Background: Vancomycin-resistant Enterococcus (VRE) is a leading cause of nosocomial infections that carries an increased risk of mortality when compared to vancomycin-sensitive Enterococcus (VSE). Data on the frequency of conversion between VSE and VRE in patients are scarce. Among patients presenting with VSE infections, little is known about the subsequent risk of conversion to VRE in the initial treatment period. Methods: A descriptive analysis of VSE to VRE conversion and a retrospective case-control study were performed examining cases of VSE that had subsequent cultures positive for VRE within 90 days within a quaternary healthcare system. Cases were obtained from June 2013 through December 2018. Controls were patients who had VSE culture followed by another VSE culture and were matched by organism (E. faecalis or E. faecium), time between cultures, and initial culture site. Age, gender, healthcare, antibiotic, Clostridiodes difficile, proton pump inhibitor (PPI) exposure, and H2 blocker exposures, and prior VRE infection or colonization were abstracted from the electronic medical record. A univariate analysis with the Fisher exact test was performed with significance considered for P < .05. Results: In total, 8,913 cases of E. faecalis and 2,322 cases of E. faecium were included in the study. Of 8,913 cases of E. faecalis, 51 of 8,503 (0.6%) cultured VRE after VSE, and 47 of 403 (11.7%) cultured VSE after initial VRE. Of E. faecium, 51 of 783 (6.5%) cultured VRE after VSE, and 76 of 1,532 (5.0%) cultured VSE after initial VRE. In total, 76 cases were matched with 99 controls. Patients converting from VSE to VRE were more likely to have prior admission to an intensive care unit (P = .0207), prior positive swab or culture for VRE (P = .0114), previous C. difficile infection (P = .0155), prior vancomycin (P = .0022) and cefepime (P = .0089) exposure. Patients receiving vancomycin after initial VSE culture were more likely to have subsequent cultures positive for VRE (P = .0053). There was no difference in age (P = .966) or male sex (P = .7588). Conclusions: Conversion from VSE to VRE is common, and E. faecium is more likely to become resistant than E. faecalis. Reversion to a vancomycin-sensitive phenotype is also common, and E. faecalis is more likely to show subsequent sensitivity than E. faecium. Previous admission to an intensive care unit, prior colonization or infection with VRE, prior C. difficile infection, and exposure to vancomycin and cefepime are risk factors for emergence of VRE after treatment for vancomycin-sensitive Enterococcus.

Funding: None

Disclosures: None

Background: Surgical site infections (SSIs) among cardiothoracic (CT) patients are associated with high rates of morbidity and mortality. Data are limited regarding SSI incidence among pediatric patients undergoing primary reparative procedures for congenital cardiac disease. Published evidence on targeted interventions to prevent pediatric CT-surgery SSI is lacking. We aimed to establish standard metrics for measuring CT-surgery SSI incidence and to implement bundled interventions for SSI prevention. Methods: A dedicated CT-surgery SSI prevention workgroup was established, consisting of hospital leadership, CT surgeons, cardiac critical care unit staff, anesthesia, perfusion, environmental services, instrument sterile processing, risk management, infection prevention and antibiotic stewardship. We created a standard definition for CT-surgery SSI and calculated retrospective SSI rates over a 24-month period (2017–2019). The outcome measured was incidence of CT-surgery SSI per 100 primary cardiac procedures with delayed ( 3 days after primary surgery) or non-delayed chest closure. The difference in proportion of SSI was reported separately for delayed closure and non-delayed closure; statistical significance was tested using a Fisher’s Exact test. We identified many potential improvement opportunities, including gaps in SSI surveillance, poor compliance with daily bathing, inconsistent perioperative antimicrobial prophylaxis, lack of controlled environment for bedside chest closures, and lapses in environmental cleaning. These issues informed the enhanced SSI prevention bundle, which included education on sterility with the operating room (OR) staff. Protocols for care of cardiac patients with delayed chest closures focused on universal daily and preoperative chlorhexidine baths. In addition, the bundle incorporated stringent environmental cleaning interventions including scheduled decluttering of patient rooms and clinical spaces, terminal cleaning of patient rooms prior to returning from the OR, and use of adjunctive ultraviolet light for the daily cleaning of operating rooms and patient rooms at discharge. Results: Surveillance definition of microbiological growth from a clinical sample obtained within 30 days of primary cardiac procedure sufficiently captured all CT-surgery SSIs. Of 551 CT-surgery procedures prior to intervention, 91 (17%) had delayed final operative closures. Prior to the intervention, 16 SSIs were identified from July 2017 – May 2019 for a rate of 2.90 per /100 procedures, and was higher among patients with delayed chest closure 6.59 per /100 procedures (6 SSIs/91 procedures) versus those with primary chest closure 2.17 per /100 procedures (10 SSIs/460 procedures; P = 0.034). Gram-positive organisms, including coagulase coagulase-negative Staphylococci, were most frequently identified as the causative organisms for SSIs. Compliance with bundled intervention, rolled out over a 2-month period, was associated with an immediate decrease in the number of SSIs for primary and delayed chest closures 6SSIs /185 procedures in the initial quarters (August – December 2019) of the post-intervention period. However, this decrease was not reflected in the overall rate (3.24 per /100 procedures) due to fewer procedures performed. Data collection to measure sustainability is ongoing. Conclusions: Bundled interventions targeting skin antisepsis and environmental cleaning may be associated with a decrease in SSIs among pediatric CT-surgery patients. Ongoing surveillance is required to determine sustainability of these interventions.

Funding: None

Disclosures: None

Background: Clostridioides difficile infection (CDI) is a major contributor to morbidity and mortality in patients with hematologic malignancy. Due to both immunosuppression and frequent antibiotic exposures, up to one-third of inpatients receiving chemotherapy or stem-cell transplant develop CDI. Transmission of C. difficile in healthcare facilities occurs due to environmental surface contamination and hand carriage by healthcare workers from colonized and infected patients. We investigated the effectiveness of enhanced room cleaning in collaboration with environmental services (EVS) staff to prevent CDI transmission and infection.

Methods: From April 1, 2018, to September 30, 2018, a multimodal enhanced cleaning intervention was implemented on 2 oncology units at the Hospital of the University of Pennsylvania. This intervention included real-time feedback to EVS staff following ATP bioluminescence monitoring. Additionally, all rooms on the intervention units underwent UV disinfection after terminal cleaning. We performed a system-level cohort study, comparing rates of CDI on the 2 study units to historic and 2 concurrent control units. Historic and concurrent control units received UV disinfection only for rooms with prior occupants with MRSA or CDI. All units during the intervention period received education on the importance of environmental cleaning for infection prevention. Mixed-effects Poisson regression was used to adjust for system-level confounders. Results: A median of 1.34 CDI cases per 1,000 patient days (IQR, 1.20–3.62) occurred during the 12-month baseline period. There was a trend toward a reduced rate of CDI across all units during the intervention period (median, 1.19; IQR, 0.00–2.47; P = .13) compared with all units during the historical period. Using mixed-effects Poisson regression, accounting for the random effects of study units, the intervention was associated with an incidence rate ratio for C. difficile of 0.72 compared to control units (95% CI, 0.53–0.97; P = .03). Average room turnaround time (TAT) increased across all units during the study period, from 78 minutes (IQR 74–81) to 92 minutes (IQR, 85–96; P < .001). Within the intervention period, TAT was higher on intervention units (median, 94 minutes; IQR, 92–98) compared to concurrent control units (median, 85; IQR, 80–92; P = .005). Conclusions: Enhanced environmental cleaning, including UV disinfection of all patient rooms and ATP bioluminescent monitoring with real-time feedback, was associated with a reduction in the incidence of CDI.

Funding: None

Disclosures: None

Background: Influenza virus illness affects millions. The virus’s ability to change yearly and its dependence on private manufacturing leads to major challenges immunizing populations. Currently the United States expects to distribute close to 162–168 million doses of influenza vaccine to the market (CDC, October 25, 2019) to immunize the population for the 2019–2020 season. In line with international best practice, our healthcare facility mandated the seasonal influenza vaccine for all employees. We set out to introduce a novel nursing workflow to enhance our Flu Campaign utilizing our Epic Systems healthcare software. Methods: The Epic Flu Express Lane function boosts provider satisfaction and efficiency by minimizing documentation time. Nurses can document the entire flu visit on a single screen with a few clicks. Workflow: An employee presents at the flu station registration desk and a scheduler creates an appointment as a Flu Visit type. After a nurse opens the chart, the system determines whether a Flu Express Lane order is a fit for the visit, launching the Express Lane activity. The nurse updates all relevant information and finally drops in the administration charge associated with the appropriate diagnosis code for the vaccine, to sign into the Express Lane. Establishment of dedicated and mobile employee flu stations: Immunizations are provided at several stations located inside the hospital and alternate sites, including the our cafeteria. Additionally, the flu team extended immunizations off site at our administrative building and an unoccupied inpatient floor, giving hospital employees easier access. Integration of Epic and human capital resource planning systems: By integrating the Epic and Mawared Human Capital systems, the flu team developed an automated workflow to capture employee compliance. A daily extract from Epic is imported to Mawared updating employee compliance status to completed. For those with incomplete status, Mawared sends an automated notification via e-mail reminding employees to complete the flu vaccination process for the season. Results: The Epic Nursing Express Lane workflow was well received, it facilitated compliance and practice efficiency for all mobile services provided. We achieved our target to provide influenza vaccination to >95% of employees in 2019 (Fig. 2). Conclusions: The electronic health record Epic Flu Express Lane workflow is an effective tool to complement practice efficiency and to facilitate a successful mobile Flu Campaign compliance in healthcare facilities.

Funding: None

Disclosures: None

Background: Extended-spectrum β-lactamase–producing (ESBL) Escherichia coli infection incidence is increasing in the United States. This increase may be due to the rapid expansion of ST131, which is now the predominant ESBL strain globally, often multidrug resistant, and has been shown to establish longer-term human colonization than other E. coli strains. We assessed potential risk factors that distinguish ST131 from other ESBL E. coli. Methods: From October 1 through December 31, 2017, 5 CDC Emerging Infections Program (EIP) sites pilot tested active, laboratory-based surveillance in selected counties in Colorado, Georgia, New Mexico, New York, and Tennessee. An E. coli case was defined as the first isolation from a normally sterile body site or urine in a surveillance area resident in a 30-day period resistant to 1 extended-spectrum cephalosporin antibiotic and susceptible or intermediate to all carbapenem antibiotics tested. Epidemiologic data were collected from case patients’ medical records. A convenience sample of 117 E. coli isolates from case patients was collected. All isolates underwent whole-genome sequencing to determine sequence type and the presence of ESBL genes. We compared ST131 E. coli epidemiology to other ESBL E. coli. Results: Among 117 E. coli isolates, 97 (83%) were ESBL producers. Of the 97 ESBL E. coli, 52 (54%) were ST131 (range, for 4 EIP sites submitting >10 isolates: 25%–88%; P < .001). Other common STs were ST38 (12%) and ST10 (5%). ST131 infections were more likely to be healthcare-associated than non-ST131 (56% vs 36%; P = .05) (Table 1). Among specific prior healthcare exposures, only residence in long-term care facilities (LTCFs) in the year before culture was more common among ST131 case patients (29% vs 11%; P = .03). Notably, 85% of ESBL E. coli collected from LTCF residents were ST131. ST131 E. coli were more common among patients with underlying medical conditions (81% vs 60%; P = .02). No statistically significant difference by sex, race, age, culture source, location of culture collection, and frequency of antibiotic use in the prior 30 days was observed. Conclusions:The prevalence of ST131 E. coli varies regionally. The association between ST131 and LTCFs suggests that these may be particularly important settings for ST131 acquisition. Improving infection control measures that limit ESBL transmission in these settings and preventing dissemination in facilities receiving patients from LTCFs may be necessary to contain ST131 spread.

Funding: None

Disclosures: None

Background: Little is known about the epidemiology, microbiology, and clinical management of carbapenem-resistant Enterobacteriaceae (CRE) in outpatient settings. In Connecticut, all clinical CRE isolates are submitted to the state public health laboratory (SPHL) for a customized panel of antimicrobial susceptibility and carbapenemase gene testing. We describe all outpatient cases of CRE in Connecticut in 2018, including location of presentation, risk factors, microbiology and aspects of treatment. Methods: Outpatient CRE cases were defined as CRE infection in a patient not hospitalized at the time of positive CRE culture or within 30 days after culture collection. Outpatient cases were identified from routine statewide CRE reporting by reviewing clinical and laboratory data. A questionnaire was sent to outpatient providers who ordered the cultures that yielded CRE to collect additional clinical information. Antimicrobial susceptibility and carbapenemase gene detection results from the SPHL were also summarized. Results: Among 53 outpatient CRE cases (1 blood, 52 urine), the most common organisms were Enterobacter (25, 47%), Klebsiella (12, 23%), and E. coli (11, 21%). Overall, 21 (39.6%) patients presented in primary care settings, 8 (15%) in urology offices, 6 (11%) in women’s health/OBGYN clinics; the remainder presented across various clinical settings (Fig. 1). Of 42 patients for whom clinical data were available, 45% had been hospitalized within the prior year and 19% had a chronic indwelling device. Among outpatient CRE cases, infectious diseases consultation was reported in 9.5% and lab consultation in none. Median patient age was 83 years. Of 36 CRE samples for which lab data were available, 9 (25%) were carbapenemase-producers (CP-CRE), of which 8 were blaKPC positive. Sensitivity rates to oral antimicrobials ranged from 43% to 75% (Table 1). Conclusions: CRE infections occur in several different outpatient settings, and formal ID consultation in the management of these patients is infrequent. These findings highlight the critical need for providers across different outpatient specialties to be familiar with the clinical management and infection control practices needed in caring for patients with CRE. Hospitalization within the year prior to presentation was frequent among patients who developed subsequent CRE-positive cultures. Most outpatient CP-CRE cases in CT are due to KPC production. Overall, sensitivity to oral antimicrobials frequently prescribed in outpatient settings is low, providing additional challenges to the outpatient management of patients with CRE infection. Fosfomycin, though only approved for E. coli infections, may be an acceptable antibiotic choice for the treatment of these patients.

Funding: None

Disclosures: None

Background: Acute-care hospitals began reporting methicillin-resistant Staphylococcus aureus (MRSA) LabID facility-wide inpatient events to the NHSN in 2013. Few data are available regarding the epidemiology of these patients. Methods: We conducted a retrospective cohort study of patients who developed hospital onset Staphylococcus aureus bloodstream infections (HO-SA-BSIs) to describe the epidemiology (characteristics and outcomes) from January 2014 through June 2019 and to compare MRSA LabID BSIs to HO-MSSA BSIs. Proportions were compared using 2 and continuous variables using the Kruskal-Wallis test (EpiInfo). Results: Overall, 264 HO-SA BSIs occurred over the study period (2.21 per 10,000 patient days), 160 HO-MSSA BSIs (1.34 per 10,000 patient days), and 104 MRSA LabID BSIs (0.869 per 10,000 patient days). These rates have not significantly changed over time (Fig. 1). Most of these patients were men (64%); 42.4% were African-American; mean age was 43.5 years; mean Charlson comorbidity index was 3.2; 67.8% were admitted for medical care (vs surgical); and 13.3% had a previous history of S. aureus infection. Of all HO-SA-BSIs, 49.2% were acquired in the ICU, 53.8% were primary BSIs, and 37.9% were catheter associated. Patients were hospitalized a mean of 19.9 days prior to HO-SA BSI, and the mean overall length of stay was 48.5 days. Compared to HO-MSSA BSIs, there were no significant differences in these characteristics among MRSA LabID BSIs except that a significantly greater proportion were catheter associated (46.2% vs 32.5%; OR, 1.78; 95% CI, 1.07–2.96; P = .04). Overall, 101 patients (38.3%) died: 41 with MRSA LabID BSI (39.4%) and 60 with HO-MSSA BSI (37.5%). Mortality rates have not changed significantly over time. The mean number of days to death was 154.2, and 59 patients (22.3%) died during incident hospitalization: 26.9% of MRSA patients and 19.4% of MSSA BSI patients. Moreover, 28.3% of patients were readmitted within 30 days of discharge from incident hospitalization, and compared to HO-MSSA BSI, this rate was significantly higher among MRSA LabID BSI patients (34.2% vs 24.8%; OR, 2.07; 95% CI, 1.09–3.93; P = .03). Among those who died, 58.4% died during hospitalization, 52.5% died within 30 days, 66.3% died within 60 days, and 74.3% had died within 90 days. Also, 47.5% died as a result of their HO-SA BSI, and compared to HO-MSSA BSI, this rate was significantly higher among those with MRSA LabID-BSI (63.4% vs 36.7%; OR, 2.99; 95% CI, 1.31–6.83; P = .02). Conclusions: Among patients with HO-SA BSI, methicillin-resistance continues to be associated with higher attributable mortality, and in our study, higher rates of 30-day readmission. There has been no significant change in HO-SA BSI rates (MSSA or MRSA) since reporting for MRSA LabID events began. Furthermore, mortality rates have not changed and remain high for both MRSA BSI and MSSA BSI patients. Given these findings, MSSA LabID event reporting should be considered.

Funding: None

Disclosures: None

Background: Carbapenem-resistant Enterobacteriaceae (CRE) are classified as an urgent antibiotic-resistant threat by the CDC, and they are listed on the critical priority list by the World Health Organization due to the lack of antibiotic treatment options. New-Delhi metallo-β-lactamase (NDM) is an emerging mechanism of carbapenem resistance in the United States. We sought to understand the risk factors and clinical characteristics of patients with NDM CRE in Michigan to improve surveillance. Methods:A retrospective descriptive study was conducted in collaboration with the Michigan Department of Health and Human Services (MDHHS). CRE isolates submitted to MDHHS between April 2014 and July 2019 were tested for the presence of NDM using CDC PCR protocols. Additional information on case demographics, laboratory results, healthcare and antibiotic exposure history, and travel history were collected. Results: In total, 30 NDM cases were identified in Michigan during the study period. Of these 30 cases, 15 (50%) were men, and the median age was 73.5 years (range, 20–88 20). Also, 2 of these patients (6.6%) were immunocompromised; 2 patients (6.6%) had had extensive abdominal surgery, and 2 patients (6.6%) had recurrent hospitalization. Furthermore, 12 case isolates (40%) were collected in outpatient settings, whereas 16 (53%) were collected from inpatient settings. In addition, 13 (43%) patients were admitted from home and 4 (13%) presented from an extended-care facility. Urine was the most common site of isolation in 19 of 30 (63%) cases, followed by blood and tissue culture in 4 of 30 (13%) each. Escherichia coli was the most common organism (17 of 30, 57%), followed by Klebsiella pneumoniae (9 of 30, 30%). Also, 15 of 30 cases (50%) had a recent history of international travel, and of these, 9 of 15 (60%) reported travel to India. Among these 15 cases, 12 (80%) sought medical care in the countries they visited. Two cases (6.6%) had a documented history of multidrug-resistant organism colonization or infection. The mortality rate was 6.6% (2 of 30). The mean time from admission to implementation of contact precautions was 7.3 days (range, 0–20). Conclusions: Suspicion of NDM CRE strains should remain high in patients with a travel history from areas known as major reservoirs of NDM. Delay in implementing contact precautions, as noted in the present study, can lead to a greater risk of transmission. Early detection and subsequent isolation of NDM patients are essential strategies for preventing transmission within healthcare facilities. Future efforts include performing whole-genome sequencing of these isolates to assist in identifying potential epidemiological links among the affected patients.

Funding: None

Disclosures: None

Background: Antibiotic prophylaxis choice for transrectal prostate biopsy (TRPB) has been affected by the emergence of fluoroquinolone-resistant Escherichia coli (FQRE). Prebiopsy FQRE screening and targeted antibiotic prophylaxis may reduce post-TRPB bloodstream infection (BSI). We assessed the impact of a FQRE screening program on post-TRPB BSIs at an academic medical center. Methods: We implemented a FQRE screening program and targeted TRPB antimicrobial prophylaxis guidelines on May 1, 2017 (Fig. 1). We performed a retrospective cohort study of all TRPB and compared the incidence of post-TRPB BSI (within 7 calendar days) per 100 procedures before the intervention (January, 1, 2016, to April 30, 2017) and to the incidence after the intervention (May 1, 2017, to August 31, 2019). We used a subanalysis to compare BSI incidence between patients with positive (+) and negative () FQRE screens and appropriate prophylaxis use, defined as administration of guideline-recommended antibiotics. The Fisher exact test of independence was used to analyze nominal data. Results: The analysis included 2,157 TRPB procedures: 647 in the preintervention period and 1,510 in the postintervention period. FQRE screening compliance was 61% (n = 914) in the postintervention group (Fig. 2); 168 FQRE screens (18%) were positive. The median time from FQRE screen to procedure was 40 days (IQR, 13–69). Postprocedure BSI rates were higher in than those in the preimplementation group; however, this difference was not statistically significant (0.86 vs 0.46; OR, 2.01; P = .42). Among FQRE-screened patients, BSI rates differed significantly between FQRE+ and FQRE patients (2.98 vs 0.54; OR, 5.67; 95% CI, 1.21–28.94; P = .01). Screened patients receiving appropriate prophylaxis had lower BSI rates than those receiving inappropriate prophylaxis; however, this was not statistically significant (1.10 vs 2.02; OR, 0.54; P = .35). The most common BSI pathogen was E. coli (2 (67%) before implementation and 10 (77%) after implementation). Also, 5 E. coli BSIs (50%) were fluoroquinolone resistant in the postimplementation group compared to 1 (33%) in the preimplementation group. Of 13 postimplementation BSIs, 6 occurred in patients who received aminoglycosides perioperatively; however, all 6 BSI pathogens were aminoglycoside sensitive. Conclusions: Compliance with our FQRE screening program and antimicrobial prophylaxis protocol was moderate. Although pre- and postimplementation differences in BSI rates were not statistically significant, the high failure rate among patients receiving aminoglycosides was concerning and led to a change in TRPB prophylaxis guidelines. Reasons for increased BSI risk among FQRE+ patients may include prophylaxis agent, dose, timing, or other confounding factors associated with drug-resistant pathogens. Facilities implementing FQRE screening protocols should evaluate the efficacy of their program and periodically review screening compliance, prophylaxis dosing and timing adherence, and impact on patient-level outcomes.

Funding: None

Disclosures: None

Background: Hand-hygiene sink drains in healthcare facilities may provide an environment for the survival and dissemination of various multidrug-resistant organisms (MDROs), including carbapenemase-producing Klebsiella pneumoniae (CPKP). We developed a sink model system to establish and test native drinking water biofilms containing CPKP in the p-traps of hand-hygiene sink drains. Methods: A handwashing sink gallery was designed to consist of 6-wall mounted stainless-steel sink basins connected to the same municipal water line. Each sink’s plumbing included a chrome-plated brass p-trap. Healthcare facility conditions were simulated to include handwashing events with the addition of hand-soap and municipal water 4 per day, and nutritional shake (simulating liquid waste) 1 per day. Resultant biofilms in the p-traps of each sink were harvested after 28 days for community analysis. Microbial community analyses were performed on selected biofilm samples using 16S rRNA sequencing of the V4 hypervariable region of genomic DNA. Another experiment evaluated 28-day p-trap biofilm inoculated with CPKP CAV1016 (10 mL 7.010E 7 CFU/mL) and was assessed over 14 days. Heterotrophic plate counts (HPCs) were determined on R2A medium (7 days of incubation at 25C). CPKP was quantified on mEndo selective medium (48 hours of incubation at 36C). Results: Biofilms developed in all p-traps, but biofilm HPC (5.78 mean log CFU/cm2, range 4.35–7.16) and community diversity (15–20 genera per p-trap) varied with sink position. Community analysis showed similarities in bacterial community composition and diversity between sinks 1 and 2, and between sinks 3, 5 and 6, but with differences between the 2 groups. The most abundant family in sinks 3, 5, and 6 was Erythrobacteriaceae (76%, 78%, and 55% of the total reads, respectively), whereas sinks 1 and 2 were dominated by Sphingomonadaceae (63% and 36%) and Methylobacteriaceae (19% and 55%). Also, 16S sequencing revealed the presence of potential opportunistic pathogens in the biofilms, including reads attributed to Pseudomonas and Acinetobacter. CPKP CAV1016 inoculated into 28-day p-trap biofilms colonized and persisted in all 6 sinks for 12 days after inoculation. Conclusions: Despite all 6 sinks sharing an incoming water line, soap, and carbon and energy source, there was a significant variation in the bacterial community composition observed between the sinks. CPKP can colonize and persist in the p-trap biofilms; however, additional work is needed to achieve a reproducible model system. Once this is achieved, the sink gallery will be used to investigate interventions to mitigate colonization or persistence of CPKP in p-trap biofilms.

Funding: None

Disclosures: None

Background: In the field of public health, network models are useful for understanding the spread of both information and infectious diseases. Collecting network data requires determining network boundaries (ie, the entities selected for data collection). These decisions, if not made carefully, have potential outsized downstream effects on study findings. In practice, collaboration and coordination between healthcare organizations are often dictated by historical or geopolitical boundaries (eg, state or county boundaries), which may distort the underlying network under study, and thereby affect the reliability and/or accuracy of the network model. Objective: We compared natural communities in a patient-sharing network with those induced by geopolitical boundaries. Methods: Using data from the Healthcare Cost and Utilization Project (HCUP), we constructed a patient-sharing network among hospitals in California, splitting the data into a training set and a holdout set. We performed edge-betweenness clustering on the training set, and with the holdout set, we compared the resulting partition with the partition by counties using modularity. We also clustered contiguous counties that might function more cohesively together than individually. We performed spatially constrained hierarchical clustering on the network constructed from total patient flow between pairs of counties. The results were again compared via modularity on the holdout set to the county partition. Lastly, we built an individual-based model (IBM) using HCUP and American Hospital Association data to perform epidemic simulations. For each of several counties, we implemented this model to estimate the proportion of patients infected over time. We then reran the individual-based model using the entire state while dividing the results into corresponding counties. Results: In total, 680,485 patients transferred between 374 hospitals in 55 counties from 2003 to 2011. The out-of-sample modularity for the edge-betweenness clustering partition was 464% higher than that of the county partition. Aggregating the counties into half as many contiguous clusters was 319% higher, and aggregating them into 6 clusters was 489% higher (Fig. 1). The epicurves from the individual-based model ranged from small to significant deviations between state versus county boundaries (Fig. 2) . Conclusions: Collecting network data using externally imposed boundaries may lead to inaccurate network models. For example, counties serve as a poor proxy for their underlying communities, resulting in poor overall disease spread simulation results when county boundaries are allowed to drive network construction. These issues should be considered when building coordination partnerships such as the Accountable Communities for Health.

Funding: None

Disclosures: None

Background: The ventriculoperitoneal shunt is the main procedure used for to treat communicating hydrocephalus. Surgical site infection associated with the shunt device is the most common complication and a cause of morbidity and mortality of related to the treatment. We sought to answer 3 questions: (1) What is the risk of meningitis after ventricular shunt operations? (2) What are the risk factors for meningitis? (3) What are the main microorganisms causing meningitis? Methods: We conducted a retrospective cohort study of patients undergoing ventricular shunt operations between July 2015 and June 2018 from 12 hospitals at Belo Horizonte, Brazil. Data were gathered by standardized methods defined by the CDC NHSN. Our sample size was 926, and we evaluated 26 preoperative and operative variables by univariate and multivariate analysis. Our outcome variables of interest were meningitis and hospital death. Results: In total, 71 cases of meningitis were diagnosed (risk, 7.7%; 95% CI, 6.1%–9.6%). The mortality rate among patients without infection was 10%, whereas hospital mortality of infected patients was 13% (P = .544). The 3 main risk factors for meningitis after ventricular shunt were identified by logistic regression model: age <2 years (OR, 3.20; P < .001), preoperative hospital stay >4 days (OR, 2.02; P = .007) and >1 surgical procedure, in addition to ventricular shunt (OR, 3.23; P = .043). Almost 1 of 3 of all patients was <2 years old (290, 31%). Also, 430 patients had >4 preoperative days (46%). Patients aged ≥2 years who underwent surgery 4 days after hospital admission had an increased risk of meningitis, from 4% to 6% (P = .140). If a patient <2 years old underwent surgery 4 or more days after hospital admission, the risk of meningitis increased from 9% to 18% (P = .026; Fig. 1). We built a risk index using the number of main risk factors based on a logistic regression model (0, 1, 2 or 3; Fig. 2). Conclusions:We identified 2 intrinsic risk factors for meningitis after ventricular shunt, age <2 years and multiple surgical procedures, and 1 extrinsic risk factor, the preoperative length of hospital stay.

Funding: None

Disclosures: None

Background: In 2016, the European Union adopted unanimously Council Conclusions on the next steps to combat antimicrobial resistance under a One Health approach. To implement some of the provisions laid down in the Council Conclusions, a European Joint Action on Antimicrobial Resistance (AMR) and Healthcare-Associated Infections (HCAI) or EU-JAMRAI was set up, gathering 44 partners. Methods: As part of EU-JAMRAI, 13 participating European countries set up a country-to-country peer review system to evaluate each other’s national action plans (NAPs). This review system entailed a self-assessment, strengths–weaknesses–opportunities–threats (SWOT) analysis, and country visits. All steps were executed with representatives from both the human and the veterinary domains (One Health approach). Special attention was given to supervision and the way supervision can enhance the implementation of guidelines on AMR, both at the policy level and within healthcare institutions. Results: Despite differences in the stage of developing and implementing NAPs, all 13 countries are working on NAPs. In this process, country visits function as a moment to exchange best practices and to provide an outsider’s point of view. At the end of 2019, 13 country-to-country visits had taken place, resulting in tailor-made recommendations for each country. These recommendations were shared with the competent authority. An example is a country that used the recommendation to improve infection prevention as an immediate reason to get the topic on the agenda of the Ministry of Health. During the country visits, intersectoral participation was perceived as desirable, but in some cases it was challenging to arrange. For some highly relevant topics, it has been recognized that discussion should take place on a European level. Examples of such topics include supervision, infection prevention guidelines, funding, surveillance, and regular audits of antibiotic prescriptions for physicians including feedback loops. Conclusions: Peer review is a cooperative and friendly working method compared to common audits. The country visits function as an agenda setting tool to get or to keep AMR on the political agenda and presenting the most relevant topic(s) to address for each country.

Funding: None

Disclosures: None

Background: Contamination of healthcare workers and patient environments likely play a role in the spread of antibiotic-resistant organisms. The mechanisms that contribute to the distribution of organisms within and between patient rooms are not well understood, but they may include movement patterns and patient interactions of healthcare workers. We used an innovative technology for tracking healthcare worker movement and patient interactions in ICUs. Methods: The Kinect system, a device developed by Microsoft, was used to detect the location of a person’s hands and head over time, each represented with 3-dimensional coordinates. The Kinects were deployed in 2 intensive care units (ICUs), at 2 different hospitals, and they collected data from 5 rooms in a high-acuity 20-bed cardiovascular ICU (unit 1) and 3 rooms in a 10-bed medical-surgical ICU (unit 2). The length of the Kinect deployment varied by room (range, 15–48 days). The Kinect data were processed to included date, time, and location of head and hands for all individuals. Based on the coordinates of the bed, we defined events indicating bed touch, distance 30 cm (1 foot) from the bed, and distance 1 m (3 feet) from the bed. The processed Kinect data were then used to generate heat maps showing density of person locations within a room and summarizing bed touches and time spent in different locations within the room. Results: The Kinect systems captured In total, 2,090 hours of room occupancy by at least 1 person within ~1 m of the bed (Table 1). Approximately half of the time spent within ~1 m from the bed was at the bedside (within ~30 cm). The estimated number of bed touches per hour when within ~1 m was 13–23. Patients spent more time on one side of the bed, which varied by room and facility (Fig. 1A, 1B). Additionally, we observed temporal variation in intensity measured by person time in the room (Fig. 1C, 1D). Conclusions: High occupancy tends to be on the far side (away from the door) of the patient bed where the computers are, and the bed touch rate is relatively high. These results can be used to help us understand the potential for room contamination, which can contribute to both transmission and infection, and they highlight critical times and locations in the room, with a potential for focused deep cleaning.

Funding: None

Disclosures: None

Background: The prevalence of healthcare-acquired infections (HAIs) and rising levels of antimicrobial resistance place a significant burden on modern healthcare systems. Cultures are typically used to track HAIs; however, culture methods provide limited information and are not applicable to all pathogens. Next-generation sequencing (NGS) can detect and characterize pathogens present within a sample, but few research studies have explored how NGS could be used to detect pathogen transmission events under HAI-relevant scenarios. The objective of this CDC-funded project was to evaluate and correlate sequencing approaches for pathogen transmission with standard culture-based analysis. Methods: We modeled pathogen transfer via hand contact using synthetic skin. These skin coupons were seeded with a community of commensal organisms to mimic the human skin microbiome. Pathogens were added at physiologically relevant high or low levels prior to skin-to-skin contact. The ESKAPE pathogens: E. faecium, S. aureus, K. pneumoniae, A. baumannii, P. aeruginosa, and Enterobacter spp plus C. difficile were employed because they are the most common antibiotic resistant HAIs. Pathogen transfer between skin coupons was measured following direct skin contact and fomite surface transmission. The effects of handwashing or fomite decontamination were also evaluated. Transferred pathogens were enumerated via culture to establish a robust data set against which DNA and RNA sequence analyses of the same samples could be compared. These data also provide a quantitative assessment of individual ESKAPE+C pathogen transfer rates in skin contact scenarios. Results: Metagenomic and metatranscriptomic analysis using custom analysis pipelines and reference databases successfully identified the commensal and pathogenic organisms present in each sample at the species level. This analysis also identified antibiotic resistance genes and plasmids. Metatranscriptomic analysis permitted not only gene identification but also confirmation of gene expression, a critical factor in the evaluation of antibiotic resistance. DNA analysis does not require cell viability, a key differentiator between sequencing and culturing reflected in simulated handwashing data. Sensitivity remains a key limitation of metagenomic analysis, as shown by the poor species identification and gene content characterization of pathogens present at low abundance within the simulated microbial community. Species level identification typically failed as ratios fell below 1:1,000 pathogen CFU:total community CFU. Conclusions: These findings demonstrate the strengths and weaknesses of NGS for molecular epidemiology. The data sets produced for this study are publicly available so they can be employed for future metagenomic benchmarking studies.

Funding: None

Disclosures: None

Background: The use of real-time polymerase chain reaction (RT-PCR) as a first-line test for the diagnosis of Clostridioides difficile may result in overdiagnosis and overtreatment because the test is not capable of distinguishing infection from carriage. Toxin EIA assays have impeditive low sensitivity. Some algorithms using enzyme immunoassay for glutamate dehydrogenase (GDH) antigen and toxins A and B as the first step have been proposed to increase diagnostic performance. However, cost-effectiveness of different diagnostic algorithms would depend on the cost of each test and on the pretest probability in different settings. The objective of the present study was to evaluate the cost-effectiveness of 2 algorithms proposed by current guidelines to diagnose C. difficile infection by developing a mathematical model that would take into account the epidemiology and costs in our hospital. Methods: The study was conducted in a 480-bed tertiary-care teaching hospital in So Paulo, Brazil. All suspected C. difficile infection cases from January to December of 2017 were evaluated for pretest probability analysis. All stools collected from patients with a requested PCR test for suspected C. difficile infection were selected for additional testing to measure the specificity and sensitivity of each different test: C. diff GDH/Toxin A/B combined test, Toxin A/B Microplate Assay, GDH, and PCR. Toxigenic stool culture for C. difficile was considered the gold standard. A mathematical model was developed and simulations were done. The outcomes evaluated were: final annual costs with diagnostic tests in US dollars and number of patients receiving a false-positive or a false-negative diagnosis in a year simulation. Results: In total, 1,441 stool samples were tested by PCR for C. difficile in our institution from January 2017 to December 2017. Overall, 206 had a positive result, with a pretest probability of 14.3%. In our simulations, the PCR-based algorithm had an annual cost of US$279,914.25, with 4 false-negative results and 8 false-positive results. The implementation of a GDH/Toxin/PCR stepwise algorithm would have reduced the annual cost to US$160,488.75, with 6 false-negative results and 1 false-positive result. Simulations of annual cost and performance of the 2 algorithms have shown that the stepwise algorithm would still be advantageous in settings with higher pretest probabilities (Fig. 1). Conclusions: A stepwise algorithm based on GDH/Toxin before PCR seems to be more cost-effective, even in settings with higher pretest probabilities.

Funding: None

Disclosures: None

Background: One of the limitations of current cleaning and disinfection strategies is that cleaned surfaces rapidly become recontaminated. In laboratory testing, a novel quaternary ammonium disinfectant provided sustained antimicrobial activity against multiple pathogens on surfaces after 24 hours. We hypothesized that this continuously active disinfectant would be effective in reducing contamination of portable medical equipment in a real-world healthcare setting.

Methods: In a hospital and affiliated long-term care facility, 114 portable devices were randomized to receive no treatment (N = 38) or a single spray application of a quaternary ammonium-alcohol disinfectant (N = 38) or of the continuously active disinfectant (N = 38). The devices were cultured at baseline and on days 1, 4, and 7 after treatment for total aerobic colony counts, methicillin-resistant Staphylococcus aureus (MRSA), and enterococci.

Results: As shown in Fig. 1, both spray disinfectants significantly reduced total aerobic colony counts in comparison to the untreated controls. The continuously active disinfectant resulted in sustained significant reductions in aerobic colony counts in comparison to baseline levels (P < .05), whereas counts returned to baseline levels by day 4 in the quaternary ammonium-alcohol disinfectant group. Recovery of MRSA and enterococci was significantly reduced on days 1–7 in the continuously active disinfectant group versus untreated controls (3 of 93, 3% vs 20 of 97, 21% respectively; P = .002), but not in the quaternary ammonium-alcohol disinfectant (11 of 97, 11%; P = .12). Conclusions: A single spray application of a continuously active disinfectant resulted in sustained reductions in total aerobic colony counts over 7 days and reduced recovery of MRSA and enterococci. The continuously active disinfectant could potentially reduce the risk for transmission of pathogens by portable devices.

Funding: None

Disclosures: None

Background: Urinary tract infections (UTIs) are often misdiagnosed and mismanaged. Disease state stewardship initiatives targeting UTIs through the development of institutional guidelines and real-time prospective audit and feedback (PAAF) on provider management may have a significant impact on the overuse of antimicrobials. Objective: Our study evaluated the effectiveness of a UTI focused disease state stewardship intervention by assessing institutional guideline adherence before and after implementation. Methods: This retrospective quasi-experimental study was conducted at a tertiary-care academic medical center. Patients >18 years of age receiving antimicrobials for a UTI were included. A previously performed retrospective review of UTI management from September-November 2017 was used as the baseline. The UTI management guideline was implemented in July 2018, and service lines were educated. A PAAF initiative began in June 2019, whereby the antimicrobial stewardship team performed daily reviews of patients receiving antimicrobials for UTIs. Data was collected on their management, and providers were contacted in real time with recommendations based on the institutional guideline. Patients reviewed June–October 2019 were included in the postimplementation analysis. Patients were excluded if they were pregnant, underwent a urological procedure with risk of mucosal bleeding, or were an outside hospital transfer already on UTI therapy. The primary outcome of this study was to evaluate guideline adherence before and after the implementation of PAAF for the management of UTIs. Results: In total, 198 patients in the preintervention group and 246 in the PAAF group were included. The emergency department was the primary ordering service of urinalyses (60.1% vs 66.1%; P = .2287) in both periods and altered mental status as the main indication for testing (35.2% vs 31.3%; P = .5465). Treatment of asymptomatic bacteriuria and pyuria decreased significantly between the 2 periods: 74.8% versus 36.2% (P = .0001). Appropriate ordering of urinalyses (33.8% vs 68.3%; P = .0001) and urine cultures (29.3% vs 61.0%; P = .0001) also improved in the PAAF group. Recommendations made during PAAF included therapy discontinuation (66.7%), antimicrobial therapy change (15.5%), or duration modification (15.5%), and 59.5% of first interventions were accepted. Overall guideline compliance significantly improved from 13.1% in the preintervention period to 26.1% in the PAAF period (P = .0011). Conclusions: A UTI disease state intervention was associated with significant reductions in the treatment of asymptomatic presentations as well as an improvement in overall guideline adherence. We believe that this approach represents a powerful stewardship strategy for decreasing unnecessary antimicrobial usage.

Funding: None

Disclosures: None

Background: Since 2009, Vidant Health has used a Sick Employee Online Log (SEOL) system to track illnesses among employees and to capture this information in real time. The CDC assessed the 2017–2018 influenza season as a high-severity influenza season, whereas the 2018–2019 influenza season was of moderate severity. Objective: In this research project, we sought to determine whether the influenza season severity would affect either the hospital-based employee illness surveillance system results or would correlate with state influenza-like illness (ILI) visits. Methods: The SEOL system is an internet-based system initiated in December 2008. When a hospital employee calls in sick, the department manager records whether the employee reports the following symptoms: nausea, vomiting, diarrhea, upper respiratory infection, fever, sore throat, headache, conjunctivitis, rash, and/or cough. The information is confidential, with raw data access restricted to review by occupational health and infection control leadership. The correlation value was determined for each symptom using the North Carolina Division of Human Services (NC DHHS) percentage of ILI visits in statewide emergency departments.1 The data collection dates covered January 1–May 31 for each year. In this study, only symptoms related to influenza were included: upper respiratory infection, fever, influenza-like illness, cough and self-reported influenza. Correlation values were calculated using MS Excel software. Results: There were no breaks in confidentiality. All of the correlation values had a correlation value of 0.5 or better (Fig. 1), showing good correlation with the NC DHHS ILI data for both years; however, the more severe 2017–2018 influenza season had correlation values of 0.7 for all symptoms, versus 0.52–0.59 for URI and ILI, respectively, only during 2018–2019. Conclusions: The higher-severity influenza season did correlate with a higher r values when compared to North Carolina’s DHHS ILI emergency department data than did the influenza season of moderate severity. A possible explanation is that a higher-severity influenza season would correlate better than a moderate influenza season because it shows fewer ILI peaks and troughs. In conclusion, the SEOL system served as an early warning that influenza is present among our staff, and it correlates well with the state system for ILI surveillance. Potential limitations of SEOL are that respiratory symptoms are not specific to influenza; thus, they are subject to variation due to other respiratory viruses circulating among our employees.

1. The North Carolina Disease Event Tracking and Epidemiologic Collection Tool. NC DETECT website. http://www.ncdetect.org. Accessed Nov 8, 2019.

Funding: None

Disclosures: None

Objective: The purpose of this study was to evaluate antibiotic use in patients undergoing urological procedures. Methodology: This single-center, IRB-approved, retrospective, observational study was conducted at Grady Health System. Patients were included if they underwent their first inpatient urologic procedure between April 1, 2016, and April 1, 2018. Patients were excluded if they were <18 years old, pregnant, or a prisoner. The primary outcome was percentage of overall adherence to our institutional guidelines for surgical prophylaxis as a composite of antibiotic selection, dose, preoperative timing, and postoperative duration. Secondary outcomes include individual components of the composite outcome, nephrotoxicity, Clostridium difficile infection, and discharge antibiotic prescriptions. Descriptive statistics were used. Results: Of the 100 patients evaluated, 11% achieved adherence with the primary outcome. Of the 89 patients who did not achieve composite outcome, only 8 selected the appropriate perioperative antibiotic. Overall, 30% were dosed appropriately, 47% were administered at the appropriate time with respect to time of incision, and 46% received perioperative antibiotics for no more than 24 hours. Also, 19 patients did not receive perioperative antibiotics. Overall, 14 different perioperative antibiotic regimens were utilized, despite institutional guidelines recommending 1 of 3 options. All 9 patients who developed nephrotoxicity received noncompliant perioperative prophylaxis. No patient developed Clostridium difficile infection within 30 days of surgery. Moreover, 58 patients were discharged with a prescription for at least 1 antibiotic. Conclusions: Most perioperative antibiotic prophylaxes for genitourinary surgeries are not compliant with institution guideline recommendations. Despite having institutional guidelines, there was a large variety in the antibiotic regimens that patients received. All of the patients identified as having an evaluated antibiotic-related adverse effect did not receive appropriate perioperative antibiotic prophylaxis. More than half of the patients received a prescription at discharge for at least 1 antibiotic.

Funding: None

Disclosures: None