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

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 of rooms 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.

Nosocomial outbreak of vancomycin-resistant Enterococcus faecium (VRE) ST796, Switzerland, 2017 to 2020

“…Therefore, we cannot make any statement on the role of healthcare workers in the transmission chain. Novel screening approaches based on machine learning were prototyped during the outbreak, and these are now in place to better determine who may have a link to other cases and should therefore be screened in similar nosocomial outbreaks in future [ 37 , 38 ]. Finally, given the costs for sequencing, we performed WGS only for approximately 1 year.…”

Perspectives on research needs in healthcare epidemiology, infection prevention, and antimicrobial stewardship: what’s on the horizon—Part II

“…A recent example involves outbreak investigation. 20 Beyond the local setting, applications in global scenarios are within reach. 21 As more routine medical data become available in dedicated data science centers in healthcare, innovative approaches for using and visualizing it should be employed.…”

Innovative Techniques for Infection Control and Surveillance in Hospital Settings and Long-Term Care Facilities: A Scoping Review

“…Despite the year of publication, we also observed the last year in which the study was conducted. We observed that most of the studies were conducted in 2019 (n = 7, [10,18,20,31,37,41,57]), followed by 2017 (n = 6, [14,22,23,33,44,55,56]), 2018 (n = 6, [17,28,32,34,42,43]), and 2020 (n = 6, [11,15,21,26,27,48]). Only a few studies were conducted in earlier and later years (2021 = 4 [18,24,40,46]; 2022 = 3 [29,30,39]; 2016 = 1 [16]; 2015 = 3 [35,36,38]; 2014 = 1 [19]; 2013 = 2 [12,49]; 2012 = 1 [45]; 2010 = 1 [13]; Unknown = 1 [47]).…”

“…For the 43 studies included in the analysis [55][56][57], most (n = 16) were published in the last year [11,15,17,18,20,21,26,[28][29][30]32,39,40,46,48,57], four in 2022 [24,31,34,42], six in 2021 [10,27,33,35,37,49], seven in 2020 [12,14,25,36,38,43,44], six in 2019 [13,22,23,41,55,56], and four in 2018 [16,19,45,47]. Despite the year of publication, we also observed the last year in which the study was conducted.…”

“…The mean duration of the studies was 43.7 months (median 38.5; IQR 12-68). The studies were conducted across multiple continents, with sixteen conducted in North America (fourteen in the US [10,14,16,17,19,20,23,34,38,42,[45][46][47]55] and two in Canada [26,35]); twenty-one in Europe (six in the UK, [24,27,29,30,36,57], four in Sweden [12,15,21,49], three in Germany [28,41,44], two in France [13,32] and Italy [37,43], and one each in the Netherlands [56], Spain [40], Norway [39], and Switzerland [11]); five in Asia (three in China [18,25,31], one in Pakistan [48], and one in Thailand [22]), and one in Africa (Rwanda [33]). The majority of studies were single-centre studies (n = 26, [10,11,1...…”