outcome: Optimal Model Selected
The customer implemented their simulation pipelines using the selected and enhanced Monte Carlo library and was able to develop and test initial versions of their complete software application.
Let’s talk
Tell us about your toughest R&D challenge and let’s discuss how bPrescient can solve it.
