Assessment of Clostridium difficile-associated disease surveillance definitions, North Carolina, 2005

objective. To determine the timing of community-onset Clostridium difficile-associated disease (CDAD) relative to the patient's last healthcare facility discharge, the association of postdischarge cases with healthcare facility-onset cases, and the influence of postdischarge cases on overall rates and interhospital comparison of rates of CDAD. Design. Retrospective cohort study for the period January 1, 2005, through December 31, 2005. Setting. Catchment areas of 6 acute care hospitals in North Carolina. Methods. We reviewed medical and laboratory records to determine the date of symptom onset, the dates of hospitalization, and stool C. difficile toxin assay results for patients with CDAD who had diarrhea and positive toxin-assay results. Cases were classified as healthcare facility-onset if they were diagnosed more than 48 hours after admission. Cases were defined ascommunity-onset if they were diagnosed in the community or within 48 hours after admission, and were also classified on the basis of the time since the last discharge: if within 4 weeks, community-onset, healthcare facility-associated (CO-HCFA); if 4-12 weeks, indeterminate exposure; and if more than 12 weeks, community-associated. Pearson's correlation coefficient was used to assess the association between monthly rates of healthcare facility onset, healthcare facility-associated (HO-HCFA) cases and CO-HCFA cases. We performed interhospital rate comparisons using HO-HCFA cases only and using both HO-HCFA and CO-HCFA cases. Results. Of 1046 CDAD cases, 442 (42%) were HO-HCFA cases and 604 (58%) were community-onset cases. Of the 604 community-onset cases, 94 (15%) were CO-HCFA, 40 (7%) were of indeterminate exposure, and 208 (34%) community-associated. A modest correlation was found between monthly rates of HO-HCFA cases and CO-HCFA cases across the 6 hospitals (r=0.63, P<.001). Interhospital rankings changed for 6 of 11 months if CO- HCFA cases were included. Conclusions. A substantial proportion of community-onset cases of CDAD occur less than 4 weeks after discharge from a healthcare facility, and inclusion of CO-HCFA cases influences interhospital comparisons. Our findings support the use of a proposed definition of healthcare facility-associated CDAD that includes cases that occur within 4 weeks after discharge.