Measuring integrated team-based oncology care: Developing a novel chart abstraction tool. Issue 28 (1st October 2022)

• Record Type: Journal Article
• Authors: Richardson, Robin; Schear, Rebekkah; Munoz, Rebecca; Kvale, Elizabeth Ann
• Is Part Of: Journal of clinical oncology. Volume 40:Issue 28(2022)Supplement
• Page Start: 316
• Page End: 316
• Publication Date: 2022-10-01
• Subjects: 227-149-1069; 5; Oncology -- Periodicals; Cancer -- Periodicals; Oncology; Medical Oncology; Cancérologie -- Périodiques; Cancer -- Périodiques; Cancérologie; Cancer; Oncology; Oncologia; Càncer; Periodicals; 616.994
• DOI: 10.1200/JCO.2022.40.28_suppl.316
• Languages: English
• ISSNs: 0732-183X

Abstract

Abstract : 316 Background: In December 2018, the Livestrong Cancer Institutes (LCI) launched the CaLM Model of Whole-Person Cancer Care. Compared to traditional models of oncology care, CaLM offers integrated interdisciplinary supportive care. As team-based care delivery models like CaLM begin to become more predominant, health IT infrastructure and systems are failing to adequately capture and measure simple clinical data on team-based encounters. Athena, our EHR, lacked discrete fields to capture necessary data from shared appointments in order to evaluate patient utilization of team-based longitudinal care; requiring us to develop a manual chart abstraction to evaluate the clinical model. Methods: We performed a literature review and environmental scan on chart abstraction for team-based care. When limited results were found, we sought to develop a novel tool in REDCap. We developed a list of variables and domains that required measurement and worked with IT to determine which, if any, could be generated automatically in reports in the Athena EHR. Remaining domains (12 domains and 151 unique data points) were strategically embedded in a methodological abstraction tool that mapped closely to the logic and flow of the EHR. We tested the tool and iterated over three months in 2019. Before abstraction began, we developed and executed a two module, four hour training for data collectors to ensure quality data collection. Over a one year period, four abstractors pulled data from over 400 individual patient charts in the EHR. Results: We developed a unique two-phase quality assurance (QA) algorithm. Phase 1- We randomly selected 10% of patient records; two abstractors completed dual data entry (same chart). Dual data entry controls for discrepant values that can be common in large datasets. A third-party team reviewed both charts, recording discrepancies by category, harmonizing charts, and data was used to retrain data collectors during weekly meetings. Highest discrepancies occurred when documenting treatment history (95%). For 89% of records, there were differences in psychosocial utilization. As a result of these discrepancies, we opted to extend QA- the remaining 90% of records were reviewed by third party scanning half of domains for accuracy. The chart abstraction was successfully completed on over 400 patient charts; the process took over 18 months to complete. Conclusions: Analyzing patters of team-based care is likely a manual process, regardless of EHR vendor. Our tool is well-structured with the ability to be open source, so that other clinics can utilize our tool in lieu of replicating the same painstaking process.