Developing a Standardized Process to Visualize, Analyze, and Communicate NSQIP Data Using an Advanced Visual Data Analytics Tool.

Abstract

Background: To help surgeons improve quality, the American College of Surgeons National Quality Improvement Program (ACS NSQIP) Semiannual Reports and Interim Semiannual Reports provide high-level views of 30-day morbidity and mortality rates. Surgeons at one hospital requested the ability to visualize data with interactive navigation and analysis of comorbidities monthly. Using advanced visual data analytics, the authors constructed a surgical scorecard to provide the desired feedback.

Methods: The authors undertook a proof-of-concept project tracking surgical site infections (SSIs) and associated medical comorbidities. An anonymized training dataset of 3,438 patients was sampled between January 1, 2021, and October 31, 2022, from the hospital's NSQIP data. For proof-of-concept interface/system testing and to maintain data privacy, a synthetic 5,000-patient NSQIP database was generated using the Synthetic Data Vault, Python 3.7. Comorbidity variables were: diabetes mellitus, HgbA1c, immunosuppressive therapy, hypertension requiring medication, body mass index, and smoking within one year. The primary outcome was SSI. The research team generated scorecards for SSIs as a function of time, surgical department, and medical comorbidity. Odds ratios with confidence intervals and chi-square tests were used to analyze the relationships between SSI and comorbidities.

Results: Advanced visual data analytics improved the timeliness of NSQIP Semiannual Reports and Interim Semiannual Reports from 6 months to 45 days. The scorecard allowed for visualization of data trends as a function of time, specialty, and procedural group. Statistical testing allowed for the identification of surgeons who were statistical outliers with regard to SSIs.

Conclusion: Implementation of an on-demand scorecard for data visualization and analysis allowed for up-to-date analysis of the relationship between medical comorbidities and SSI and identification of performance outliers.