ALGORITHM DEVELOPMENT IN R TO CATEGORIZE RESPONSE TO ADVANCED THERAPUTICS IN THE STUDY OF A PROSPECTIVE ADULT RESEARCH COHORT WITH IBD

Abstract Objective IBD Plexus is a research tool that houses multiple datasets, including The Study of a Prospective Adult Research Cohort with Inflammatory Bowel Disease (SPARC IBD). SPARC IBD is a multicentered, longitudinal cohort that follows patients through their standard of care. Within IBD Plexus, patient-reported data (eCRF) are linked with clinical, electronic medical record (EMR), and multi-omics data. Despite endoscopic remission being a central outcome of interest when assessing treatment efficacy, there is currently no consensus definition when using real-world data, highlighting the need for a consistent framework. To address the need for standardized definitions of treatment response, we developed an algorithm implemented through an R function, tx_response, which compiles, cleans, and organizes available clinical and laboratory outcomes to help evaluate a patient’s response to therapy. Algorithm Development The tx_response R function integrates outcomes from both eCRF and EMR sources for patients on advanced therapies which includes biologics and targeted small molecules. Available measures include endoscopic scores (Mayo Endoscopic Score, Simple Endoscopic Score for Crohn’s Disease), symptom scores (6-Point Ulcerative Colitis Disease Activity Index, Short Crohn’s Disease Activity Index), physician global assessments (PGA), fecal calprotectin (FCal), and C-reactive protein (CRP) or high-sensitivity C-reactive protein (hsCRP). Each measure was assigned threshold values to categorize active versus inactive disease. Clinical outcomes are classified as Remission or Non Remission, while laboratory outcomes are designated as Active or Inactive. The various treatment response measurements are ordered by clinical rigorousness, with endoscopic scores being considered the most rigorous measurement, followed by symptom scores, PGA, FCal and lastly CRP or hsCRP. The function generates an Excel file with three worksheets: ALL SCORES, which contains all available clinical outcomes for each patient-medication combination; CLOSEST SCORE, which identifies the outcome nearest to a user-specified timepoint after medication start; and MOST RIGOROUS, which reports only the highest-priority outcome available. Additional parameters allow researchers to filter specific medications or restrict outcomes to a defined time window after medication start date. CONCLUSION The tx_response function, now available in the ibdplexus R package (https://github.com/ccf-tfehlmann/ibdplexus), establishes a framework for defining remission and non-remission post initiation of an advanced therapy in real-world SPARC IBD data. By leveraging a range of outcomes and ranking them by rigor, this approach maximizes the data available for analysis while prioritizing the most clinically meaningful measures.