Extending outbreak investigation with machine learning and graph theory - Benefits of new tools with application to a nosocomial outbreak of a multidrug-resistant organism

Q: What hospital outbreak and pathogen context does the study address?

The study focuses on a vancomycin-resistant enterococci (VRE) outbreak in a hospital network from January 2018 through July 2020, emphasizing a nosocomial investigation approach.

Q: How were risk factors for VRE colonization determined?

Risk factors were identified using standard statistical methods and extended with a decision-tree machine-learning approach based on medical records from January 2018 to December 2019.

Q: What transmission pathways were identified using graph theory?

Graph network analysis established three main connections among VRE cases: healthcare personnel, medical devices, and patient rooms.

Aug 2, 2026

Upload by Jiven

Objective: From January 1, 2018, until July 31, 2020, a hospital network outbreak of vancomycin-resistant enterococci (VRE) prompted development of improved investigation processes. The study applies machine-learning and graph-theoretical methods to strengthen nosocomial outbreak analysis. Methods: Medical records from January 2018 through December 2019 were assembled, risk factors for VRE colonization were identified with standard statistics and decision-tree learning, and transmission pathways were inferred using network graph analysis. Results: Among 560 VRE patients and 86,684 controls, logistic models and decision trees confirmed predictors including age, ICU admission, comorbidity burden, antibiotic exposure, and ward mobility; graph analysis identified three major links. Conclusions: The work supports data science for outbreak understanding, while emphasizing interpretability needs, data maturity, and confounding considerations for screening and isolation.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Objective

Methods

Data and risk-factor modeling

Graph network pathway detection

Results

Predictors of VRE colonization

Identified transmission pathways

Conclusions

Implications for screening and isolation strategies

Infection Control & Hospital Epidemiology (2022), 1–7 doi:10.1017/ice.2022.66 Original Article Extending outbreak investigation with machine learning and graph theory: Benefits of new tools with application to a nosocomial outbreak of a multidrug-resistant organism Andrew Atkinson PhD1,a , Benjamin Ellenberger MSc2,a , Vanja Piezzi MD1, Tanja Kaspar MPH1, Luisa Salazar-Vizcaya PhD1, Olga Endrich MD3, Alexander B. Leichtle MD2,4 and Jonas Marschall MD1,5 1Department of Infectious Diseases, Bern University Hospital, Inselspital, University of Bern, Bern, Switzerland, 2Insel Data Science Center, Bern University Hospital, Inselspital, University of Bern, Bern, Switzerland, 3Medical Directorate, Bern University Hospital, Inselspital, University of Bern, Bern, Switzerland, 4University Institute of Clinical Chemistry, Bern University Hospital, Inselspital, University of Bern, Bern, Switzerland and 5Division of Infectious Diseases, Washington University School of Medicine, St Louis, Missouri, United States Abstract Objective: From January 1, 2018, until July 31, 2020, our hospital network experienced an outbreak of vancomycin-resistant enterococci (VRE). The goal of our study was to improve existing processes by applying machine-learning and graph-theoretical methods to a nosocomial outbreak investigation. Methods: We assembled medical records generated during the first 2 years of the outbreak period (January 2018 through December 2019). We identified risk factors for VRE colonization using standard statistical methods, and we extended these with a decision-tree machine-learning approach. We then elicited possible transmission pathways by detecting commonalities between VRE cases using a graph theoretical network analysis approach. Results: We compared 560 VRE patients to 86,684 controls. Logistic models revealed predictors of VRE colonization as age (aOR, 1.4 (per 10 years), with 95% confidence interval [CI], 1.3–1.5; P < .001), ICU admission during stay (aOR, 1.5; 95% CI, 1.2–1.9; P < .001), Charlson comorbidity score (aOR, 1.1; 95% CI, 1.1–1.2; P < .001), the number of different prescribed antibiotics (aOR, 1.6; 95% CI, 1.5–1.7; P < .001), and the number ofrooms the patient stayed in during their hospitalization(s) (aOR,1.1; 95%CI,1.1–1.2; P < .001). The decision-tree machinelearning method confirmed these findings. Graph network analysis established 3 main pathways by which the VRE cases were connected: healthcare personnel, medical devices, and patient rooms. Conclusions: We identified risk factors for being a VRE carrier, along with 3 important links with VRE (healthcare personnel, medical devices, patient rooms) . Data science is likely to provide a better understanding of outbreaks, but interpretations require data maturity, and potential confounding factors must be considered. (Received 13 October 2021; accepted 18 February 2022) Electronic medical records contain information relevant for outbreak investigations; consequently, by integrating the relevant data sources, we can potentially inform and improve patient screening and isolation strategies. However, this integration necessarily leads to large quantities of data, which can be difficult to analyze using standard statistical techniques. Machine-learning or “artificial intelligence” methods comprise a toolbox of approaches that have become popular for analyzing such “big data”.1,2 To date in the field Author for correspondence: Andrew Atkinson, E-mail: [andrew.atkinson@insel.ch](andrew.atkinson@insel.ch)[a](aAuthors of equal contribution.)[Authors of equal contribution.](aAuthors of equal contribution.) PREVIOUS PRESENTATION. Parts of these results were presented at the IDWeek conference in Washington, DC, on October 2–6, 2019 . Cite this article: Atkinson A, et al. (2022) . Extending outbreak investigation with machine learning and graph theory: Benefits of new tools with application to anosocomial outbreak of a multidrug-resistant organism.

FAQs

What hospital outbreak and pathogen context does the study address?

The study focuses on a vancomycin-resistant enterococci (VRE) outbreak in a hospital network from January 2018 through July 2020, emphasizing a nosocomial investigation approach.

How were risk factors for VRE colonization determined?

Risk factors were identified using standard statistical methods and extended with a decision-tree machine-learning approach based on medical records from January 2018 to December 2019.

What transmission pathways were identified using graph theory?

Graph network analysis established three main connections among VRE cases: healthcare personnel, medical devices, and patient rooms.

Aug 2, 2026

Upload by Jiven