How do you screen thousands of abstracts with transparency and reproducibility using minimal training data?
In this case study, MadeAi partnered with the Roche HEOR team to design and validate an explainable AI pipeline for title and abstract screening in a systematic literature review. Using a structured PICO-based framework, the team screened over 3,300 abstracts across multiple datasets and generated clear inclusion and exclusion criteria, ensuring reproducibility and transparency aligned with publication-grade standards.
The approach aligns with methodologies reflected in a Value in Health Journal publication, demonstrating how explainable, structured workflows support rigorous evidence generation.
See how AI-driven screening, SME validation, and research-ready outputs came together to enable faster, reliable evidence generation.
