Background: Modeling is a cost-effective way to evaluate interventions pertaining to hospital infection acquisitions, such as staffing levels. Increasing the number of nurses in an intensive care unit affects rates of HAI transmission. The way compartmental models are often formulated assumes that there is a never-ending series of tasks for workers to perform. Our previous models used a baseline of 1:3 nurse:patient ratio, and we kept the number of tasks the same across staffing ratios. We wanted to understand how having a finite number of tasks, using this baseline number, across staffing levels affected HAI acquisitions. Methods: We used a stochastic mathematical model of methicillin-resistant Staphylococcus aureus (MRSA) to study the impact of changes in staffing and a finite pool of tasks on hospital-associated acquisitions. For a 15-bed intensive care unit (ICU), we have 1 intensivist, and we set the nurse:patient ratios at 1:1, 1:1.5, 1:2.5, 1:3, 1:5, and 1:7.5, to represent the extreme ends of staffing levels and more moderate values in line with critical care society guidelines. Each model was run 1,000 times. The outcome of each scenario is the median number of hospital-associated MRSA acquisitions in 1 year from those 1,000 runs. Results: Treating the 1:3 nurse:patient ratio as the baseline, with 45 MRSA acquisitions per year, increasing the number of nurses from 5 to 6 (moving to a 1:2.5 nurse:patient ratio) had a relative risk (RR) of 0.77, suggesting that a small change in nurse staffing levels might have an outsized impact on rates. More dramatic changes had correspondingly larger swings in MRSA acquisition rates, with 1:1 nurse:patient ratio scenarios having an RR of 0.17, and at the other extreme, a 1:7.5 nurse:patient ratio having an RR of 4.66. Comparing the infinite to finite models, the ratios with more nurses had lower acquisition rates, with decreases ranging from 20% to 50%. Ratios with fewer nurses in the ICU showed 100%–400% increases in the number of acquisitions. All results were statistically significant. Conclusions: As nurse:patient ratios go up, the burden of direct-care tasks fall on fewer people, which has a direct impact on HAI rates. Our model demonstrates this hypothesis. Therefore, appropriate staffing should be considered in infection control guidelines, and the cost of staffing should be weighed against its impact on infection prevention as well as other areas of patient care. In this study, we considered only the impact from changes in contact patterns emerging from different staffing levels. Further insights may exist when considering other outcomes that also accompany increased staffing.

Disclosures: None

This quality improvement project aimed to reduce institutional incidence of Clostridioides difficile infection (CDI) following autologous stem cell transplantation. CDI incidence per transplant was .17 in a baseline period and .09 following the implementation of postdischarge ultraviolet room cleaning (χ2 = 2.11, p = .15).

Background: Prevention of healthcare-associated infections (HAIs) requires timely feedback to and input from all staff involved in patient care to best identify practice gaps and improvement targets. However, multidisciplinary review of HAI events can be challenging to promptly complete given staffing shortages and the excess administrative burden of emailed and printed forms and disjointed analyses, reporting, and visualization tools. Plagued by a lack of feedback from attending and ordering physicians, difficulty transcribing and analyzing nonstandardized data, and challenges in summarizing and distributing actionable findings, we sought to reduce turnaround time (TAT), improve data collection, and broaden communication of HAI contributing factors and proposed solutions. Methods: A secure web application for electronic data capture and reporting, Research Electronic Data Capture (REDCap), was used; the software application is free to nonprofit organizations. The review process is now initiated by an infection preventionist entering HAI information into an initial survey, which automatically cascades information into 4 subsequent surveys, distributed through automated email links, providing an opportunity for individual responses from the nursing unit, the attending provider, an infectious disease physician, and the ordering provider for the positive test that detected the HAI. Survey questions focus on evaluation of adherence to CDC and SHEA HAI prevention strategies. Reminders are automatically generated and continue to be sent to involved staff until their portion is completed. Survey responses are automatically summarized upon completion of all reviews and are shared with several stakeholders, including hospital leadership, the care team, infection prevention staff, and quality-control partners (Fig.). Discrete qualitative and quantitative data are exported in a standard application-programming interface (API) format for immediate analysis and interpretation. Results: After the review process was launched using new electronic technology, the average TAT and completion rate improved from 23 days and 40% to 7 days and 95%, respectively. Input from ordering and attending physicians, once extremely rare, became frequent. Nuanced insight into causative and preventive factors, previously unachievable, occurred during review of all 38 HAIs reported in December 2022. Reviewers believed that 48% of HAIs reviewed could have been prevented. Conclusions: Applying electronic technology to HAI case review improved completion and timeliness of reviews by both providers and nurses. By sharing data and insights with all stakeholders in real time, the new procedure permitted multidirectional communication between the care teams and increased awareness of patient harm as well as ownership of patient safety. Our process is freely and readily generalizable to any nonprofit healthcare facility.

Disclosures: None

Background: Parkland Health is a 900-bed safety-net hospital that serves Dallas County, Texas. It has an OPAT program in which patients are managed via self-administration (S-OPAT), home-health/hemodialysis (H-OPAT), and skilled nursing facilities (SNF-OPAT). We evaluated the reasons for unscheduled emergency department (ED) visits by patients in these groups to identify strategies to decrease unexpected healthcare utilization and to improve safety. Methods: We performed a retrospective chart review of all adult patients discharged from Parkland Health on OPAT between April and June 2021. Demographic, medical, and healthcare utilization information, including the date and reason of first unscheduled ED visit after discharge, was collected utilizing a standardized instrument. The institutional review board approved this study. Results: In total, 184 patients were discharged with OPAT. Among them, 32% were female and 55% identified as Hispanic; 41% were non-English speakers, and 45% were treated for a musculoskeletal infection. Among all OPAT models of care, 43.4% were S-OPAT patients, 31.5% were H-OPAT patients, and 25% were SNF-OPAT patients (Table 1). The groups differed, and fewer African Americans received H-OPAT. Also, 45% were being treated for musculoskeletal infections and were more likely to be discharged with H- or SNF-OPAT. In addition, 41% were being treated for endovascular infections and 21.7% were being treated for genitourinary infections. The total length of stay in the hospital was longer for SNF-OPAT patients and shorter for S-OPAT patients (Table 2). Among 184 OPAT patients, 41 patients (22.2%) had an ED visit: 17.3% SNF-OPAT patients, 27.6% H-OPAT patients, and 21.3% S-OPAT patients (Table 2). ED visits were attributed to intravenous (IV) access–related problems (12 of 41, 29.0%), worsening of known infection (3 of 41, 7.3%), and abnormal blood test results (2 of 41, 4.9%). Also, 24 ED visits (58%) were not related to underlying infection or OPAT. However, when examined by the OPAT care model, 41% of ED visits among S-OPAT patients, 20% among H-OPAT visits, and 25% among SNF-OPAT visits were related to IV access issues. Among S-OPAT ED visits pertaining to IV access, 71% were for minor issues such as dressing changes or line occlusion or malfunction. Conclusions: One-fifth of OPAT patients had an unscheduled ED visit, of whom 20%–41% had issues with IV access. Many of these visits could be avoided with enhanced outreach to patients discharged with OPAT and improved ambulatory capabilities to provide standard services related to maintenance of IV access.

Disclosures: None

Background: Prevention of hospital-onset bacteremia (HOB) in all settings is a healthcare priority. The CDC is developing a neonatal-specific HOB quality metric, but the epidemiology of neonatal HOB is poorly understood. Our objective was to validate a prior single-center finding that HOB risk varies by birthweight and postnatal age in a multicenter cohort. Methods: We performed a multicenter, retrospective cohort study of neonates admitted to 4 neonatal intensive care units (NICUs) for ≥4 days between July 1, 2016, and July 1, 2021. HOB was defined as a positive blood culture for bacteria or fungi on day ≥4 of admission. The first HOB event in the hospitalization was counted per neonate. Repeat HOB events during a neonate’s admission were excluded. Poisson regression models with robust variance estimates were used to estimate the incidence rate (IR) of HOB, expressed as HOB events per 1,000 patient days and IR ratios (IRRs), within strata defined by CDC birthweight categories and 4-week postnatal age intervals, adjusting for central venous catheter (CVC) presence at time of HOB and study site. Results: The analysis included 9,267 neonates, contributing 191,295 patient days and 470 HOB events, with an unadjusted IR of 2.46 per 1,000 patient days (Table 1). Of 477 infants born ≤750 g, 153 (30.1%) had a HOB with an IR of 13.3 (95% CI, 10.5–16.0) events per 1,000 patient days in the first 4 weeks after birth (Fig. 1). After adjusting for CVC presence and study site, infants ≤750 g had a higher HOB rate in the first 4 weeks of life (IRR, 7.45; 95% CI, 3.81–14.56) compared to infants ≥2,500 g. After 8 weeks of life, there was no difference in HOB rate in the 2 groups (IRR, 0.8, 95% CI, 0.3–2.7). Conclusions: Neonates born ≤750 g were at highest risk for HOB within the first 4 weeks after birth; however, risk for HOB was not consistent over time. Postnatal age should be considered in a neonatal HOB quality metric.

Disclosures: None

Background: The University of Colorado Health (UCHealth) metropolitan region is composed of 4 hospitals. Therein, 10 infection preventionists (IPs) retrospectively review all cases of potential central-line–associated bloodstream infection (CLABSI), catheter-associated urinary tract infection (CAUTI), and surgical site infection (SSI) to adjudicate whether each case meets the NHSN definitions for hospital-acquired infection (HAI). In August 2021, the UCHealth IP team structure transitioned from a subject-matter expert model (in which each IP reviewed a specific HAI) to a unit-based model (in which each IP reviewed all HAIs and SSIs on their assigned units) to create redundancy in knowledge and skill. The IP team subsequently instituted a weekly meeting to review all potential cases of HAI. We hypothesized that this review structure would result in increased consistency in the application of NHSN definitions across the UCHealth hospitals and units. Methods: From August 17, 2022, through March 3, 2023, the UCHealth IPs, managers, and medical directors met weekly for 1 hour via teleconferencing. Each IP presented key details for all near-miss and confirmed cases of SSI or HAI on their respective units and received questions and feedback from their peers and medical directors. Case determination was based on team discussion and consensus. If there was discordance in the interpretation of an NHSN case definition, a formal inquiry was sent to resolve the uncertainty. The number of cases reviewed, case determinations changed, and formal inquiries to NHSN were tracked. Results: During the study period, the IP team convened weekly meetings and reviewed 248 patient cases—of which 208 (83.9%) were confirmed HAIs. Based on collaborative team discussion, 14 cases (5.6%) were changed from reportable to nonreportable. Three cases (1.2%) originally thought to be nonreportable were changed to reportable. The HAI determination of a reportable case (eg, revision of a “superficial” SSI to “deep” SSI) was changed for 9 (6.0%). Following team discussion, 13 formal inquiries were sent to the NHSN to clarify case definitions, and these responses were collated for future reference. Conclusions: Team-based IP review of HAI cases improves consistency in application of NHSN case definitions and highlights areas of uncertainty in their interpretation. This team-based model of case review is a useful educational and practical tool to increase interrater reliability in case adjudication across large teams of IPs, to create a systematic way to query NHSN, and to ensure that knowledge gained is disseminated for future benefit.

Disclosures: None

Effective de-implementation models often include replacement of an ineffective practice with an alternative. We co-developed patient education materials as a replacement strategy for inappropriate post-procedural antibiotics in cardiac device procedures. Lessons learned and developed materials may be used to promote infection prevention in other periprocedural settings.

Background: We estimated the incidence of hospital-acquired respiratory virus infections (HARVIs) by viral species, and we identified risk factors for and outcomes of HARVIs. Methods: We identified a cohort of all inpatient admissions of ≥24 hours duration to University of Michigan hospitals during 3 study years (2017–2018, 2018–2019, and 2019–2020). HARVIs were defined as initial respiratory virus detection (adenovirus, coronaviruses, human metapneumovirus, influenza A and B, parainfluenza viruses, respiratory syncytial virus, or rhinovirus-enterovirus) in a clinical test ordered after the 95th percentile of the virus-specific incubation period. Incidence was calculated as the number of HARVIs per 10,000 patient admission days. Patient demographic and clinical characteristics were assessed as risk factors for HARVI in Cox proportional hazards models of the competing outcomes of HARVIs and hospital discharge. The association between time-varying HARVI status and the competing outcomes of discharge and in-hospital death was estimated in covariate-adjusted Cox-proportional hazards models. All analyses were performed separately for adult patients (aged ≥18 years) and pediatric patients (aged <18 years). Results: The overall incidences of HARVI were 8.5 and 3.0 per 10,000 admission days for pediatric and adult patients, respectively. Rhinovirus was the most common HARVI in both pediatric and adult patients, with incidences of 5.1 and 1.1 infections per 10,000 admission days, respectively. With the exception of influenza A, the incidence of HARVI was higher in pediatric patients than adult patients for all viral species. For adults, congestive heart failure, renal disease, and cancer all increased HARVI risk independent of their associations with extended hospital stays. Risk of HARVI was also elevated for patients admitted September through June relative to July admissions. For pediatric patients, chronic cardiovascular and respiratory conditions, cancer, medical-device dependence, and December admission increased risk of HARVI. Age, sex, and race were not associated with risk of HARVI for children or adults. Inpatient lengths of stay were longer for adults with HARVI compared to those without (range of virus-specific hazard ratios, 0.48– 0.77). However, estimated effects were not statistically significant for human metapneuomovirus, parainfluenza, or adenovirus. Only influenza A was associated with an increased risk of in-hospital death within 30 days of infection for adults. No HARVIs were associated with increased length of stay or risk of death for pediatric patients. Conclusions: The incidence of HARVI varied by viral species and was higher among pediatric patients. HARVIs increased the length of hospital stays for adults but not for pediatric patients.

Disclosures: None

Background: Globally, the 30-day cumulative incidence of surgical-site infections (SSI) was 11% (95% CI, 10%–13%) based on the systematic review and meta-analysis derived from 57 studies. SSIs are poorly studied in the Philippines. Americares and its hospital partner, Camarines Norte Provincial Hospital, Philippines, sought to reduce SSIs through (1) establishing SSI surveillance in the hospitals’ surgical departments, (2) implementing quality improvement processes, and (3) developing and implementing an SSI prevention care bundle. Methods: A quality improvement methodology was used to introduce SSI surveillance and care-bundle checklist in partnership with Americares. Using paired t tests, pre- and posttest scores of the SSI care bundle training were analyzed. SSI surveillance was established based on the adapted CDC criteria. All clean surgeries were monitored except orthopedic surgeries. The number of surgeries performed, monitored, and SSIs identified were documented using the surveillance forms and plotted using Microsoft Excel software. A care bundle based on WHO evidence-based interventions for SSI prevention was designed and implemented. Compliance with the SSI care bundle was documented using Microsoft Excel. The relationship between the use of a care bundle and SSIs was analyzed using the Pearson correlation coefficient. Results: An online SSI care bundle training session was conducted. Overall, 150 participants had a mean pretraining test score of +6.46. After the training was conducted, the same participants had a mean posttraining test score of + 1.76). a statistically significant increase of 5.29 (95% CI). Thereby, the mean score difference after training showed that knowledge increased overall. These findings show an average of 90.43% compliance with the SSI care-bundle checklist over the 18-month window from May 2021 to November 2022. From a baseline of 0%, compliance increased from 80% upon its introduction in May 2021. Lastly, the SSI incidence rate from May 2021to November 2022 averaged 1.89%. The days between reported SSIs averaged 16.85. No baseline was available for comparison prior to the introduction of the surveillance and care bundle. A Pearson r data analysis (n = 1,850) was used to determine the relationship between the use of the care bundle and SSIs. The data illustrated a moderate negative correlation (r = −.31). Therefore, higher care-bundle compliance yielded fewer SSI cases. Conclusions: The use of an evidence-based care bundle paired with a local quality improvement process significantly improved SSI prevention and surveillance. Future studies are needed that include clean-contaminated, contaminated, and dirty surgical cases to test the degree of SSI reduction possible.

Disclosures: None

Background: Cefazolin is the preferred antimicrobial for prevention of surgical-site infections in most procedures at our institution. Our first alternative is vancomycin which is associated with higher adverse events and infection rates. The presence of penicillin allergies can influence prescribing of vancomycin despite a low risk of cross-reactivity between penicillin and cephalosporins. Nebraska Medicine implemented a systemwide change in April 2022 that suppressed alerts for non–IgE-mediated penicillin allergies in the electronic medical record (EMR, Epic Systems) upon cephalosporin prescribing. We evaluated changes in perioperative antimicrobial surgical infection prophylaxis after this change. Methods: We conducted a quasi-experimental study of all patients undergoing procedures for which cefazolin is considered preferred per institutional guidance. Preintervention data were from April 1, 2021, to March 31, 2022, and postintervention data included patients from April 11, 2022, to October 31, 2022, after guidance was distributed to surgeons, operating room staff, and pharmacists. Patients were excluded if they were aged <19 years, had a hospital length of stay <24 hours, underwent procedures after their first throughout the time frame, or received both vancomycin and cefazolin. Statistical significance was set at P < .05, determined using the Fisher exact test. Results: The study included 6,676 patients: 4,147 in the preintervention group and 2,529 in the postintervention group. We identified 15 procedure categories, with no significant differences between periods (Table 1). The average age was 61 years. Penicillin allergy was reported in 508 patients (12.3%) in the preintervention group and in 319 patients (12.6%) in the postintervention group. In individuals with penicillin allergy, cefazolin prescribing increased from 49.6% to 74.3% (P < .01) and vancomycin prescribing decreased from 50.4% to 25.7% (P < .01). The largest changes occurred in patients undergoing cardiac, spinal, neurological, and vascular procedures. For patients without penicillin allergy, prescribing remained unchanged. Overall, cefazolin prescribing increased from 92.0% to 95.0% (P < .01), and the rate of vancomycin prescribing decreased from 8.0% to 5.0% (P < .01) in procedures for which cefazolin was preferred. Conclusions: Following the suppression of EMR alerts for non–IgE-mediated allergies when ordering cephalosporins, penicillin prescribing rates of cefazolin for surgical infection prophylaxis improved significantly in procedures for which it was the preferred agent. Further research on infection rates and adverse events with these and other alternative agents are needed.

Disclosures: None

Background: Carbapenem-resistant Acinetobacter (CRA) bacteria are an urgent public health threat. Accurate and timely testing of CRA is important for proper infection control practices to minimize spread. In 2017, the CDC estimated 8,500 CRA cases among hospitalized patients, 700 deaths, and $281 million in attributable healthcare costs. Treatment options are extremely limited for carbapenem-resistant Acinetobacter baumannii (CRAB) infections, making CRAB a unique concern. Colonization screening is a valuable tool for containment but requires sampling of 4 body sites. Identifying a reliable specimen collection site for CRAB is important to inform public health recommendations as screening can cost healthcare facilities valuable time and resources. Methods: Results of all screening specimens of patients with at least 1 site positive for CRAB on a unique collection date were extracted from the Southeast Regional data of Antimicrobial Resistance Lab Network (SEARLN) data. Non-CRAB screening and screenings that did not yield at least 1 positive result on a single collection date were excluded. We also limited our data to include only the following screening sites, which have been validated by the Tennessee Department of Health’s State Public Health Laboratory: axilla and groin, rectal, sputum, and wound. For each specimen source, we calculated the percentage of positive specimen among CRAB-colonized patients. Data were extracted and analyzed using SAS version 9.4 software. Results: The SEARLN data contained 594 CRAB screening specimens collected over 4 years, 2018 through 2021, and 486 of those specimens yielded CRAB. For CRAB-colonized patients screened in this study, wound specimens had the highest positivity rate at 93.4% (95% CI, 89.9%–96.9%) of samples culturing CRAB. Sputum followed at 87.7%, then axilla and groin at 77.6% and rectal at 59.7%. Conclusions: Wound specimens produced the highest proportion of positive cultures among CRAB-positive patients, making them the sample type with the highest prevalence in our study. For healthcare facilities with limited time and resources seeking to optimize their CRAB screening process, wound specimens may be the most reliable single site for detecting CRAB colonization in patients with an open wound. When a wound is not present, sputum may be a good alternative single-source collection site. More research should be conducted before CRAB screening recommendations are updated.

Disclosures: None

Background: Candida auris, an emerging fungal pathogen, is frequently drug resistant and spreads rapidly in healthcare facilities. Screening to identify patients colonized with C. auris can prevent further spread by prompting aggressive infection prevention and control measures. The CDC recommends C. auris screening based on local epidemiological conditions, patient characteristics, and facility-level risk factors; such screening might help facilities in higher burden areas to mitigate transmission and those in lower-burden areas to detect new introductions before spread begins. To describe US screening practices and challenges, we surveyed a network of infection disease practitioners, comparing responses by local C. auris case burdens. Methods: In August 2022, we emailed a survey about C. auris screening practices to ~3,000 members of the IDSA Emerging Infection Network. We describe survey results, stratifying findings by whether the healthcare facility was in a region where C. auris is frequently identified (tier 3 facility) or not frequently identified (tier 2 facility), based on CDC assessment using existing multidrug-resistant organism containment guidance (https://www.cdc.gov/hai/containment/guidelines.html). Results: We received 253 responses (tier 3 facilities: 119, tier 2 facilities: 134); overall, 37% performed screening. Tier 3 facilities more frequently performed screening than tier 2 facilities (59% vs 17%). Among facilities that performed screening, tier 3 facilities, compared with tier 2 facilities, more frequently screened patients on admission (84% vs 55%) and used an in-house laboratory for testing (68% vs 29%), most often with culture-based methods. Tier 2 facilities more frequently screened patients already admitted in the facility (eg, in response to cases or as part of point-prevalence surveys) compared with tier 3 facilities (59% vs 49%). Among facilities performing screening, 72% had identified ≥1 case in the previous year (tier 3 facilities, 85%; tier 2 facilities, 33%). Barriers to screening included limited laboratory capacity, long testing turnaround times, and the perception that screening was not useful. Conclusions: Most facilities surveyed did not perform C. auris screening. However, most facilities that performed screening, including those in regions of higher and lower C. auris burden, detected cases during the previous year. Admission screening, which might help detect new introductions before spread begins, was uncommon in facilities in lower-burden areas. Improving ease of C. auris screening through access to in-house laboratory testing with rapid turnaround times might increase the adoption of C. auris screening by facilities, thereby increasing detection and preventing spread.

Disclosures: None