Development, Evaluation, and Multisite Deployment of a Machine Learning Decision Tree Algorithm To Optimize Urinalysis Parameters for Predicting Urine Culture Positivity.

PittUDT, a recursive partitioning decision tree algorithm for predicting urine culture (UC) positivity based on macroscopic and microscopic urinalysis (UA) parameters, was developed in support of a broader system-wide diagnostic stewardship initiative to increase appropriateness of UC testing. Reflex algorithm training utilized results from 19,511 paired UA and UC cases (26.8% UC positive); the average patient age was 57.4 years, and 70% of samples were from female patients. Receiver operating characteristic (ROC) analysis identified urine white blood cells (WBCs), leukocyte esterase, and bacteria as the best predictors of UC positivity, with areas under the ROC curve of 0.79, 0.78, and 0.77, respectively. Using the held-out test data set (9,773 cases; 26.3% UC positive), the PittUDT algorithm met the prespecified target of a negative predictive value above 90% and resulted in a 30 to 60% total negative proportion (true-negative plus false-negative predictions). These data show that a supervised rule-based machine learning algorithm trained on paired UA and UC data has adequate predictive ability for triaging urine specimens by identifying low-risk urine specimens, which are unlikely to grow pathogenic organisms, with a false-negative proportion under 5%. The decision tree approach also generates human-readable rules that can be easily implemented across multiple hospital sites and settings. Our work demonstrates how a data-driven approach can be used to optimize UA parameters for predicting UC positivity in a reflex protocol, with the intent of improving antimicrobial stewardship and UC utilization, a potential avenue for cost savings.

Assessment of the effectiveness of ultraviolet-C disinfection on transmission of hospital-acquired pathogens from prior room occupants.

Objective: To evaluate the effectiveness of ultraviolet-C (UV-C) disinfection as an adjunct to standard chlorine-based disinfectant terminal room cleaning in reducing transmission of hospital-acquired multidrug-resistant organisms (MDROs) from a prior room occupant. Design: A retrospective cohort study was conducted to compare rates of MDRO transmission by UV-C status from January 1, 2016, through December 31, 2018. Setting: Acute-care, single-patient hospital rooms at 6 hospitals within an academic healthcare system in Pennsylvania. Methods: Transmission of hospital-acquired MDRO infection was assessed in patients subsequently assigned to a single-patient room of a source occupant with carriage of 1 or more MDROs on or during admission. Acquisition of 5 pathogens was compared between exposed patients in rooms with standard-of-care chlorine-based disinfectant terminal cleaning with or without adjunct UV-C disinfection. Logistic regression analysis was used to estimate the adjusted risk of pathogen transfer with adjunctive use of UV-C disinfection. Results: In total, 33,771 exposed patient admissions were evaluated; the source occupants carried 46,688 unique pathogens. Prior to the 33,771 patient admissions, 5,802 rooms (17.2%) were treated with adjunct UV-C disinfection. After adjustment for covariates, exposed patients in rooms treated with adjunct UV-C were at comparable risk of transfer of any pathogen (odds ratio, 1.06; 95% CI, 0.84-1.32; P = .64). Conclusion: Our analysis does not support the use of UV-C in addition to post-discharge cleaning with chlorine-based disinfectant to lower the risk of prior room occupant pathogen transfer.

Method for Economic Evaluation of Bacterial Whole Genome Sequencing Surveillance Compared to Standard of Care in Detecting Hospital Outbreaks.

BACKGROUND: Whole genome sequencing (WGS) surveillance and electronic health record data mining have the potential to greatly enhance the identification and control of hospital outbreaks. The objective was to develop methods for examining economic value of a WGS surveillance-based infection prevention (IP) program compared to standard of care (SoC). METHODS: The economic value of a WGS surveillance-based IP program was assessed from a hospital's perspective using historical outbreaks from 2011-2016. We used transmission network of outbreaks to estimate incremental cost per transmission averted. The number of transmissions averted depended on the effectiveness of intervening against transmission routes, time from transmission to positive culture results and time taken to obtain WGS results and intervene on the transmission route identified. The total cost of an IP program included cost of staffing, WGS, and treating infections. RESULTS: Approximately 41 out of 89 (46%) transmissions could have been averted under the WGS surveillance-based IP program, and it was found to be a less costly and more effective strategy than SoC. The results were most sensitive to the cost of performing WGS and the number of isolates sequenced per year under WGS surveillance. The probability of the WGS surveillance-based IP program being cost-effective was 80% if willingness to pay exceeded $2400 per transmission averted. CONCLUSIONS: The proposed economic analysis is a useful tool to examine economic value of a WGS surveillance-based IP program. These methods will be applied to a prospective evaluation of WGS surveillance compared to SoC.

Prospective assessment of the effectiveness of standard high-level disinfection for echoendoscopes.

BACKGROUND AND AIMS: Duodenoscopes have been implicated in the transmission of multidrug-resistant organisms (MDROs). Echoendoscopes could potentially transmit infection. The aim of this study was to assess the effectiveness of standard high-level disinfection (HLD) for radial and linear echoendoscopes and to compare it with that of duodenoscopes. METHODS: We performed a prospective single-center study sampling echoendoscopes immediately before use, from the working channel (radial and linear echoendoscopes) and the transducer (radial echoendoscope) or elevator mechanism and transducer (linear echoendoscope). The primary outcome was the proportion of echoendoscopes with any culture showing ≥1 MDRO; secondary outcomes included bacterial growth >0 colony forming units (CFUs) and ≥10 CFUs on either sampling location. We compared these findings with duodenoscope cultures from the previously published DISINFECTS trial. RESULTS: During the study period, 101 echoendoscopes were sampled (n = 50 radial echoendoscopes, n = 51 linear echoendoscopes). No MDROs were recovered. Bacterial growth >0 CFUs was noted in 6% and ≥10 CFUs in 3% of all echoendoscopes. There was no significant difference in growth between radial and linear echoendoscopes (P = .4 for >0 CFU growth; P = .6 for ≥10 CFUs growth). The proportion of transducer and/or elevator mechanism positive for bacterial growth was significantly higher in duodenoscopes as compared with echoendoscopes (P = .02). CONCLUSIONS: After standard HLD, no echoendoscope showed MDRO growth, 6% showed >0 CFUs, and 3% showed ≥10 CFUs bacterial growth. Bacterial growth was higher in duodenoscopes at the level of the transducer and/or elevator mechanism when compared with echoendoscopes.

Status of the Prevention of Multidrug-Resistant Organisms in International Settings: A Survey of the Society for Healthcare Epidemiology of America Research Network.

OBJECTIVE To examine self-reported practices and policies to reduce infection and transmission of multidrug-resistant organisms (MDRO) in healthcare settings outside the United States. DESIGN Cross-sectional survey. PARTICIPANTS International members of the Society for Healthcare Epidemiology of America (SHEA) Research Network. METHODS Electronic survey of infection control and prevention practices, capabilities, and barriers outside the United States and Canada. Participants were stratified according to their country's economic development status as defined by the World Bank as low-income, lower-middle-income, upper-middle-income, and high-income. RESULTS A total of 76 respondents (33%) of 229 SHEA members outside the United States and Canada completed the survey questionnaire, representing 30 countries. Forty (53%) were high-, 33 (43%) were middle-, and 1 (1%) was a low-income country. Country data were missing for 2 respondents (3%). Of the 76 respondents, 64 (84%) reported having a formal or informal antibiotic stewardship program at their institution. High-income countries were more likely than middle-income countries to have existing MDRO policies (39/64 [61%] vs 25/64 [39%], P=.003) and to place patients with MDRO in contact precautions (40/72 [56%] vs 31/72 [44%], P=.05). Major barriers to preventing MDRO transmission included constrained resources (infrastructure, supplies, and trained staff) and challenges in changing provider behavior. CONCLUSIONS In this survey, a substantial proportion of institutions reported encountering barriers to implementing key MDRO prevention strategies. Interventions to address capacity building internationally are urgently needed. Data on the infection prevention practices of low income countries are needed. Infect Control Hosp Epidemiol. 2016:1-8.