Hard stop on hospital infections

Illustrated by
Rachel Keeney

Pitt Med clinical scientists, with partners at Carnegie Mellon University, have found that machine learning and whole genome sequencing surveillance can stop in-hospital infections in their tracks.

They published their findings in fall 2021 in Clinical Infectious Diseases. UPMC Presbyterian is now implementing their system, designed to save lives and cut costs, in real time.

Senior author on the study Lee Harrison is an MD professor of medicine at the School of Medicine and of epidemiology at the School of Public Health; he leads the team using the Enhanced Detection System for Healthcare-Associated Transmission (EDS-HAT). EDS-HAT couples genomic sequencing with computer algorithms connected to the vast trove of data in electronic health records. When sequencing detects that two or more patients have near-identical strains of an infection, machine learning quickly mines those patients’ electronic health records for commonalities; it then alerts infection preventionists to investigate and halt further transmission.

Alexander Sundermann is a clinical research coordinator and a recently minted DrPH and assistant professor of medicine. He says the typical model hospitals use for detection is labor-intensive and often dependent on busy health care workers noticing a shared infection between patients. “These practices haven’t changed significantly in over a century,” adds Harrison.

From 2016 to 2018, UPMC Presbyterian ran EDS-HAT with a six-month lag (so as not to interfere with normal clinical practice) for several infectious pathogens often associated with health care–acquired infections. EDS-HAT detected 99 clusters of similar infections in that period and identified at least one potential transmission route in 65% of those clusters. During the same period, infection prevention used whole genome sequencing to aid in the investigation of 15 suspected outbreaks, two of which revealed genetically related infections.

If EDS-HAT had been running in real time, the team estimates as many as 63 transmissions of an infectious disease from one patient to another could have been prevented. It also would have saved the hospital up to $692,000.

Harrison says the current real-time implementation has detected one bronchoscopy-associated cluster and multiple nursing unit–based clusters. Hospital staff members were able to intervene for all of those.