BACKGROUND: Health care-associated infections due to multidrug-resistant organisms (MDROs), such as methicillin-resistant Staphylococcus aureus (MRSA) and Clostridioides difficile (CDI), place a significant burden on our health care infrastructure.
OBJECTIVE: Screening for MDROs is an important mechanism for preventing spread but is resource intensive. The objective of this study was to develop automated tools that can predict colonization or infection risk using electronic health record (EHR) data, provide useful information to aid infection control, and guide empiric antibiotic coverage.
METHODS: We retrospectively developed a machine learning model to detect MRSA colonization and infection in undifferentiated patients at the time of sample collection from hospitalized patients at the University of Virginia Hospital. We used clinical and nonclinical features derived from on-admission and throughout-stay information from the patient\'s EHR data to build the model. In addition, we used a class of features derived from contact networks in EHR data; these network features can capture patients\' contacts with providers and other patients, improving model interpretability and accuracy for predicting the outcome of surveillance tests for MRSA. Finally, we explored heterogeneous models for different patient subpopulations, for example, those admitted to an intensive care unit or emergency department or those with specific testing histories, which perform better.
RESULTS: We found that the penalized logistic regression performs better than other methods, and this model\'s performance measured in terms of its receiver operating characteristics-area under the curve score improves by nearly 11% when we use polynomial (second-degree) transformation of the features. Some significant features in predicting MDRO risk include antibiotic use, surgery, use of devices, dialysis, patient\'s comorbidity conditions, and network features. Among these, network features add the most value and improve the model\'s performance by at least 15%. The penalized logistic regression model with the same transformation of features also performs better than other models for specific patient subpopulations.
CONCLUSIONS: Our study shows that MRSA risk prediction can be conducted quite effectively by machine learning methods using clinical and nonclinical features derived from EHR data. Network features are the most predictive and provide significant improvement over prior methods. Furthermore, heterogeneous prediction models for different patient subpopulations enhance the model\'s performance.

We aim to evaluate the role of frailty and inflammatory markers in predicting the short-term outcomes after catheter-associated urinary tract infections (CAUTI) and central line-associated bloodstream infections (CLABSI).
Data regarding the patients\' characteristics, isolates on CAUTI and CLABSI, antibiotic susceptibility, frailty (11-point Modified Frailty Index), and inflammatory markers were retrospectively collected. Their impact on the short-term outcomes was assessed using regression modeling response.
One hundred and one patients with CAUTI (n = 71) and CLABSI (n = 30) between January 2018 and December 2019 were included in this study. The pooled incidence rates for CAUTI were 5.50 and for CLABSI 3.58 episodes/1000 catheter-days. We observed 74.7% drug resistance in our CAUTI isolates and 93.3% in CLABSI. In the multivariate analysis, frailty (P = 0.006), neutrophil/lymphocyte ratio (NLR) (P = 0.007) and the presence of sepsis (P = 0.029) were found to be significant predictors of in-hospital mortality in CAUTI. In patients with CLABSI, frailty (P = 0.029) and NLR (P = 0.029) were found significant and along with sepsis (P = 0.069) resulted in a regression model with good accuracy in predicting mortality. The receiver operating characteristic curve showed that 11-point Modified Frailty Index and NLR as well as the regression model significantly predicted mortality with an area under the curve of 86.1%, 81.4%, and 95.4%, respectively, in CAUTI, and 70.9%, 77.8%, and 95.2%, respectively, in CLABSI.

The Centers for Disease Control and Prevention\'s (CDC\'s) National Healthcare Safety Network (NHSN) is the most widely used health care-associated infection (HAI) and antimicrobial use and resistance surveillance program in the United States. Over 37,000 health care facilities participate in the program and submit a large volume of surveillance data. These data are used by the facilities themselves, the CDC, and other agencies and organizations for a variety of purposes, including infection prevention, antimicrobial stewardship, and clinical quality measurement. Among the summary metrics made available by the NHSN are standardized infection ratios, which are used to identify HAI prevention needs and measure progress at the national, regional, state, and local levels.
To extend the use of geospatial methods and tools to NHSN data, and in turn to promote and inspire new uses of the rendered data for analysis and prevention purposes, we developed a web-enabled system that enables integrated visualization of HAI metrics and supporting data.
We leveraged geocoding and visualization technologies that are readily available and in current use to develop a web-enabled system designed to support visualization and interpretation of data submitted to the NHSN from geographically dispersed sites. The server-client model-based system enables users to access the application via a web browser.
We integrated multiple data sets into a single-page dashboard designed to enable users to navigate across different HAI event types, choose specific health care facility or geographic locations for data displays, and scale across time units within identified periods. We launched the system for internal CDC use in January 2019.
CDC NHSN statisticians, data analysts, and subject matter experts identified opportunities to extend the use of geospatial methods and tools to NHSN data and provided the impetus to develop NHSNViz. The development effort proceeded iteratively, with the developer adding or enhancing functionality and including additional data sets in a series of prototype versions, each of which incorporated user feedback. The initial production version of NHSNViz provides a new geospatial analytic resource built in accordance with CDC user requirements and extensible to additional users and uses in subsequent versions.

BACKGROUND: The cluster detection of health care-associated infections (HAIs) is crucial for identifying HAI outbreaks in the early stages.
OBJECTIVE: We aimed to verify whether multisource surveillance based on the process data in an area network can be effective in detecting HAI clusters.
METHODS: We retrospectively analyzed the incidence of HAIs and 3 indicators of process data relative to infection, namely, antibiotic utilization rate in combination, inspection rate of bacterial specimens, and positive rate of bacterial specimens, from 4 independent high-risk units in a tertiary hospital in China. We utilized the Shewhart warning model to detect the peaks of the time-series data. Subsequently, we designed 5 surveillance strategies based on the process data for the HAI cluster detection: (1) antibiotic utilization rate in combination only, (2) inspection rate of bacterial specimens only, (3) positive rate of bacterial specimens only, (4) antibiotic utilization rate in combination + inspection rate of bacterial specimens + positive rate of bacterial specimens in parallel, and (5) antibiotic utilization rate in combination + inspection rate of bacterial specimens + positive rate of bacterial specimens in series. We used the receiver operating characteristic (ROC) curve and Youden index to evaluate the warning performance of these surveillance strategies for the detection of HAI clusters.
RESULTS: The ROC curves of the 5 surveillance strategies were located above the standard line, and the area under the curve of the ROC was larger in the parallel strategy than in the series strategy and the single-indicator strategies. The optimal Youden indexes were 0.48 (95% CI 0.29-0.67) at a threshold of 1.5 in the antibiotic utilization rate in combination-only strategy, 0.49 (95% CI 0.45-0.53) at a threshold of 0.5 in the inspection rate of bacterial specimens-only strategy, 0.50 (95% CI 0.28-0.71) at a threshold of 1.1 in the positive rate of bacterial specimens-only strategy, 0.63 (95% CI 0.49-0.77) at a threshold of 2.6 in the parallel strategy, and 0.32 (95% CI 0.00-0.65) at a threshold of 0.0 in the series strategy. The warning performance of the parallel strategy was greater than that of the single-indicator strategies when the threshold exceeded 1.5.
CONCLUSIONS: The multisource surveillance of process data in the area network is an effective method for the early detection of HAI clusters. The combination of multisource data and the threshold of the warning model are 2 important factors that influence the performance of the model.

We conducted a multicenter study to determine the clinical and microbiological characteristics of health care-associated (HCA) cellulitis in Korea. We retrospectively reviewed the medical records of patients who had been diagnosed with community-onset cellulitis. Of the 2208 cellulitis patients, 232 (10.5%) had HCA cellulitis, 1243 (56.3%) patients were hospitalized, and 15 (0.7%) died in hospital. Compared with community-acquired (CA) cellulitis, patients with HCA cellulitis were older and more frequently presented with comorbidity and septic shock. A total of 355 microorganisms were isolated from 314 patients (14.2%). Staphylococcus aureus (134 isolates) was the most common organism, followed by Streptococcus spp. (86 isolates) and Gram-negative fermenters (58 isolates). Methicillin-resistant S. aureus (MRSA) accounted for 29.1% (39/134) of S. aureus infections. None of the Gram-negative fermenters were resistant to carbapenem. The antibiotic susceptibility pattern of isolated microorganisms was not different between HCA and CA cellulitis. In patients with HCA cellulitis, S. aureus (11.2% [26/232] vs. 5.5% [108/1976], p = 0.001), including MRSA (4.3% [10/232] vs. 1.5% [29/1976], p = 0.003) and Gram-negative fermenters (6.0% [14/232] vs. 2.3% [44/1976], p = 0.002), were more common causative organisms than in CA-cellulitis patients. Age ≥ 65 years, septic shock, and HCA infection were statistically significant factors associated with in-hospital mortality.

We assessed the current status of knowledge, perception, attitude, and role model regarding hand hygiene (HH) among infection control nurses (ICNs) and identified the factors influencing these variables.
A structured questionnaire was adapted from a World Health Organization survey. Data were collected from November 8, 2017, to February 2, 2018.
ICNs showed the following scores (mean ± SD): knowledge (19.5 ± 2.3), perception (69.9 ± 8.9), attitude (46.9 ± 5.8), and role model (39.2 ± 6.0). HH performance of health care workers (HCWs) was 75.2 ± 15.5. Mean HH performance scores of HCWs (P = .007) differed significantly according to infection control experience (3 groups: ≤12 months, 13-24 months, >24 months). Perception, attitude, and role model scores showed positive correlations with each other (P < .01). The regression model for HH performance of HCWs was calculated as follows: Y1 = 31.638 + 0.067X1 (perception of ICNs) + 0.133X2 (attitude of ICNs) + 0.825X3 (role model of ICNs) (P < .001; adjusted R2 = 0.115).
Perception, attitude, and role model scores of ICNs were significant predictors of HH performance of HCWs.
Specialized well-structured HH education programs should be developed for ICNs that will help improve HH performance of HCWs.

The prevalence of Clostridium difficile spores was assessed in 48 observations of infected inpatients. Participants were randomized to hand hygiene with either alcohol-based handrub or soap and water. C difficile was recovered in 14.6% of pre-hand hygiene observations. It was still present on 5 of these 7 participants after hand hygiene (3/3 using alcohol-based handrub; 2/4 using soap and water).

BACKGROUND: Contextual factors associated with health care settings make reducing health care-associated infections (HAIs) a complex task. The aim of this article is to highlight how ethnography can assist in understanding contextual factors that support or hinder the implementation of evidence-based practices for reducing HAIs.
METHODS: We conducted a review of ethnographic studies specifically related to HAI prevention and control in the last 5 years (2012-2017).
RESULTS: Twelve studies specific to HAIs and ethnographic methods were found. Researchers used various methods with video-reflexive sessions used in 6 of the 12 studies. Ethnography was used to understand variation in data reporting, identify barriers to adherence, explore patient perceptions of isolation practices and highlight the influence of physical design on infection prevention practices. The term ethnography was used to describe varied research methods. Most studies were conducted outside the United States, and authors indicate insights gained using ethnographic methods (whether observations, interviews, or reflexive video recording) as beneficial to unraveling the complexities of HAI prevention.
CONCLUSIONS: Ethnography is well-suited for HAI prevention, especially video-reflexive ethnography, for activating patients and clinicians in infection control work. In this era of increasing pressure to reduce HAIs within complex work systems, ethnographic methods can promote understanding of contextual factors and may expedite translation evidence to practice.

There are many well-established national health care-associated infection surveillance programs (HAISPs). Although validation studies have described data quality, there is little research describing important characteristics of large HAISPs. The aim of this study was to broaden our understanding and identify key characteristics of large HAISPs.
Semi-structured interviews were conducted with purposively selected leaders from national and state-based HAISPs. Interview data were analyzed following an interpretive description process.
Seven semi-structured interviews were conducted over a 6-month period during 2014-2015. Analysis of the data generated 5 distinct characteristics of large HAISPs: (1) triggers: surveillance was initiated by government or a cooperative of like-minded people, (2) purpose: a clear purpose is needed and determines other surveillance mechanisms, (3) data measures: consistency is more important than accuracy, (4) processes: a balance exists between the volume of data collected and resources, and (5) implementation and maintenance: a central coordinating body is crucial for uniformity and support.
National HAISPs are complex and affect a broad range of stakeholders. Although the overall goal of health care-associated infection surveillance is to reduce the incidence of health care-associated infection, there are many crucial factors to be considered in attaining this goal. The findings from this study will assist the development of new HAISPs and could be used as an adjunct to evaluate existing programs.

Antibiotic resistance is a challenge in long-term care facilities (LTCFs). The objective of this study was to demonstrate that a novel, minimally invasive program not interfering with activities of daily living or socialization could lower methicillin-resistant Staphylococcus aureus (MRSA) disease.
This was a prospective, cluster-randomized, nonblinded trial initiated at 3 LTCFs. During year 1, units were stratified by type of care and randomized to intervention or control. In year 2, all units were converted to intervention consisting of universal decolonization using intranasal mupirocin and a chlorhexidine bath performed twice (2 decolonization-bathing cycles 1 month apart) at the start of the intervention period. Subsequently, after initial decolonization, all admissions were screened on site using real-time polymerase chain reaction, and those MRSA positive were decolonized, but not isolated. Units received annual instruction on hand hygiene. Enhanced bleach wipe cleaning of flat surfaces was done every 4 months.
There were 16,773 tests performed. The MRSA infection rate decreased 65% between baseline (44 infections during 365,809 patient days) and year 2 (12 infections during 287,847 patient days; P <.001); a significant reduction was observed at each of the LTCFs (P <.03).
On-site MRSA surveillance with targeted decolonization resulted in a significant decrease in clinical MRSA infection among LTCF residents.