Online Program

Reproducible Performance Report Generation to Provide an Approach for Communicating Quality Metrics Longitudinally to Wisconsin Health Care Providers (304194)

Tudor Borza, University of Wisconsin - Department of Urology 
Caprice Greenberg, University of Wisconsin - Department of Surgery 
Bret Hanlon, University of Wisconsin - Department of Surgery 
Jonathan Kohler, University of Wisconsin - Department of Surgery 
Elise Lawson, University of Wisconsin - Department of Surgery 
*Nicholas Marka, University of Wisconsin - Department of Surgery 
Sudah Pavuluri Quamme, University of Wisconsin - Department of Surgery 
Jessica Schumacher, University of Wisconsin - Department of Surgery 
Manasa Venkatesh, University of Wisconsin - Department of Surgery 
Dou-Yan Yang, University of Wisconsin - Department of Surgery 

Keywords: Reproducible research, R programming, Knitr, Health services research, Performance evaluation

The Surgical Collaborative of Wisconsin (SCW) is a statewide practice change community of surgeons and quality leaders engaged in initiatives to improve surgical quality. SCW’s initiatives rely on accurate, individualized, timely, and risk-adjusted surgeon- and hospital-level performance reports to guide efforts. These reports provide comparative benchmarks every 6 months with updated data from centralized data warehouses.

This work highlights an automated process for estimating and building reports employing R with knitr. With our software pipeline, we can synthesize disparate data sources (e.g. claims, hospital discharge data). Automating processes in R allows for improved efficiency, reduced error, volume checking, and flexibility to add features as interventions are implemented and information needs change. In Nov 2019, our team generated reports for 3 initiatives and 1178 providers across 276 facilities.

New initiatives (e.g. low volume reports) that require varied statistical estimation approaches and increasing SCW membership will increase the type and volume of reports needed. A key challenge is reducing computational bottlenecks to improve scalability.