How do they stack up? Using NORS, PulseNet, and NARMS data to estimate the number of shigellosis outbreaks in the United States, 2009 –2013

BACKGROUND: Shigellosis is the third most common enteric bacterial infection in the United States, with approximately 500,000 infections per year. Although shigellosis is a nationally notifiable condition, the burden of shigellosis outbreaks has not been well described. Using a capture-recapture approach, we aimed to assess the number of shigellosis outbreaks that occurred in the United States from 2009 –2013. 

METHODS: Data were obtained from three surveillance systems: the National Outbreak Reporting System (NORS), which reports epidemiologically-defined outbreaks; the National Antimicrobial Resistance Monitoring System (NARMS), which intends to test isolates from all outbreaks; and PulseNet, which identifies clusters of genetically similar isolates that may not be epidemiologically related. To enable capture-recapture analysis, multistate outbreaks were reclassified as an outbreak in each state with cases and multi-residency outbreaks as single-state outbreaks in the states where exposure occurred. We linked outbreaks reported in multiple surveillance systems by comparing jurisdiction and onset date, as well as PulseNet cluster code and pulsed-field gel electrophoresis (PFGE) pattern when available.   We used a log-linear model to calculate the unbiased estimate of the total number of shigellosis outbreaks.

RESULTS: Overall, 363 single-state and 17 multistate outbreaks were reported among the three systems, which were recoded as a total of 468 single-state outbreaks.  NORS reported 316 single-state outbreaks, PulseNet reported 14 single-state and 14 multistate clusters, and NARMS tested isolates from 33 single-state and 3 multistate outbreaks.Multistate clusters captured by PulseNet were often reported as multiple single-state outbreaks in NORS.  PulseNet captured the highest number of multistate clusters, while NORS captured the highest number of single-state outbreaks. Isolates from 46/316 (15%) NORS outbreaks and 27/96 (28%) PulseNet clusters were tested by NARMS.  The unbiased estimate of the total number of shigellosis outbreaks is approximately 3 times greater than the number captured by any of the three systems from 2009 –2013.  

CONCLUSIONS: The total number of shigellosis outbreaks in the United States each year may be 3 times greater than currently reported in surveillance systems. Clearly defining outbreaks and linking common events across systems was difficult. In accordance with their intended purposes, NORS and PulseNet differed in their capacity to detect single- and multistate outbreaks.  Isolates from a small proportion of outbreaks are submitted to NARMS. Additional efforts should be undertaken to link data in different surveillance systems, facilitate reporting, and prevent and mitigate shigellosis outbreaks.