Background: The incidence of CA-CDI is increasing. Patients admitted with CA-CDI increase the risk for subsequent hospital-onset CDI. Non-medical and environmental risk factors for patients with CA-CDI have not been well characterized.
Methods: This study summarizes the first use of GIS and spatial statistics to evaluate factors associated with acquisition of CA-CDI. The study population included all patients admitted to one of three study hospitals with CA-CDI from 2006-2011. Study hospitals included a tertiary care center (bed size 950) and two community hospitals (bed sizes 240 and 120). CA-CDI was defined used standard NHSN definitions. Only the first case of CA-CDI per patient was included. We used GIS and spatial statistics techniques to analyze transmission patterns and risks of CA-CDI. All patients had their residential addresses verified, geo-referenced, and set to a system of XY coordinates using ArcMap. Patients were excluded if home address geographic coordinates were not available. Data from the 2010 U.S. Population Census were used to determine socioeconomic status (SES). Getis-Ord procedures were used to detect statistically significant CA-CDI “hot spots”. Hot spots were mapped for population density and SES.
Results: A total of 312 episodes of CA-CDI met all inclusion criteria. Map 1A displays CA-CDI hot-spots involving 231 patients. While clusters of CA-CDI occurred more commonly in high population density areas than non-clustered cases (13% v. 1%, p<0.0001), 190 (82%) of clustered cases occurred in medium or low density areas. Patients in hot spots were on average older than non-clustered patients (mean age 60.2 v. 50.4, p<0.0001) and more commonly African American race (30% v. 14%, p=0.02). Map 1B displays CA-CDI hotspots in relation to SES. Lower SES was not associated with CA-CDI hot spots.
Conclusion: Our study represents the first use of GIS and spatial statistics to evaluate factors associated with CA-CDI. Numerous clusters were identified, particularly among older and African American patients. Future research with larger datasets will expand on this study to evaluate novel risk factors for CA-CDI, such as proximity to water or farmland.
L. F. Rojas,
L. P. Knelson, None
S. Pruitt, None
S. S. Lewis, None
R. W. Moehring, None
L. F. Chen, None
D. J. Sexton, None
D. Anderson, None
See more of: Oral Abstract Session