Background: Pathogens that remain on healthcare surfaces after standard cleaning have been reported to cause 10%–30% of all healthcare-associated infections among admitted patients. The study reported here evaluated the effect of a continuous application of low-level gaseous hydrogen peroxide (H2O2) on microbial surface contamination in a pulmonary specialty unit (PSU) of a large magnet-designated urban hospital. Methods: A baseline surface contamination level was measured by obtaining cultures of 5 high-touch points in 8 patient rooms, a nurse’s station, and a nurse’s charting room to determine the bacterial and fungal microbial burden levels (period 1). No revisions to cleaning practices were made during any period. Also, 8 continuous decontamination devices were installed in the HVAC near the area to be treated. The devices convert humidified air into predominately H2O2, which then exited the ducts, covering all surfaces. Environmental sampling was conducted every 4 weeks for 4 months after activation to measure the effects on the intrinsic microbial burden on surfaces as well as the incidence of MRSA and VRE (periods 2–5). In addition, the unit manager tracked the level of absenteeism of employees in the PCU during the trial to compare results to the same 4-month period from the prior year. Results: In concert with regular cleaning and disinfecting practices, the average bacterial microbial burden levels found on 52 locations prior to the activation of the devices were 6,446 CFU/1002. Incidence of MRSA and VRE were 31% and 10%, respectively. During the intervention, the levels of both bacterial and fungal microbial burden on surfaces were significantly lower, as was the incidence of MRSA and VRE. The continuous application of low-levels of oxidizing molecules exerted a significant 78% and 97% reduction in the average bacterial and fungal microbial burden found on PSU surfaces (1,415 CFU/100 2 and 328 CFU/1002 for bacteria and fungi, respectively). The incidences of MRSA and VRE on surfaces were reduced by 63% and 70%, respectively. In addition, the unit manager reported a significant decline in absenteeism, with a reduction of 550 hours (1,313 vs 752) during the 4-month trial versus the same 4-month period the previous year. Conclusions: The introduction of continuous low-level gaseous H2O2 on all surfaces throughout a PSU reduced the overall bacterial and fungal microbial burden as well as lowered the incidence of both MRSA and VRE on surfaces, and it reduced employee absenteeism, providing a potentially safer environment for patients, staff, and visitors.

Funding: None

Disclosures: None

Background: One of the main global public health challenges is the fight against microbial resistance, according to the World Health Organization. Inadequate use of antimicrobials is considered one of the main factors related to the phenomenon and is quite common in the hospital environment. Managing the use of antimicrobials in hospitals has become a necessity and has shown positive results in many ways, such as maximizing the effects of pharmacotherapy, preventing the emergence of resistant microorganisms, and reducing healthcare costs. Methods: The prescriptions for patients admitted to a 380-bed nonprofit private hospital in Belo Horizonte, Brazil were monitored from January 1, 2019, to August 31, 2019, with a monthly average of 251 patients followed by the antimicrobial stewardship (AMS) team (1 infectious diseases doctor and 2 clinical pharmacists). Patients selected for follow-up and intervention were those submitted to intravenous, intramuscular, and/or oral antibiotic therapy with the following antimicrobial agents: piperacillin/tazobactam, carbapenem, polymyxin B, tigecycline, vancomycin, teicoplanin, daptomycin, third- and fourth-generation cephalosporins, quinolone, and aminoglycosides. Patients on prophylactic or antimicrobial treatment not mentioned above were excluded from surveillance. Interventions were dose adjustments, drug adjustment by culture results, intravenous to oral treatment switch, and discontinuation of therapy. Results: There were 318 interventions, and 64.82% of the interventions performed by the AMS team were accepted by prescribers. The interventions provided a total savings of BR$ 119,706 (~US$30,000) in direct antimicrobial spending. Correlating the interventions with the defined daily dose (DDD) measurement and comparing data from the same period in 2018, we detected a reductions in the consumption of several antimicrobials: ceftriaxone (25.6%), ciprofloxacin (45.7%), meropenem (34%), piperacillin/tazobactam (12.7%), teicoplanin (18.8%), vancomycin (20.6%), cefepime (23.9%) and polymyxin B (26%). We also detected reductions in days of therapy (DOT) for most of these drugs, such as polymyxin B, with an average reduction of 2 DOT. Conclusions: Reducing antimicrobial use is one of the key strategies for avoiding unnecessary exposure and selective pressure leading to the emergence of resistant microorganisms. The measured data point to a favorable trend in the rational use of antimicrobials in our institution with simple interventions. The results presented were used to reaffirm the importance of the AMS team in our institution. More data on length of stay, indirect costs, reduction of side effects, mortality, and occurrence of microbial resistance should be made.

Funding: None

Disclosures: None

Background: Toys in playrooms are often shared among patients in pediatric healthcare settings; they can present a risk for transmission of bacterial and viral pathogens. Effective cleaning and disinfection of toys using disinfectant wipes is labor intensive and difficult due to irregular surfaces. Methods: We conducted a point-prevalence culture survey to determine the frequency of contamination of in-use toys and high-touch surfaces in playrooms in a pediatric healthcare facility with methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), and Clostridioides difficile. Using a variety of toys inoculated with pathogens, we evaluated efficacy and ease-of-use of 3 novel “no-touch” technologies: (1) an electrostatic sprayer, (2) a small ultraviolet-C (UV-C) box (18.9 × 9.9 × 1.8 inches) for smaller toys, and (3) a high-level disinfection cabinet using ultrasonic submicron droplets of peracetic acid and hydrogen peroxide. Test pathogens included C. difficile, MRSA, and Candida auris. Results: Of 135 items cultured in playrooms, 6 (4.4%) were contaminated with MRSA, 1 (0.7%) was contaminated with VRE, and none were contaminated with C. difficile. Each of the technologies reduced all pathogens by >4 log10 CFU on all types of toys tested (plastic, soft rubber, and tablet). The electrostatic sprayer was considered the easiest to use by all users because large numbers of toys could be processed much more quickly (ie, spray for 20 seconds and allow to air dry) than with disinfectant wipes. The disinfection cabinet required 21 minutes for cycle completion, whereas the decontamination cycle for the UV box was only 30–90 seconds but with limited capacity to hold toys. Conclusions: Three “no-touch” technologies were effective for disinfection of toys contaminated with healthcare-associated pathogens. The electrostatic spray application of disinfectant was considered the easiest to use for rapid decontamination of toys.

Funding: None

Disclosures: None

Background: Carbapenemase-producing organisms (CPOs) are a growing antibiotic resistance threat. Colonization screening can be used to identify asymptomatically colonized individuals for implementation of transmission-based precautions. Identifying high-risk patients and settings to prioritize screening recommendations can preserve facility resources. To inform screening recommendations, we analyzed CPO admission screens and screening conducted on point-prevalence surveys (PPSs) performed through the Antibiotic Resistance Laboratory Network’s Southeast Regional Laboratory (SE AR Lab Network). Methods: During 2017–2019, the SE AR Lab Network collected data via a REDCap survey for a subset of CPO screens on a limited set of easily determined patient risk factors. Rectal swabs were collected and tested with the Cepheid Carba-R. Specimens collected within 2 days of admission were classified as admission screening and the remainder were classified as PPS. Index cases were excluded from analyses. Odd ratios (ORs) and 95% confidence intervals were calculated, and a value of 0.1 was used for cells with a value of zero. Results: In total, 520 screens were conducted, which included 366 admission screens at 2 facilities and 154 screens from 27 PPSs at 8 facilities. CPOs were detected in 14 (2.7%) screens, including in 10 (2.7%) admission screens and in 4 (2.6%) contacts during PPSs; carbapenemases detected were Klebsiella pneumoniae carbapenemase (KPC) (n = 12), New Delhi Metallo-β-lactamase (NDM) (n = 1) and Verona Integron-Encoded Metallo-β-lactamase (VIM) (n = 1). One long-term acute care hospital (LTACH) performed universal admission screening, which accounted for 96% of admission screens and all 10 CPOs detected by admission screening. Mechanical ventilation (OR, 5.0; 95% CI, 1.4–18.0) and the presence of a tracheostomy (OR, 5.4; 95% CI, 1.5–19.4) were associated with a positive admission screen. Moreover, 8 facilities conducted PPSs: 4 acute care hospitals, 2 long-term acute care hospitals, and 2 nursing homes. CPO prevalence in long-term acute care hospitals was 4.8% (2 of 42), 2.4% (1 of 41) in acute care hospitals, and 1.5% (1 of 69) in nursing homes. Requiring assistance with bathing (OR, 4.8; 95% CI, 1.6–8.0) and stool incontinence (OR, 16.6; 95% CI, 13.4–19.8) were associated with a positive screen on PPSs. All 7 roommates of known cases tested negative for CPO colonization. Conclusions: Findings suggest that patients with certain easily assessed characteristics, such as mechanical ventilation, tracheostomy, or stool incontinence or who require bathing assistance, may be associated with CPO positivity during screening. Further data collection and analysis of such risk factors may provide insight for the development of more targeted admission and contact screening strategies.

Funding: None

Disclosures: None

Background: Contaminated healthcare surfaces can serve as reservoirs for the transmission of pathogens. Sensitive sampling methods are needed to investigate sources of pathogens for implementing effective disinfection strategies and thereby preventing environmental transmission. Conventional approaches employ swabs to sample environmental surfaces. Contact plates represent an alternative approach for sampling healthcare surfaces that does not require lab processing, though little is known about their performance. A contact plate is an agar plate that is overfilled with selective or nonselective media. It can be gently applied to the surface, then simply incubated at a temperature optimal for target organism (s), thus saving time and resources. Methods: In this study, contact plates containing trypticase soy agar with 5% sheep blood (TSAII), foam swabs, and flocked swabs were evaluated for their ability to recover 4 pathogens that persist on healthcare surfaces. Stainless-steel coupons (4 in2) were inoculated with the following pathogens (102 CFU): Acinetobacter baumannii (AB, strain type 12), carbapenemase-producing KPC+ Klebsiella pneumoniae (KP; ATCC BAA-1705); methicillin-resistant Staphylococcus aureus (MRSA; ATCC 43300); and vancomycin-resistant Enterococcus faecalis (VRE; Van A + 256). The plates were allowed to dry 1 hour. Sampling with CPs was performed in 2 ways; (1) a single contact plate was used to sample 1 stainless-steel surface and (2) a composite was collected by 3 sequential contact-plate samplings of the same stainless-steel surface. The contact plates were then incubated at 351C. Foam and flocked swabs were premoistened with phosphate-buffered saline + 0.02% polysorbate 80 (PBST) and were used to sample the stainless-steel coupons. Swabs were held for 1 hour and processed by sonication and vortexing in 5 mL of PBST, then the eluent was cultured and CFU counted. Mean percentage recoveries (%R) relative to the inoculum were calculated and compared. Results: When the %R for all 4 pathogens were pooled, the composite contact-plate sampling method yielded the highest, (P < .05) (66.0%; SD, 0.22), followed by the single contact plate method (39.7%; SD, 0.12), foam swab (32.9%; SD, 0.18), and flocked swab (20.3%; SD, 0.20). The composite contact plate method yielded the highest %R for VRE (102.1 %; SD, 0.17), and the lowest %R was observed when using flocked swabs to recover KP (6.3%; SD, 0.05). Conclusions: The contact-plate composite method may provide investigators with minimal environmental microbiology capacity an alternative method for environmental sampling and detection of organisms from surface areas (4 in 2) with low bioburden.

Funding: None

Disclosures: None

Background: In October 2013, the University of Maryland Medical Center established a formal antibiotic prophylaxis protocol for patients undergoing ventricular assist device (VAD) placement, replacing a previous system of various broad-spectrum antibiotic combinations typically for prolonged durations based on surgeon preference. This new protocol consisted of a standardized regimen of 72 hours of vancomycin and ceftriaxone after the procedure. The objective of this project was to evaluate the rate of surgical site infection (SSI) related to VAD placement to ensure that implementing the new protocol did not cause an increase in SSI rates. Methods: The study was a retrospective cohort study of patients who had undergone VAD placement before the protocol change (January 1, 2011, to October 1, 2013) and after the change (October 1, 2013, to November 15, 2015). The primary outcomes was the difference in SSI rate before and after the protocol change using CDC NHSN definitions. Pertinent data points of interest included reason for VAD placement, duration/type of antibiotics used, delayed sternal closure, SSI, characterization of infection (bloodstream, driveline, or pocket), organism identified on culture and mortality at 30 days and 1 year. SSI rates were assessed using the Fischer exact test, and descriptive statistics were used for other outcome variables. Results: In total, 75 patients were included before the protocol and 46 after the protocol change. Overall, 27% and 17% of patients were on therapeutic antibiotics prior to the VAD placement, respectively (P = 0.23). Also, 8 (6.6%) patients in the preintervention group had an SSI compared to 1 patient (0.8%) in the postintervention group (P = .15). Adherence to the protocol was suboptimal, with 27% of patients in the postintervention group receiving non–protocol-adherent antibiotics and 65% of patients receiving antibiotics >96 hours postoperatively. When evaluating the patients collectively, SSI rates were the same when antibiotics were discontinued <72 hours postoperatively versus when antibiotics were continued beyond 72 hours postoperatively or were not given at all postoperatively (3.1% vs 10.7% vs 0%; P = .24). SSI rates were also no different among patients who received cefazolin monotherapy (0%), vancomycin and ceftriaxone (2.7%), vancomycin and piperacillin tazobactam (2%), and other antibiotic combinations (7.7%) for surgical prophylaxis (P = 0.1). Conclusions: No change in SSI rates was noted after a protocol change narrowing the spectrum and duration of antibiotic prophylaxis was implemented. Evaluation of optimal surgical prophylaxis in this patient population is difficult due to low event rates and frequent therapeutic indications for antibiotics outside the standard prophylaxis. Despite these challenges, this study supports the safety of studying SSI prophylaxis reduction in the VAD population. Further studies are reasonable and warranted.

Funding: None

Disclosures: None

Background: Whole-genome sequencing (WGS) is becoming the method of choice for outbreak analysis of microbial pathogens. However, the main challenge with WGS for microbial strain typing is the conversion of raw sequencing data to actionable results for epidemiology and surveillance analysis. We evaluated the bioMrieux EPISEQ-CS, a cloud-based WGS data analysis software for outbreak detection to compare the results for 4 groups of different species previously characterized by strain typing and commonly isolated in hospital-acquired infections. Methods: In total, 30 methicillin-resistant Staphylococcus aureus (MRSA), 15 Clostridioides difficile (CDIFF), 17 Pseudomonas aeruginosa (PSA), and 10 Acinetobacter baumannii (ACB) isolates were included in this study. All isolates had been previously characterized by rep-PCR using the DiversiLab system (bioMrieux, France) and saved at 70C. Before testing, samples were thawed and plated, and DNA extraction was performed on the QIAcube (Qiagen, Hilden, Germany) using the DNEasy Ultra Clean Microbial kit extraction protocol. DNA libraries were prepared using the Nextera DNA Flex Kit and sequenced on the Illumina iSeq100 platform according to manufacturer’s recommendations (Illumina, San Diego, CA). Generated sequences were uploaded into EPISEQ-CS, and wgMLST-based analysis was performed. We compared clusters generated by the DiversiLab system and EPISEQ-CS. Results: DiversiLab identified 9 MRSA clusters among 30 isolates. EPISEQ-CS reclassified 14 of 30 isolates into 5 MRSA clusters and the remaining 16 isolates were unrelated. DiversiLab identified 2 CDIFF clusters among 15 isolates. EPISEQ-CS reclassified 3 isolates into 1 CDIFF cluster and determined the remaining 12 to be unrelated. DiversiLab identified 5 PSA clusters among 17 isolates, whereas EPISEQ-CS reclassified all 17 isolates as unrelated. DiversiLab identified 2 ACB clusters among 10 isolates, whereas EPISEQ-CS reclassified 2 ACB isolates into 1 cluster and determined 8 to be unrelated. Analysis using Simpson’s diversity index (D) suggested that the EPISEQ-CS showed increased diversity when compared to DiversiLab clustering across all bacterial species analyzed in this study. Conclusions: EPISEQ-CS enabled a comprehensive wgMLST analysis, including quality control and comparative epidemiological analysis, thereby providing a more reliable method for bacterial strain typing. As WGS becomes more affordable and applicable to routine epidemiological surveillance, EPISEQ-CS provides an informative tool in the monitoring of hospital-acquired infections.

Funding: None

Disclosures: None

Background: Hand hygiene (HH) is the most important measure for preventing healthcare-associated infections. The objective is to gain insight into the evolution of the degree of compliance with recommendations (DCR) on HH and its associated factors in the surgical areas of a tertiary-care hospital. Methods: This observational, cross-sectional study, was repeated over time, with direct observation of the DCR on HH during the daily activity of healthcare workers in surgical areas: general surgery, urology, vascular surgery, traumatology, neurosurgery, thoracic surgery, heart surgery, pediatric surgery, otorhinolaryngology, gynecology and obstetrics, ophthalmology. Over 14 years (from 2005 to 2018), 15,946 HH opportunities were registered, together with different additional variables (age, sex, professional position, surgical area ). The 2 test was used to study the association and the crude, and adjusted odds ratios were used to quantify its magnitude. Results: The DCR on HH in surgical areas was 49.7% (95% CI, 48.9%–50.5%), and in the group of nonsurgical areas it was 53.4% (95% CI, 53.1%–54.1%). The area with the highest degree of compliance was urology (56.7%; 95% CI, 53.9%–59.6%), and the area with the lowest degree of compliance was traumatology (43.3%; 95% CI, 40.4%–46.2%). Some associated factors were the indications after an activity has been performed (58.6%; aOR, 2.7; 95% CI, 2.5–2.9) and the availability of pocket-size alcohol-based disinfectant (63.8%; aOR, 2.4; 95% CI, 2.2–2.5). Conclusions: The DCR on HH in surgical areas is lower than in other hospital areas, and there is still some margin for improvement. We have identified some modifiable factors that have an independent association with HH compliance in surgical areas. Focusing on them will increase compliance with HH with the ultimate goal of reducing healthcare-associated infections.

Funding: None

Disclosures: None

Background: Domestically, the integration of public health into healthcare-associated infection (HAI) and antibiotic resistance (AR) prevention activities represents a major development. We describe CDC Funding: of public health HAI/AR programs through the Epidemiology and Laboratory Capacity (ELC) cooperative agreement to improve local capacity to prevent HAIs and detect and contain the spread of AR threats. Methods: We reviewed ELC budget reports and program documents to summarize the evolution of funded activities and programs from 2009 to 2018. Results: In 2009, 51 programs (49 states, 2 cities and territories) received US$35.8 million through the American Recovery and Reinvestment Act for an initial 28-month period. These funds supported each jurisdiction to establish an HAI coordinator and a multidisciplinary HAI advisory group, coordinate and report HAI prevention efforts, conduct surveillance and report HAI data, and maintain an HAI plan; ~27 programs were also funded to coordinate multicenter HAI prevention collaboratives among acute-care hospitals. Through 2011, 188 state or local HAI/AR program positions were at least partially funded by the CDC. From 2011 to 2015, investments from the Affordable Care Act (~US$10–11 million annually) were used to maintain the HAI/AR programs, with some expansion of program goals related to non–acute-care settings and antibiotic stewardship. In 2015, following the Ebola outbreak in West Africa, supplemental ELC funds were awarded to 61 programs (50 states, 11 cities and territories) totaling US$85 million over 36 months. These awards marked an expansion of HAI/AR program activities to develop healthcare provider inventories, to conduct data-driven education and training, and to perform onsite infection control assessments in healthcare facilities. In 2016, through its AR Solutions Initiative, CDC invested US$57.3 million in Funding: to 57 programs (50 states, 7 cities and territories), expanding laboratory capacities for AR threat detection (via the AR Laboratory Network) and epidemiologic activities to rapidly contain novel and targeted multidrug-resistant organisms. As of 2018, >500 state or local HAI/AR program positions were at least partially funded by the CDC. Conclusions: State and local HAI/AR programs have grown substantially over the 10 years of their existence, as reflected in major increases in funding, staffing, scope, and partnerships. CDC investments and guidance have supported the development of HAI/AR epidemiology prevention and response capacity.

Funding: None

Disclosures: None

Background: Patient sharing between hospitals has long been known to be a contributor to the regional transmission of hospital-acquired infections. This inter–healthcare-facility connectedness suggests that regional, as opposed to local, control and surveillance strategies should be favored. However, the absence of easily and universally accessible patient transfer data hampers researchers and public health agencies who wish to build accurate network models of interfacility transmissions. Medicare data offer only a biased subsample of the full patient transfer network, but because it is widely available, this data source has historically been used for inference and simulation studies. Objective: The purpose of this study was to determine whether Medicare data could successfully be recalibrated to more closely resemble 100% inpatient capture data. Methods: We used data from the Healthcare Cost and Utilization Project (HCUP) to construct 100% capture and Medicare-only patient sharing networks among hospitals in Iowa, Wisconsin, and Nebraska. We used matrix decomposition techniques on the Medicare-only networks along with hospital characteristics from the American Hospital Association (AHA) data for feature construction in a truncated Poisson regression model, and we used Monte Carlo integration to obtain predicted values. These predicted values served as calibrated Medicare-only networks. We split the patient transfer data into training and testing sets and computed the mean squared prediction error (MSPE) for the testing data. We also built an individual based model (IBM) using HCUP and AHA data to perform epidemic simulations that depended on a matrix of patient transfer rates between hospitals. We then compared epicurves from these IBMs resulting from 100% capture networks, Medicare-only networks, and our calibrated networks. Results: Our calibrated networks reduced the MSPE with respect to Medicare-only networks by 84%, 47%, and 88% for Iowa, Wisconsin, and Nebraska, respectively. Although the epicurves from Medicare-only networks differed considerably from that from 100% capture networks, our calibrated networks retained high fidelity to the curves obtained from 100% capture networks. Conclusions: Medicare-only networks greatly underestimate the number of patients transferred between hospitals. Our approach allows us to use Medicare data to estimate networks when 100% inpatient capture is unavailable.

Funding: None

Disclosures: None

Background: The CDC National Healthcare Safety Network (NHSN) is the nation’s most widely used healthcare-associated infection (HAI) and antibiotic use and resistance (AUR) surveillance system. More than 22,000 healthcare facilities report data to the NHSN. The NHSN data are used by facilities, the CDC, health departments, the CMS, among other organizations and agencies. In 2017, the CDC updated the NHSN Agreement to Participate and Consent (Agreement), completed by facilities, broadening health department access to NHSN data and extending eligibility for data use agreements (DUAs) to local and territorial health departments. DUAs enable access to NHSN data reported by facilities in the health department’s jurisdiction and have been available to state health departments since 2011. The updated agreement also enables the CDC to provide NHSN data to health departments for targeted prevention projects outbreak investigations and responses. Methods: We reviewed the current NHSN DUA inventory to assess the extent to which health departments use the NHSN’s new data access provisions and used semistructured interviews with health department staff, conducted via emails, phone, and in person conversations, to identify and describe their NHSN data uses. Results: As of late 2019, the NHSN has DUAs with health departments in 17 states, 7 local health departments (including municipalities and counties), and 1 US territory. The NHSN also has received requests from 2 state health departments for data supporting HAI prevention projects. Health departments with DUAs described improved relationships with facilities in their jurisdictions because of new opportunities to offer NHSN data analysis assistance to facilities. One local health department analyzed their NHSN carbapenem-resistant Enterobacteriaceae (CRE) data to identify (1) facilities in its jurisdiction with comparatively high CRE infection burden and (2) geographic areas to target for a CRE isolate submission program. Outreach to facilities with high CRE burden led to enrollment of 15 clinical laboratories into a voluntary isolate submission program to analyze CRE isolates for additional characterization. Examples of health departments’ use of data for action include: notifying facilities with high standardized infection ratios (SIRs) and sharing Targeted Assessment for Prevention (TAP) reports. Conclusions: The NHSN’s role as a shared surveillance resource has expanded in multiple public health jurisdictions as a result of new data access provisions. Health departments are using NHSN data in their programmatic responses to HAI and AR challenges. New access to NHSN data is enabling public health jurisdictions to assess problems and opportunities, provide guidance for prevention projects, and support program evaluations.

Funding: None

Disclosures: None

Background: Inappropriate antibiotic prescription leads to increased Clostridiodes difficile infections, adverse effects including organ toxicity, and generation of antibiotic-resistant bacteria. Despite efforts to improve antibiotic use in acute-care settings, unnecessary and inappropriate prescription still occur in 30%–50% of patients. Objectives: We assessed factors associated with inappropriate antibiotic prescription at 2 time points: (1) initial, empiric therapy and (2) 3–5 days after therapy initiation. Methods: As part of a multicenter study investigating strategies to reduce antibiotic therapy after 3–5 days of use, antibiotic prescription data were collected from 11 adult and pediatric intensive care and general medical units at 6 hospitals in Maryland in 2014 and 2015. We performed a retrospective cohort study of all hospitalized patients who received any of 23 common antibiotics for at least 3 days. Each medical record was reviewed for demographics, admission and discharge dates, patient comorbidities, and antibiotic regimen by at least 1 infectious disease physician or pharmacist. Classification of antibiotic inappropriateness was based on each institution’s guidelines and standards. Bivariate analyses were performed using logistic regression for both initial therapy and therapy at days 3–5. Multivariable logistic regression was performed using covariates meeting the significance level of P < .05. Results: In total, 3,436 antibiotic courses were assessed at time of initial therapy, and 1541 regimens were continued and reviewed again at days 3–5 of therapy. For the initial therapy, 1,255 regimens (37%) were inappropriate; 45% of these were considered unnecessary and 41% were too broad in spectrum. In the multivariable regression, older age and antibiotic prescription during the summer were associated with the receipt of inappropriate antibiotics (Table 1). Having end-stage renal disease as a comorbid condition was protective against inappropriate use. At days 3–5 of therapy, 688 (45%) of the antibiotic courses were inappropriate. Reasons regimens were considered inappropriate included unnecessary antibiotic prescriptions (49%) and antibiotics being too broad (38%). Older age and receiving cefepime or piperacillin-tazobactam on day 3 of therapy were factors associated with inappropriate use (Table 2). Having undergone a transplant or a surgical procedure was protective of inappropriate antimicrobial use at days 3–5 of therapy. Conclusions: Older patients are more likely to receive inappropriate antibiotics at both initial regimen and 3–5 days later. Patients receiving cefepime or piperacillin-tazobactam are at greater risk of receiving inappropriate antibiotics at days 3–5 due to failure to de-escalate. Antibiotic stewardship strategies targeting these patient populations may limit inappropriate use.

Funding: None

Disclosures: None

Background: An antibiogram is a tool for tracking and reporting antimicrobial resistance; the CDC has endorsed as part of a comprehensive antimicrobial stewardship program in nursing homes. We have previously shown that antibiogram utilization has increased in nursing homes, but most facilities employ tools that are not based on facility-specific data. In this study, we investigate the feasibility to develop antibiograms using facility-specific data and compare these results with antibiograms developed using data from multiple facilities that share the same lab and geographic region. Methods: Raw, de-identified culture results from January 1 through December 31st, 2018 were collected from participating nursing homes and their consulting microbiology laboratories under an IRB-exempt protocol. Culture results were entered and stored in REDCap. Number of isolates per species was examined based on nursing home, nursing home laboratory network, and region. Percentage sensitivities of the most frequently isolated species to commonly used antibiotics were calculated at the nursing home and regional level and compared. T tests of the absolute difference between nursing home- and regional level percentage sensitivities were performed. All data analyses were performed in R software. Results: The mean annual cultures per nursing home was 23.5 (SE, ±3.29). Grouping cultures by lab and region increased the mean culture count 6-fold and 12-fold, respectively. The most commonly isolated species were Escherichia coli (29.7%), Enterococcus spp (11.6%), Proteus spp (10%), Klebsiella spp (8.5%). None of the nursing homes had >30 isolates of a single species (Fig. 1). Escherichia coli was the only species that exceeded the 30-isolate threshold when aggregated at the laboratory network level (Fig. 2). Grouping nursing home cultures by region provided the greatest average isolate count across the most common species. The greatest differences in percentage sensitivity between nursing homes and their region were noted for Escherichia coli and Proteus spp to fluoroquinolones (>20% difference; P < .01). The difference in sensitivity was <5% for Escherichia coli to nitrofurantoin. Conclusions: Nursing homes do not generate enough annual isolates to create antibiograms compliant with Clinical Laboratory Standard Institute guidelines. Grouping isolates from multiple nursing homes at the regional level does reliably exceed the 30-isolate threshold for multiple bacterial species but leads to susceptibility estimates that may vary substantially from those observed at the facility level. Alternative tools for tracking antibiotic resistance and guiding antibiotic prescribing decisions at the local level are needed.

Funding: None

Disclosures: None

Background: Since 1991, US tuberculosis (TB) rates have declined, including among health care personnel (HCP). Non–US born persons accounted for approximately two-thirds of cases. Serial TB testing has limitations in populations at low risk; it is expensive and labor intensive. Method: We moved a large hospital system from facility-level risk stratification to an individual risk model to guide TB screening based on Tuberculosis Screening, Testing, and Treatment of US Health Care Personnel: Recommendations from the National Tuberculosis Controllers Association and CDC, 2019. This process included individual TB risk assessment, symptom evaluation, TB testing for M. tuberculosis infection (by either IGRA or TST) for HCP without documented evidence of prior LTBI or TB disease, with an additional workup for TB disease for HCP with positive test results or symptoms compatible with TB disease. In addition, employees with specific job codes deemed high risk were required to undergo TB screening. Result: In 2018, this hospital system of ~10,000 employees screened 7,556 HCP for TB at a cost of $348,625. In 2019, the cost of the T Spot test increased from $45 to $100 and the cost of screening 5,754 HCP through October 31, 2019, was $543,057. In 2020, it is anticipated that 755 HCP will be screened, saving the hospital an estimated minimum of $467,557. The labor burden associated with employee health personnel will fall from ~629.66 hours to 62.91 hours. The labor burden associated with pulling HCPs from the bedside to be screened will be reduced from 629.66 hours to 62.91 hours as well. Conclusion: Adoption of the individual risk assessment model for TB screening based on Tuberculosis Screening, Testing, and Treatment of US Health Care Personnel: Recommendations from the National Tuberculosis Controllers Association and CDC, 2019 will greatly reduce financial and labor burdens in healthcare settings when implemented.

Funding: None

Disclosures: None

Background: Sepsis is currently one of the important global health issues due to its complexity from pathophysiologic, clinical, and therapeutic viewpoints. Most sepsis-related studies are from the West, where all the patients were grouped together failing to identify specific patient populations that may actually benefit from a particular intervention. We investigated the characteristics and impact of the source of infection on sepsis-related ICU outcomes among critically ill adult patients Methods: A prospective ICU based observational study was conducted over 15 months in a tertiary-care hospital in southern India. Our study included all critically ill patients (18 years old) who were admitted either with existing a new episode of sepsis with suspected or documented bacterial infections within 24 hours of ICU admission with SOFA score 2. Basic demographics, the clinical presentation with the anatomical site of infection and outcome were noted. Categorical variables were compared using the 2 test, and continuous variables were compared using 1-way analysis of variance (ANOVA). Cox regression was used to determine the effect of sepsis source on 28-day mortality. Results: Among the 4,548 patients screened during the study period, 400 were recruited, with a mean age of 55.716 years, among whom 276 (61%) were men. The mean SOFA score at admission was 9.92.7. Bacteremia was observed among 99 cases (24.8%), predominantly gram-negative sepsis (65 of 99, 65.6%). The source for blood culture positivity was determined in 48 of 99 cases (48.4%). Successful isolation of the bacteria was achieved from other anatomical sites in 115 patients (37.8%) where blood culture remained negative. The most common source of sepsis was lung (67 of 400, 16.7%) followed by skin and soft-tissue infection (56 of 400, 14%). Patients treated with steroids were more prone to develop a respiratory infection (P = .001), whereas renal impairment was correlated with urinary tract infection (P = .001). Patients with respiratory infections had a longer ICU stay (P < .001). The overall in-hospital mortality was 37.2%. Multivariable Cox regression showed patients with genitourinary infection (HR, 2.23; P = .04) and implantable devices (HR, 11.30; P = .17) were at higher risk of death. Conclusions: Site-specific infection was a significant predictor of mortality in our study. These factors should be taken into consideration and warrant further evaluation to understand their specific roles in adverse outcomes among a cohort of patients diagnosed with sepsis.

Funding: None

Disclosures: None

Background:Candida auris is an invasive, multidrug-resistant pathogen that can cause outbreaks in hospitals. The mode of transmission is through contaminated hospital items such as fomites and staff interventions. The outbreak of Candida auris affecting 13 patients at the medical intensive care unit (MICU) and medical inpatient unit, either in the form of infection or colonization, is the first documented C. auris outbreak in the Qatar. Methods: The first case was identified in November 2018 in a patient colonized in the respiratory tract. Candida auris biweekly tests were conducted. The second to fourth cases were confirmed in the MICU admitted in the same room (room 2). The fifth case was identified incidentally and was not part of the screening in another ward (6 North Medical Inpatient Unit), and 4 weeks later, Candida auris was isolated from the urine and throat of a patient on this ward. The realization that case 5 was an index case changed the direction of the outbreak investigation, and expanded screening was started among the medical inpatients. When the IPC team identified cases 6–11 and 13, all had possible epidemiological links with case 5. Results: Our root-cause analysis suggests that the index case came from another general hospital. However, because no screening protocol has been established for Candida auris, interventions have not been in place to effectively prevent and control this organism. A strong collaborative outbreak team worked to end this outbreak using the following evidence-based IPC interventions: (1) patient screening and decolonization; (2) environmental screening; (3) enhanced environmental disinfection using peracetic acid wipes, 1% chlorine, and hydrogen peroxide vapor disinfection; (4) prophylactic contact precautions; (5) enhanced hand hygiene with bare below elbows protocol; and (6) a no white gown policy. Conclusions: The outbreak of Candida auris was declared to have nbeen terminated on August 22, 2019. Despite the long period involved in this outbreak, we succeeded in ending it through the concerted efforts of a multidisciplinary team utilizing the latest scientific evidence.

Funding: None

Disclosures: None

Background: Antibiotic overuse has led to increasing rates of antibiotic resistant infections and unnecessary antibiotic costs. Clinical pharmacists can play a key role in optimizing appropriate use of antimicrobials and reducing antimicrobial resistance. However, the role of clinical pharmacists in antimicrobial stewardship is new and not well established in Viet Nam. Objective: We evaluated the use of clinical pharmacists for improved antimicrobial prescribing. Methods: We assembled an antibiotic stewardship program (ASP) team consisting of a clinical pharmacist and a specialist in infection prevention and control in a 60-bed medical intensive care unit (MICU) at Hue Central Hospital in central Viet Nam. During January–September 2018, the ASP team collected baseline antibiotic prescribing days of therapy (DOT) for all antibiotics administered in the MICU. Then, from October 2018 through June 2019, the ASP team reviewed daily positive clinical bacterial cultures and susceptibility results for all patients present in the MICU. They reviewed medical charts, including antimicrobial prescriptions, during week days and only if patient was still in the ICU at the time of ASP rounds. The team recommended changes to antibiotic therapy verbally to physicians and left the decision to change antibiotic therapy to their discretion. The ASP team documented whether their recommendations were accepted or rejected. Statistical significance was determined using the Student t test. Results: The ASP team reviewed 160 medical charts and made 169 ASP recommendations: 122 (72%) to continue current treatment; 24 (14%) to monitor drug levels or obtain diagnostic tests; 10 (6%) to discontinue therapy; 6 (4%) to de-escalate therapy; 5 (3%) to adjust doses; and 2 (1%) to broaden therapy. Only 8 of the recommended changes (5%) were declined by the clinicians. The average monthly DOT for all types of antibiotics declined significantly from 2,213 to 1,681 (24% decrease; P = .04). Reductions in DOT for the most common broad-spectrum antibiotics included colistin from 303 to 276 (P = .75); imipenem-cilastatin 434 to 248 (P = .06); doripenem 150 to 144 (P = .85). Piperacillin-tazobactam increased from 122 to 142 (P = 0.75). Conclusions: We demonstrated that daily review of cultures and antibiotic use decreased overall antibiotic prescribing. Given that few recommendations included discontinuation of therapy, ASP rounds likely raised awareness for clinicians to optimize antibiotic use.

Funding: None

Disclosures: None

Background: Inappropriate prescribing behavior can be associated with higher rates of antibiotic resistance, calling for detailed studies on how physicians make prescribing decisions. We conducted a mixed-methods study to investigate physician antibiotic prescribing behavior in a 141-bed pediatric hospital. Methods: We applied a mixed-methods research design. The quantitative phase was conducted over a 6-month period to identify cases of inappropriate prescribing. The qualitative phase comprised 22 qualitative interviews with clinical teaching units (CTU) and pediatric intensive care unit (PICU) team members (physicians and pharmacists). Two coders analyzed the data deductively using the theoretical domain framework (TDF), as well as the social determinants of antimicrobial prescribing (SDAP). Results: In 52.9% of the 36 identified cases in the CTU and 31.4% of the 37 cases in the PICU, an infectious diseases (ID) consultation occurred. Compliance rates with ID recommendations were 79% and 91% in the CTU and PICU, respectively. The CTU and PICU expressed appreciation for ID involvement when ID supported their de-escalation choices in complex cases and in cases in which less commonly known antibiotics were used. However, the ID service involvement was perceived as detrimental to antimicrobial prescribing decisions for CTU and PICU across 3 of the 4 SDAP domains (Fig. 1, qualitative research quotes). Relationship between clinicians: CTU physicians and pharmacists perceived ID involvement as negatively impacting the relationship of the team. Antimicrobial decisions were automatically defaulted to ID, whereas pharmacy involvement was disregarded and the decisions were delayed. Risk, fear, and emotion: These were experienced across all respondents’ groups that identified ID specialists’ egos and personalities as contrary to open collaborative discussion on antimicrobial decisions. (Mis)perception of the problem: ID physicians were identified as more conservative in their antimicrobial choices, leading to prolonged duration of treatment, broader choices, and longer hospitalizations. The CTU and pharmacy respondents felt that ID recommendations were inconsistent among physicians and deviated from guidelines with little justification. Conclusions: Although CTU and PICU teams tend to comply with ID prescribing recommendations and ID involvement with complicated cases, pharmacists, CTU physicians, and PICU physicians perceived ID consultations to negatively affect collaborative efforts for stewardship. These findings offer novel insights into how an ID service can improve its role to positively affect appropriate prescribing. CTU and PICU respondents called for a supportive and trusting relationship with the ID service as a major driver for behavioral change and enhanced stewardship.

Funding: None

Disclosures: None

Background: An important aspect of antimicrobial stewardship is the qualitative assessment of antimicrobial prescribing. Owing to lack of standardized tools and resources required to design, conduct and analyze qualitative audits, these assessments are rarely performed. Objective: We designed an audit tool that was appropriate for all Australian hospital types, suited to local user requirements and including an assessment of the appropriateness of antimicrobial prescribing. Methods: In 2011, a pilot survey was conducted in 32 Australian hospitals to assess the usability and generalizability of a qualitative audit tool. The tool was revised to reflect the respondents’ feedback. A second study was performed in 2012 in 85 hospitals. In 2013, following further feedback and refinement, an online auditing tool, the Hospital National Antimicrobial Prescribing Survey (NAPS), was developed. Early audits demonstrated that surgical prophylaxis had the highest rates of inappropriate prescribing. In 2016, the Surgical NAPS was developed to further investigate reasons for this, and the NAPS program was further expanded to audit antimicrobial prescribing practices in Australian aged-care homes (ie, the Aged Care NAPS). Results: Between January 1, 2013, and November 12, 2019, 523 Australian public and private hospitals (53.8%) utilized the Hospital NAPS; 215 (22.1%) have utilized the Surgical NAPS; and 774 of Australian aged-care homes (29.0%) have utilized the Aged Care NAPS. National reporting has identified key target areas for quality improvement initiatives at both local and national levels. The following initiatives have been outlined in 14 public reports: improved documentation; prolonged antimicrobial prophylaxis; compliance with prescribing guidelines; appropriateness of prescribing; access to evidence-based guidelines; and improved microbiology sampling. Conclusions: By utilizing the Plan-Do-Study-Act cycle for healthcare improvement and by involving end users in the design and evaluation, we have created a practical and relevant auditing program to assess both quantitative and qualitative aspects of antimicrobial prescribing in a wide range of settings. This voluntary program is now endorsed by the National Strategy for Antimicrobial Resistance Surveillance, partners with the Antimicrobial Use and Resistance in Australian Surveillance System, and is utilized by facilities to meet mandatory national accreditation standard requirements. With the success of the NAPS program in Australia, it has now been implemented in New Zealand, Canada, Malaysia, Fiji, and Bhutan, with plans for other countries to implement the program soon. Current research is being conducted to expand the program to include audits for family physicians, veterinarians, and remote indigenous communities, and for antifungal use.

Disclosures: None

Funding: None

Background: The epidemic NAP1/027 Clostridioides difficile strain (MLST1, ST1) that emerged in the mid-2000 is on the decline. The current distribution of C. difficile strain types and their transmission dynamics are poorly defined. We performed whole-genome sequencing (WGS) of C. difficile isolates in 2 regions to identify the predominant multilocus sequence types (MLSTs) in community- and healthcare-associated cases and potential transmission between cases using whole-genome single-nucleotide polymorphism (SNP) analysis. Methods: Isolates were collected through the CDC Emerging Infections Program population-based surveillance for C. difficile infections (CDI) for 3 months between 2016 and 2017 in 5 Minnesota counties and 1 New York county. Isolates were limited to incident cases (CDI in a county resident with no positive C. difficile test in the preceding 8 weeks). Cases were classified as healthcare associated (HA-CDI) or community associated (CA-CDI) based on healthcare exposures as previously described. WGS was performed on an Illumina Miseq. The CFSAN (FDA) pipeline was used to compute whole-genome SNPs, SPAdes was used for assembly, and MLST was assigned according to www.pubmlst.org. Results: Of 431 isolates, 269 originated from New York and 162 from Minnesota; 203 cases were classified as CA-CDI and 221 as HA-CDI. The proportion of CA-CDI cases was higher in Minnesota than in New York: 62% vs 38%. The predominant MLSTs across both sites were ST42 (9%), ST8 (8%), and ST2 (8%). MLSTs more frequently encountered in HA-CDI than CA-CDI included ST1 (note that this ST includes PCR Ribotype 027; 76% HA-CDI), ST53 (84% HA-CDI), and ST43 (80% HA-CDI). In contrast, ST110 (63% CA-CDI) and ST3 (67% CA-CDI) were more commonly isolated from CA-CDI cases. ST1 accounted for 7.6% of circulating strains and was more common in New York than Minnesota (10% vs 3%) and was concentrated among New York HA-CDI cases. Also, 412 isolates (1 per patient) were included in the final whole-genome SNP analysis. Of these, only 12 pairs were separated by 0–3 SNPs, indicating potential transmission, and most involved HA-CDI cases. ST1, ST17, and ST46 accounted for 8 of 12 pairs, with ST17 and ST46 potentially forming small clusters. Conclusions: This analysis provides a snapshot of the current genomic epidemiology of C. difficile across 2 geographically and epidemiologically distinct regions of the United States and supports other studies suggesting that the role of direct transmission in the spread of CDI may be limited.

Funding: None

Disclosures: None