This article is more than 5 years old. Information may no longer be current.
Artificial intelligence helps identify patients at risk for C. difficile
Click Here to Manage Email Alerts
Hospital-specific risk models built with the help of machine learning allowed physicians to more quickly and accurately identify patients with the highest risk for Clostridium difficile, according to new research published in Infection Control and Hospital Epidemiology.
Investigators said previous C. diff risk models were designed as one-size-fits-all approaches and did not take important hospital-specific factors into account, whereas their new approach allowed for the development of a tailor-made model for each individual health care facility.
“Despite substantial efforts to prevent C. difficile infection and to institute early treatment upon diagnosis, rates of infection continue to increase,” Erica Shenoy, MD, PhD, of the Massachusetts General Hospital (MGH) division of infectious diseases, said in a press release. “We need better tools to identify the highest risk patients so that we can target both prevention and treatment interventions to reduce further transmission and improve patient outcomes.”
Shenoy and colleagues extracted patient demographics, admission details, patient history and daily hospitalization details from the electronic health records of 191,014 adults admitted to the University of Michigan Hospitals and 65,718 admitted to MGH.
The investigators used their machine-learning-based model to analyze the collected data and then generate daily risk scores for each patient. They further categorized patients as high risk for C. diff if they reached a certain threshold.
The models were highly successful in predicting which patients would eventually receive a diagnosis of C. diff, achieving area under the receiver operating characteristic curve (AUROC) values of 0.82 (95% CI, 0.8 – 0.84) at University of Michigan and 0.75 (95% CI, 0.73–0.78) at MGH. The investigators noted that many of the top predictive factors differed between the two facilities’ risk models.
If validated in future studies, the risk prediction score could help physicians guide early screening for the infection, they concluded. They also made their algorithm code freely available for others to adapt and tailor to their own individual facilities, according to the press release.
“This represents a potentially significant advance in our ability to identify and ultimately act to prevent infection with C. difficile,” Vincent Young, MD, PhD, study investigator and the William Henry Fitzbutler professor in the department of internal medicine at the University of Michigan said in the press release. “The ability to identify patients at greatest risk could allow us to focus expensive and potentially limited prevention methods on those who would gain the greatest potential benefits.” – by Alex Young
Disclosures: The authors report no relevant financial disclosures.
Click Here to Manage Email Alerts