Quality Improvement Project: Workflow, Standardization and Data Validation of Comorbidity Index at a Small to Medium Volume Transplant Center

The Hematopoietic Cell Transplant Comorbidity Index (HCT-CI) developed in Seattle predicts morbidity and mortality due to hematopoietic stem cell transplant (HSCT) and is relevant to both patient outcome management and data reporting. At Saint Louis University (SLU), we performed a retrospective analysis to inform a process integrating pre-transplant workflows with rigorous collection and validation of HCT-CI data for a new analytic database. Herein we report data process corrections and quality improvement strategy for pre-transplant comorbidity assessment. 101 transplants between 3/2014 and 3/2016 were queried for any reference to “HCT-CI”, “Sorror Score” or “comorbidity index”. Data and reported scores from a variety of disparate source documents were extracted into a spreadsheet that recapitulated HCT-CI data fields. We analyzed score variances from data sources and re-assessed each chart according to the most recent HCT-CI model. Data was analyzed for elements which did not conform (i.e: outside ideal time window, etc.) Of 101 transplants, 81 had sufficient data for analysis. In 36% (Figure 1) scores were consistent across all source documents, 40% varied by a score of 1, and 25% varied by >1. Of the 37 charts reassessed using the current published method (Figure 2A), 35% conformed and were concordant with prior scoring, 8% (3/37) were scored but were inconsistent with prior data, and 57% could not be scored. The primary reason (Figure 2B) for non-scoring was hepatic (54%) and renal (30%) data outside the correct date range.Figure 2Retrospective analysis: concordance and incomplete data.View Large Image Figure ViewerDownload Hi-res image Download (PPT) A modified workflow (Figure 3) is intended to improve accuracy of HCT-CI scoring. Our ongoing intervention includes: program education based on retrospective data; use of an online scoring tool to resolve inconsistencies; and integration of HCT-CI scoring into our real time validation process previously published. Only post validated data is placed into the core SQL database as shown (Figure 3). As a core programmatic function, prognostication is central to individual patient care decisions and predicting outcome. Accuracy in pre-transplant comorbidity risk assessment facilitates program development goals by allowing informed protocol modifications. This audit driven analysis/correction of our HCT-CI workflow initiates a precise process for scoring and reporting of comorbidities. We expect our novel quality improvement process to provide a superlative data management tool for future quality improvement initiatives in service of our patients, as we have recently demonstrated in the management of acute graft vs. host disease (Ravulapati et al, 2016).