Electronic Method to Improve Identification of Salmonella Clusters - Nebraska, 2012

BACKGROUND:  Prior to 2013, the Nebraska Department of Health and Human Services (NDHHS) faced challenges in timely identification of PFGE-matched salmonellosis clusters. NDHHS uses the National Electronic Disease Surveillance System (NEDSS) for receiving electronic laboratory reports (ELR) from laboratories throughout the state including Nebraska Public Health Laboratory (NPHL) salmonella serotype and PFGE results. Technical message format issues delayed review of NPHL-reported serotype data until the corresponding PFGE results were forwarded. This precluded timely patient follow-up and thus compromised exposure recall. Objectives were 1) to eliminate delays in sharing laboratory epi marker tests with epidemiologists; 2) to create an analyzable dataset of all salmonella laboratory tests submitted through NEDSS, including serotype/ PFGE data; 3) to identify all serotype/PFGE clusters for 2012; and 4) to use these findings to enhance future cluster identification.

METHODS:  NPHL modified ELR processing and message formatting to decouple serotype and PFGE results to ensure immediate receipt of reports upon sequential completion of testing (e.g. culture, serotype, and PFGE). SAS code was developed to extract a dataset of all 2012 Nebraska salmonella results for analysis. Comparisons were done to determine: 1) total number of salmonellosis cases; 2) percent of salmonella isolates that reached NPHL; 3) percent of isolates with serotype and PFGE results. We retrospectively determined the number of 2012 clusters and an interview-completion rate for cluster-associated cases. A cluster was defined as two or more salmonellosis cases with matching PFGE patterns and illness onset no greater than 2 weeks apart.

RESULTS:  ELR shortcomings were addressed and a line-level serotype dataset was developed. During 2012, 344 salmonellosis cases were identified; 268 (77.9%) reached NPHL; 27 (7.8%) reported serotypes were from out of state laboratories and 49 (14.3%) did not report serotype or PFGE. Among these, we retrospectively identified 32 clusters involving 94 patients representing 11 and 26 unique serotype and PFGE patterns, respectively. Of the 94 cluster-associated cases, 54 (57.4%) interviews were completed.

CONCLUSIONS:  ELR shortcomings were corrected to facilitate timely receipt of reports and case follow-up to improve salmonella surveillance. We retrospectively identified 32 salmonellosis clusters which occurred in Nebraska during 2012. Future efforts include identification of clusters in real time, development of analytic tools for improved cluster detection, process evaluations and timely patient follow-up to identify common exposures.