USING R TO ADVANCE PATIENT-CENTRIC RESEARCH IN THE STUDY OF A PROSPECTIVE ADULT RESEARCH COHORT. (26th January 2023)

• Record Type: Journal Article
• Authors: Fehlmann, Tara; Harvey, Sarah; Deepak, Parakkal; O'Charoen, Sirimon
• Is Part Of: Inflammatory bowel diseases. Volume 29(2023)Supplement 1
• Page Start: S23
• Page End: S23
• Publication Date: 2023-01-26
• Subjects: Inflammatory bowel diseases -- Periodicals; Colitis, Ulcerative -- Periodicals; Crohn Disease -- Periodicals; Inflammatory Bowel Diseases -- Periodicals; 616.344
• DOI: 10.1093/ibd/izac247.043
• Languages: English
• ISSNs: 1078-0998

Abstract

Abstract: OBJECTIVE: IBD Plexus is a data exchange platform that houses four cohorts, including The Study of a Prospective Adult Research Cohort with IBD (SPARC IBD). SPARC IBD is a multicentered, longitudinal cohort that follows patients through their standard of care. In IBD Plexus, patient reported data (eCRF) is linked with clinical and electronic medical record (EMR) data. Biospecimens are continuously collected and have been used to generate a myriad of molecular and biomarker data. To facilitate inflammatory bowel disease (IBD) research, we built an R package 'ibdplexus' (https://github.com/ccf-tfehlmann/ibdplexus ), with multiple functions to load, filter and subset the data in IBD Plexus. Currently, this R package focuses primarily on SPARC IBD, with a future goal to expand to incorporate other studies. Each function is designed to help understand an IBD patient's journey including their treatment use and disease activity over time. PRIMARY FUNCTIONS: Three main functions help navigate the prescription information from the eCRF and structured EMR data. 'Sparc_medication_filter' filters the prescription data to generate a dataset with medication(s) of interest from eCRF and EMR data. For long-term medications including biologics, aminosalicylates, and immunomodulators, `sparc_medication_starts' extracts the first instance of a medication start date from eCRF and EMR data. Sparc_medication_at_index' generates a dataset with each participant and their medication at a specified time-point (e.g., enrollment, at time of biosample collection, etc.). Another set of functions map out a patient's disease activity over the course of their enrollment. 'Calculate_disease_scores' creates a list of dataframes and an optional spreadsheet containing multiple tabs with disease activity and endoscopy scores for participants. Calculations include the short Crohn's Disease activity index and the patient reported outcome (PRO) derivations PRO2 and PRO3, as well as the simple endoscopic score for Crohn's Disease. For ulcerative colitis patients, the disease activity measures include the 6 and 9-point Mayo score and Mayo endoscopy score. 'Sparc_scores' generates a dataset with each SPARC participant and their disease activity scores at a specified time-point. CONCLUSION: The ibdplexus R package has increased the speed and efficiency with which the SPARC IBD data can utilized for research projects. Twenty domains can now be easily collapsed into one table mapped to a date of interest. These reference tables are then used to complete multiple queries, reducing the time per query from up to a day to an hour. The results from the medication functions are well aligned with a subset of patients that underwent chart review. In the few cases where the medication at index date did not align, the information was only available in progress notes or other forms of unstructured EMR data.