Methods: We conducted this study in a 2-hospital, >1100-bed community-based academic healthcare system in northern Delaware. We created a CDI-specific analytic application using the Health Catalyst analytics platform, over the existing data warehouse (Cerner), using 2016-2018 data (Figure). The application refreshes daily and is able to provide near real-time patient data, including PPI and antibiotic use. We aimed to describe current PPI utilization patterns, calculate risk associated with PPI use adjusted for other risk factors for CDI, and measure the effect of interventions to decrease PPI use.
Results: Among 133,592 total inpatient encounters from 1/1/16 – 4/22/18, 39,156 (29%) received PPIs and 1146 (0.9%) had a positive PCR result for C. diff. Among the C. diff positive encounters, PPIs were used in 486 (42%), with an adjusted OR of 2.1 (95% CI 1.7-2.6). Of encounters involving high risk antibiotics who had a positive C. diff PCR, 52% (255/486) were receiving PPIs. The services most likely to prescribe PPIs were internal medicine, orthopedic surgery and general surgery. Targeted chart review indicated that most inpatients receiving PPIs lacked an identified upper gastrointestinal (GI) disorder, and 37% were on the same PPI as outpatients prior to admission. Duration of therapy varied widely, but PPI courses were longer in patients diagnosed with CDI.
Conclusion: A novel application using existing health record data confirmed the increased risk of CDI due to PPI use, and identified important opportunities to decrease HO-CDI by limiting such use. Using this analytics platform provides near real-time data and will support rapid cycle improvements and allow for early evaluation of CDI interventions.
D. Walsh, None
N. Harrington, None
M. Winiarz, None
A. Gabriel, None
E. F. Ewen, None