This article is more than 5 years old. Information may no longer be current.
Data mining EHRs can rapidly identify hospital outbreaks
Click Here to Manage Email Alerts
Data mining of an electronic health record, or EHR, accurately identified transmission routes among patients involved in hospital disease outbreaks, according to findings from a retrospective analysis.
“Traditional outbreak detection in hospitals typically involves manual review of the EHR to identify the transmission route,” Lee H. Harrison, MD, professor of medicine and epidemiology at the University of Pittsburgh, told Infectious Disease News. “Data mining of the EHR provides a novel and enhanced approach for more rapid identification of the outbreak transmission route.”
Writing in Infection Control & Hospital Epidemiology, Harrison and colleagues said they are currently developing the “Enhanced Detection System for Healthcare-Associated Transmission,” a new surveillance system that combines prospective whole genome sequencing surveillance with data mining of EHRs. Harrison said this method has the “potential to speed up investigation and interruption of outbreaks.”
For their study, Harrison and colleagues retrospectively analyzed nine hospital outbreaks that occurred between 2011 and 2016 and had been categorized according to transmission route and molecular characterization of the bacterial isolates, they explained. They assessed the ability of EHR data mining to identify the correct route of transmission, how early in the outbreak the correct route was identified, and how many cases could potentially have been prevented.
In eight of the outbreaks, correct transmission routes were identified with the second patient. One outbreak involved more than one transmission route, and the correct route was not detected until the eighth patient, Harrison and colleagues reported.
According to the study, assuming an effective intervention was implemented within 7 days of the transmission route detection, 78% of possible preventable infections could have been averted if data mining was implemented in real time. Real time data mining could have averted 66% of possible preventable infections if an intervention was implemented within 14 days of identifying the transmission route, according to Harrison and colleagues.
“Data mining tools can be applied to the electronic health record to rapidly identify the transmission route or routes that are responsible for a hospital outbreak,” Harrison said. “We are now evaluating the use of whole-genome sequencing surveillance of key hospital pathogens in combination with EHR data mining for outbreak detection and investigation, respectively.” – by Marley Ghizzone
Disclosures: The authors report no relevant financial disclosures.
Perspective
Back to Top
Bernard C. Camins, MD, MSc
Data mining of EHRs, in conjunction with whole-genome sequencing (WGS), can potentially identify health care-associated outbreaks much more rapidly than current methods. By accessing EHRs, hospital epidemiologists can rapidly identify common exposures that are shared by patients infected by the same pathogen. This new method accurately identified the correct routes of transmission when applied to previous outbreaks. If routes of transmission are identified, then further spread of infections can be avoided. The authors estimated that this new surveillance system could have prevented at least two-thirds of outbreak-related health care-associated infections.
Through data mining of EHRs and molecular typing (WGS), outbreaks can be identified rapidly with minimal manpower by infection prevention personnel with limited experience in outbreak investigation.
This study retrospectively analyzed EHR data and successfully demonstrated a proof of concept. The same investigators will have to demonstrate the feasibility of this method with real-time outbreaks. Unfortunately, WGS is not available in the majority of hospitals. The method may not be able to identify an outbreak if there is more than one mode of transmission.
Click Here to Manage Email Alerts