![[ Visit Client Website ]](images/banner.jpg)
BACKGROUND: Early outbreak identification presents an important challenge to health departments. Implementation of aberration detection modeling methods to complement normal surveillance activities may be useful for this purpose. The Utah Department of Health (UDOH) initiated an aberration detection program and began monitoring aberrations on all of Utah’s reportable communicable diseases and conditions in January 2013. The results from a retrospective analysis using this new aberration detection system on influenza-associated hospitalization data from the 2012-2013 influenza season are presented.
METHODS: Since 2009, Utah has mandated reporting of all cases of influenza-associated hospitalizations. Positive laboratory reports are sent electronically and entered into the UDOH communicable disease surveillance database, TriSano®. Data are exported daily from TriSano®and analyzed using the R surveillance package Farrington and Negative Binomial CUSUM aberration detection algorithms. Influenza-associated hospitalization data was modeled weekly from January 2012 through March 2013. Alarm signals were generated based on baseline disease rates utilizing training data from 2009 through 2011.
RESULTS: From October 2012 through March 2013, a total of 925 of influenza-associated hospitalizations were reported in Utah. The season peaked in January, with 439 cases reported that month. By contrast, only 357 cases were reported during the 2011-2012 influenza season which peaked in March with 184 reported cases. Although only eight cases of influenza-associated hospitalization were reported in Utah in October 2012, both aberration models detected signals during this month. The Farrington model was more effective and detected an aberration on October 8, 2012. The Negative Binomial CUSUM model detected an aberration in hospitalized influenza cases as early as October 29, 2012. The CDC did not report increased influenza activity until MMWR week 44 (October 28-November 3, 2012).
CONCLUSIONS: The aberration detection system could have alerted UDOH to an increase in influenza-associated hospitalizations several weeks earlier and would have been useful to our public health efforts to ramp up vaccination campaigns and alert local health department personnel to meet the increased burden. The aberration detection system will continue to be utilized on a daily basis for reportable communicable diseases in Utah. Future projects include simulation modeling, investigation and addition of other aberration algorithms, and development of department policies and procedures regarding response to aberrations detected by the system. The system was easy to implement and maintain and thus may be appealing to other state and local health departments.