Motivation

90% of all analytics projects start with the question “How many people are in our database that satisfy a specific set of clinical and observational criteria?” Can you help us find a way to reduce the amount of time it takes us to answer these questions from weeks to minutes?

Key Results

Challenge

• An industry leading provider of patient-centered real world data, needed tools to generate bespoke patient cohorts.
• Cohort specifications required both disease-specific and disease-independent criteria.
• Both coders (quantitative scientists and engineers) and non-coders (commercial and marketing) users.
• Must ensure high levels of data quality and appropriateness for any given research project.

Approach

• Met with stakeholders to identify scope in terms of usability, speed, and functionality.
• Surveyed the team’s code base to understand what existing work could be leveraged.
• Designed, prototyped, and iterated solutions based on cross-functional team feedback.

Deliverables

• An R package providing a simple code-first interface for users to generate a bespoke one-row-per-patient cohorts.
• A web-based no-code dashboard enabling point-and-click access to key package features.
• User documentation, end-to-end templates and guides and a company-wide demonstration of the tools’ capabilities.

Business Impact

• The quantitative sciences team was freed from hundreds of hours of impending development time.
• Teams were enabled to quickly generate answers to client questions live during client meetings that would have otherwise required days of resourcing.
• A vision for a new future client-facing product leveraging the “RWD cohort builder” was born.