Improving Reproducibility in Experimental Research by Using Shared Data -A Use Case in Spinal Cord Injury

Abstract Robust and reproducible treatment effects are critical for clinical impact. However, current biomedical research has limited reproducibility, ultimately stalling medical progress. We propose combining collaborative data sharing with novel statistical methods to improve inferential robustness and experimental intervention evidence to accelerate translation. We will do so in the context of spinal cord injury, a leading cause of long-term disability, by 3 aims: 1) Improve inferential robustness of experimental treatments through advanced analytics and historical borrowing; 2) Accelerate translation through a "platform pre-clinical trial" to efficiently prioritize treatments and uncover determinates of treatment response; 3) Create Open-Source data tools that are usable by the broader research community to improve reproducibility and accelerate translation.