PITTSBURGH, November 17, 2021 – By coupling machine learning with whole genome sequencing, University of Pittsburgh School of Medicine and Carnegie Mellon University scientists have dramatically improved the rapid detection of infectious disease outbreaks in hospitals compared to traditional methods of monitoring outbreaks.
The results, published today in the journal Clinical infectious diseases, point to a way for health systems to identify and then stop infectious disease outbreaks in hospitals, reducing costs and saving lives.
“The current method used by hospitals to detect and stop the transmission of infectious diseases in patients is obsolete. These practices have not changed significantly in over a century, ”said senior author Lee Harrison, MD, professor of infectious diseases at Pitt’s School of Medicine and epidemiology at the Graduate School of Public Health, Pitt. “Our process detects significant outbreaks that would otherwise go under the radar of traditional infection control surveillance. “
The Enhanced Detection System for Healthcare Associated Transmission (EDS-HAT) combines recent development in affordable genomic sequencing with computer algorithms connected to the vast mine of data contained in electronic health records. When sequencing detects that two or more patients in a hospital have nearly identical infection strains, machine learning quickly pulls those patients’ electronic health records for commonalities – whether near hospital beds. , a procedure using the same equipment or a shared healthcare provider – alerting infection control specialists to investigate and stop transmission.
Normally, this process requires clinicians to notice that two or more patients have a similar infection and alert their infection prevention team, who can then review patient charts to try to find how the infection was transmitted.
“This is an incredibly laborious process that often depends on busy healthcare workers noticing a shared infection between patients to begin with,” said lead author Alexander Sundermann, MPH, CIC, FAPIC, research coordinator clinical and doctoral student at Pitt Public. Health. “It might work if the patients are in the same unit of a hospital, but if those patients are in different units with different healthcare teams and the only shared link was a visit to a procedure room,” the chances of this outbreak being detected before other patients are infected drops dramatically.
From November 2016 to November 2018, UPMC Presbyterian Hospital performed EDS-HAT with a six-month lag for a few selected infectious pathogens often associated with hospital-acquired infections nationwide, while continuing with traditional methods of real-time infection prevention. The team then studied the performance of the EDS-HAT.
EDS-HAT detected 99 clusters of similar infections during this two-year period and identified at least one potential route of transmission in 65.7% of these clusters. During the same period, Infection Prevention used whole genome sequencing to help investigate 15 suspected outbreaks, two of which revealed genetically related infections.
Had the EDS-HAT worked in real time, the team estimates that up to 63 transmissions of an infectious disease from one patient to another could have been prevented. It would also have saved the hospital up to $ 692,500.
In a case study, EDS-HAT found a vancomycin-resistant Enterococcus faecium outbreak which she attributed to an interventional radiology procedure involving the injection of sterile contrast medium that was performed according to the manufacturer’s instructions. Due to detection of the outbreak by EDS-HAT, UPMC alerted the manufacturer to instructions that led to faulty sterilization practices.
“In this case, EDS-HAT made the link between apparently unconnected patient infections occurring in different hospital units, stopping this outbreak but also potentially preventing similar outbreaks in other hospitals,” said Harrison. “This example summarizes the value of EDS-HAT.”
UPMC plans to introduce real-time EDS-HAT at the UPMC Presbyterian Hospital and expects this innovation to benefit other infection prevention and control programs in the future. And the original EDS-HAT, which primarily focused on drug-resistant bacterial pathogens, will soon be expanded to incorporate sequencing of respiratory viruses, including COVID-19.
The other authors of this research are Jieshi Chen, MS, James K. Miller, Ph.D., and Artur Dubrawski, Ph.D., all from Carnegie Mellon University; and Praveen Kumar, B.Tech., Ashley M. Ayres, Shu-Ting Cho, MS, Chinelo Ezeonwuka, M.Sc., Marissa P. Griffith, Mustapha M. Mustapha, Ph.D., A. William Pasculle, Sc .D., Melissa I. Saul, MS, Kathleen A. Shutt, MS, Vatsala Srinivasa, MPH, Kady Waggle, Daniel J. Snyder, M.Sc., Vaughn S. Cooper, Ph.D., Daria Van Tyne, Ph.D., Graham M. Snyder, MD, Jane W. Marsh, Ph.D., and Mark S. Roberts, MD, all of Pitt or UPMC, or both.
This study was funded in part by National Institute of Allergies and Infectious Diseases grants R21 AI109459 and R01 AI127472.