Developing a cohesive grammar to utilize the suite of packages available in HADES and the OMOP Methods Library

‍

Motivation

Healthcare data science teams increasingly rely on the OMOP Common Data Model and HADES (Health Analytics Data-to-Evidence Suite) to conduct robust real-world evidence studies. HADES represents a comprehensive collection of open source R packages for large scale analytics, including population characterization, population-level causal effect estimation, and patient-level prediction. With over 20 individual packages offering extensive analytical capabilities, teams sometimes need additional support to navigate and utilize this rich ecosystem cohesively for their specific use cases. Organizations often benefit from supplementary frameworks that help bridge their particular analytical requirements with HADES' powerful existing functionality.

Solution

Plinth developed a comprehensive assessment framework to evaluate how HADES capabilities could be optimally aligned with real-world data science requirements. We conducted a systematic analysis across 23 measurement capabilities and over 100 study-specific requirements, identifying opportunities to enhance 65% of core analytical workflows through custom extensions and integrations built on HADES foundations. Our team then designed a cohesive grammar and workflow framework that leverages these capabilities while maintaining full compatibility with the broader HADES ecosystem. The solution included enhanced usability patterns, standardized analytical pipelines, and extensible architecture that allows teams to build upon HADES' robust foundations while addressing their specific analytical needs.

Impact

Our HADES enhancement framework enabled the client to reduce time-to-insight for cohort generation and analysis by an estimated 60% while maintaining the scientific rigor and reproducibility standards that make HADES so valuable. The cohesive grammar we developed allows data scientists to seamlessly transition between different HADES packages without learning disparate interfaces, significantly reducing onboarding time for new team members. Most importantly, our approach ensures teams can leverage the full power of the OMOP ecosystem and HADES' extensive capabilities while building sustainable, extensible analytical workflows that evolve with their research needs.

‍