Creating a Centralized Surveillance Program using Core Measures Data Abstractors

Abstract

BACKGROUND Healthcare facilities are working to make Infection Prevention departments more efficient and cost-effective. One approach is implementation of centralized surveillance programs (CSP) allowing teams to specialize in applying healthcare-associated infection (HAI) surveillance definitions. However, staffing CSPs with Infection Preventionists (IP) may be difficult due to national shortages and cost. This study evaluates the use of data abstractors (DA) to perform initial HAI surveillance as a more cost-effective option. METHODS A five-hospital healthcare system with more than 1800 acute care beds began a CSP in May 2018. Two full-time DAs with prior Core Measures abstraction experience were hired to conduct all HAI surveillance. A 12-week training program was developed to apply the National Healthcare Safety Network (NHSN) HAI protocols. Training consisted of video modules, case studies, supervised practice with a certified IP, and a post-test for each HAI protocol. Additional training was performed when post-test performance was less than 85%. A follow-up inter-rater reliability (IRR) program was used to validate post-training competency. All infections were confirmed by an IP. RESULTS Following the initial 12-week training program, DAs achieved IRR of higher than 90% for catheter-associated urinary tract infection, central line-associated bloodstream infection, surgical site infection (SSI), and ventilator-associated event (VAE) surveillance. Prior to implementation of the CSP, IP time constraints did not allow for daily review ofinternal reports designed to identify VAEs. An incidental finding following CSP implementation was an increase in the reported number of VAEs due to more comprehensive surveillance. CONCLUSIONS Professionals with experience in Core Measures data abstraction are highly-skilled in the application of standardized definitions. When considering implementation of a CSP, staffing with DAs rather than IPs may be a more cost-effective option while still maintaining high data quality.