Automated detection tool improves nosocomial outbreak identification

SAN DIEGO — The use of an automated outbreak detection tool may be able to identify previously undetected nosocomial outbreaks and more quickly spot those identified using traditional methods, according to results of a presentation at IDWeek 2015.

“Detecting outbreaks can be challenging,” Meghan Baker, MD, ScD, of Harvard Medical School, said during her presentation. “Detection methods often rely on temporal or spatial clustering, [or] on case counting or subjective judgment to determine whether or not a cluster actually exists. It can be time consuming and labor intensive … and because clusters, whether they’re perceived or real, often require follow-up, false clusters can waste valuable resources.”

Using the Premier SafetySurveillor tool, Baker and colleagues applied a multidimensional scan statistic to 83 years of historical microbiology data collected from 43 hospitals. Clusters of cultures yielding pathogenic species were identified, with researchers including those involving three or more patients and occurring by chance less than once per year. These data were compared with outbreaks identified by a convenience sample of hospitals using standard methods. In addition, infection preventionists from included hospitals were surveyed for opinions on the automated tool as a resource for outbreak detection.

There were 230 clusters identified by the automated tool, resulting in a mean frequency of one cluster per 100 beds. Most clusters were detected based on resistance patterns, with others identified by common ward or specialty service. Eighty-nine percent of the clusters identified were not detected by the hospital’s routine procedures, and 12 of the 15 clusters identified by the detection software and hospitals were detected earlier by the automated tool.

Eighty-one percent of infection preventionists said they would have wanted notification of the cluster in question, 47% of which the preventionists said were either moderately or highly concerning. In addition, 76% felt that the software would have moderately or greatly improved their ability to detect outbreaks, and 26 out of 29 hospitals were “pleased” to expand surveillance.

“Automated outbreak detection can increase the efficiency and scope of identification of nosocomial clusters, including pathogens not under routine surveillance,” Baker and colleagues wrote. – by Dave Muoio

Reference:

Baker M, et al. Abstract 637. Presented at: IDWeek; Oct. 7-11, 2015; San Diego.

Disclosure: Baker reports no relevant financial disclosures. Please see the full study for a list of all other authors’ relevant financial disclosures.