Hospital emergency department saturation diversion data and influenza surveillance data: Statistical analysis provides evidence for correlation

Objectives: Previous studies at the California state level have suggested correlation between emergency department saturation data and admissions for pneumonia and influenza. Seasonal trends in Orange County (OC) Rapid Emergency Digital Data Information Network (ReddiNet) emergency department saturation diversion data appear anecdotally to correlate with influenza activity but no formal evaluation in OC had been conducted. Evaluating the ReddiNet data against laboratory influenza reports and influenza-like illness trends will help to establish data reliability, thus allowing for monitoring of trends in influenza activity using near real-time hospital-based data.

Methods: We conducted a time-series cross-correlation analysis using SPSS 15.0 statistical software to compare diversion data to influenza surveillance data by CDC disease week; a correlation coefficient of 0.60 or greater was considered high for our purposes. Number of diversion hours was compared to the number of laboratory-positive influenza reports received from disease week 41 of 2003 to week 39 of 2007, and to sentinel provider influenza-like illness (ILI) reports received from disease week 40 of 2004 to week 39 of 2007.

Results: A correlation coefficient of 0.71 was generated when comparing diversion data to laboratory-positive influenza reports and a correlation of 0.60 when diversion data was compared to ILI reports.

Discussion: We found high positive correlations for diversion data versus influenza surveillance data. We concluded that ReddiNet diversion data is a reliable source for monitoring influenza activity within Orange County.