Experiments in engineering are typically conducted in controlled environments where parameters can be set to any desired value. This assumes that the same applies in a real-world setting, which is often incorrect as many experiments are influenced by uncontrollable environmental conditions such as temperature, humidity, and wind speed. When optimizing such experiments, the focus should be on finding optimal values conditionally on these uncontrollable variables. This article extends Bayesian optimization to the optimization of systems in changing environments that include controllable and uncontrollable parameters. The extension fits a global surrogate model over all controllable and environmental variables but optimizes only the controllable parameters conditional on measurements of the uncontrollable variables. The method is validated on two synthetic test functions, and the effects of the noise level, the number of environmental parameters, the parameter fluctuation, the variability of the uncontrollable parameters, and the effective domain size are investigated. ENVBO, the proposed algorithm from this investigation, is applied to a wind farm simulator with eight controllable and one environmental parameter. ENVBO finds solutions for the entire domain of the environmental variable that outperform results from optimization algorithms that only focus on a fixed environmental value in all but one case while using a fraction of their evaluation budget. This makes the proposed approach very sample-efficient and cost-effective. An off-the-shelf open-source version of ENVBO is available via the NUBO Python package.

We present a deep learning approach for near real-time detection of Global Navigation Satellite System (GNSS) radio frequency interference (RFI) based on a large amount of aircraft data collected onboard from the Global Positioning System (GPS) and Attitude and Heading Reference System (AHRS). Our approach enables detection of GNSS RFI in the absence of total GPS failure, i.e. while the receiver is still able to estimate a position, which means RFI sources with low power or at larger distance can be detected. We demonstrate how deep one-class classification can be used to detect GNSS RFI. Furthermore, thanks to a unique dataset from the Swiss Air Force and Swiss Air-Rescue (Rega), preprocessed by Swiss Air Navigation Services Ltd. (Skyguide), we demonstrate application of deep learning for GNSS RFI detection on real-world large scale aircraft data containing flight recordings impacted by real jamming. The approach we present is highly general and can be used as a foundation for solving various automated decision-making problems based on different types of Communications, Navigation and Surveillance (CNS) and Air Traffic Management (ATM) streaming data. The experimental results indicate that our system successfully detects GNSS RFI with 83$\,\cdot\,$5% accuracy. Extensive empirical studies demonstrate that the proposed method outperforms strong machine learning and rule-based baselines.

Background: New Delhi Metallo-β-lactamase (NDM)–producing Escherichia coli are highly resistant organisms that spread quickly. In the United States, organisms with blaNDM are rare and mostly associated with healthcare settings. However, in other countries, blaNDM can be relatively common and are found in community settings. State veterinary and public health partners detected NDM E. coli in a dog from Iran living at a Wisconsin animal rescue facility (ARF), where 40% of dogs had international origins. We investigated to determine spread among dog and human contacts and prevent further transmission. Methods: We screened dogs and humans at the ARF, a local veterinary clinic (clinic A), and ARF staff homes (homes A and B) for colonization with blaNDM. We reviewed veterinary records and conducted a case–control analysis to identify risk factors for blaNDM acquisition among dogs. We evaluated ARF infection control practices. Screening specimens that were positive for blaNDM were cultured. We conducted an analysis of short- and long-read whole-genome sequencing data to evaluate isolate relatedness. We compared NDM E. coli sequences from dogs to all NDM E. coli sequences from humans collected in Wisconsin and nearby states. Results: Screening identified blaNDM colonization in 27 (37%) of 73 ARF dogs and 4 (56%) of 7 dogs in home A, but not in ARF or staff in clinic A. Among ARF dogs with blaNDM, 20 (74%) 27 had international origins and 22 (81%) had ≥1 medical condition. Dogs sharing the same space (OR, 5.1; 95% CI, 1.8–14.7) were associated with blaNDM acquisition. We observed high animal density, soiled environments, and insufficient hand hygiene. ARF staff wore workwear and work shoes off site, including to home A. Sequencing identified 3 multilocus sequence types (STs) using the Achtman scheme among 27 isolates with blaNDM-5. Most isolates were ST361 (20 of 27, 74%) followed by ST167 (6 of 27, 22%) and ST1163 (1 of 27, 4%). Within-MLST cluster variability was <1–3 high-quality single-nucleotide variant differences, each harboring a ST-specific plasmid with blaNDM-5. No NDM-E. coli sequences from humans appeared related. Conclusions: Investigation of a single isolate led to identification of widespread NDM-E. coli transmission among dogs at an ARF. There were multiple NDM E. coli introductions to the ARF, likely by dogs of international origin. Poor hygiene contributed to transmission among ARF dogs and to dogs outside the ARF. Transmission of blaNDM-5 at the ARF and offsite spread to home A demonstrate the potential for unrecognized community sources to disseminate NDM E. coli in community settings. Strategies and lessons learned from interventions to prevent antibiotic resistance in human healthcare settings may inform and support prevention in animal care.

Disclosures: None

Among nursing home outbreaks of coronavirus disease 2019 (COVID-19) with ≥3 breakthrough infections when the predominant severe acute respiratory coronavirus virus 2 (SARS-CoV-2) variant circulating was the SARS-CoV-2 δ (delta) variant, fully vaccinated residents were 28% less likely to be infected than were unvaccinated residents. Once infected, they had approximately half the risk for all-cause hospitalization and all-cause death compared with unvaccinated infected residents.

Therapeutic drug monitoring of flecainide in children using plasma concentration measurements is undertaken by some clinicians. There is very little published evidence surrounding factors which influence plasma flecainide concentration, particularly in paediatric populations. We undertook a retrospective study of 45 children receiving flecainide to identify factors that influence its plasma concentration. Patients receiving a dose of 6 mg/kg/day had a higher mean plasma flecainide concentration than those receiving lower doses. Younger age and lighter weight were also associated with higher plasma flecainide concentrations. Children aged younger than 1 year receiving flecainide three times a day had a higher mean plasma flecainide concentration than older children who received flecainide twice a day. All supratherapeutic levels occurred in children aged less than 1 year who were receiving flecainide three times a day. Supratherapeutic levels were more common in those receiving 6 mg/kg/day while subtherapeutic levels were more common in those receiving 2 mg/kg/day. A supratherapeutic level did not correlate with adverse effects or clinical toxicity. Our results would suggest the need for a change of practice from prescribing flecainide at a frequency of three times a day in children aged younger than 1 year to twice a day in line with other ages.

In this large, retrospective cohort study, we used administrative data to evaluate nonpregnant adults with group B Streptococcus (GBS) bacteriuria. We found greater all-cause mortality in those with urinary tract infections compared to asymptomatic bacteriuria. Differences in patients’ baseline characteristics and the 1-year mortality rate raise the possibility that provider practices contribute to differences observed.

OBJECTIVES/GOALS: Using the covariate-rich Veteran Health Administration data, estimate the association between Proton Pump Inhibitor (PPI) use and severe COVID-19, rigorously adjusting for confounding using propensity score (PS)-weighting. METHODS/STUDY POPULATION: We assembled a national retrospective cohort of United States veterans who tested positive for SARS-CoV-2, with information on 33 covariates including comorbidity diagnoses, lab values, and medications. Current outpatient PPI use was compared to non-use (two or more fills and pills on hand at admission vs no PPI prescription fill in prior year). The primary composite outcome was mechanical ventilation use or death within 60 days; the secondary composite outcome included ICU admission. PS-weighting mimicked a 1:1 matching cohort, allowing inclusion of all patients while achieving good covariate balance. The weighted cohort was analyzed using logistic regression. RESULTS/ANTICIPATED RESULTS: Our analytic cohort included 97,674 veterans with SARS-CoV-2 testing, of whom 14,958 (15.3%) tested positive (6,262 [41.9%] current PPI-users, 8,696 [58.1%] non-users). After weighting, all covariates were well-balanced with standardized mean differences less than a threshold of 0.1. Prior to PS-weighting (no covariate adjustment), we observed higher odds of the primary (9.3% vs 7.5%; OR 1.27, 95% CI 1.13-1.43) and secondary (25.8% vs 21.4%; OR 1.27, 95% CI 1.18-1.37) outcomes among PPI users vs non-users. After PS-weighting, PPI use vs non-use was not associated with the primary (8.2% vs 8.0%; OR 1.03, 95% CI 0.91-1.16) or secondary (23.4% vs 22.9%;OR 1.03, 95% CI 0.95-1.12) outcomes. DISCUSSION/SIGNIFICANCE: The associations between PPI use and severe COVID-19 outcomes that have been previously reported may be due to limitations in the covariates available for adjustment. With respect to COVID-19, our robust PS-weighted analysis provides patients and providers with further evidence for PPI safety.

The remarkable archaeological record of Neolithic Orkney has ensured that these islands play a prominent role in narratives of European late prehistory, yet knowledge of the subsequent Bronze Age is comparatively poor. The Bronze Age settlement and cemetery at the Links of Noltland, on the island of Westray, offers new evidence, including aDNA, that points to a substantial population replacement between the Late Neolithic and Bronze Age. Focusing on funerary practice, the authors argue for interconnecting identities centred on household and community, patrilocality and inheritance. The findings prompt a reconsideration of the Orcadian Bronze Age, with wider implications for population movement and the uptake of cultural innovations more widely across prehistoric north-western Europe.

Social media platforms allow users to share news, ideas, thoughts, and opinions on a global scale. Data processing methods allow researchers to automate the collection and interpretation of social media posts for efficient and valuable disease surveillance. Data derived from social media and internet search trends have been used successfully for monitoring and forecasting disease outbreaks such as Zika, Dengue, MERS, and Ebola viruses. More recently, data derived from social media have been used to monitor and model disease incidence during the coronavirus disease 2019 (COVID-19) pandemic. We discuss the use of social media for disease surveillance.

Background: Group B Streptococcus (GBS) can cause life-threating invasive infections, yet GBS is also a normal component of the intestinal and genitourinary tract. Although it is regarded as a potential urinary pathogen, the morbidity and mortality associated with recovery of GBS from urine cultures of nonpregnant adults is not well understood. We evaluated characteristics and mortality among nonpregnant adults with urine cultures that grew GBS. Methods: Using administrative data from the Veterans’ Healthcare Administration (VHA), we conducted a retrospective cohort study of VA healthcare system users from January 1, 2008, through December 31, 2017, with monomicrobial urine cultures growing ≥100,000 colony-forming units of GBS. Urinary tract infection (UTI) cases were defined as urinalysis positive for leukocyte esterase and pyuria (≥10 white blood cells), an International Classification of Diseases (ICD) code for UTI, and an antibiotic prescription. Cases with colonization were defined as negative for leukocyte esterase and pyuria, no ICD code for UTI, and no antibiotic prescription. Cases not meeting either definition were deemed unclassifiable. We compared demographics, comorbidities, and all-cause mortality among these 3 groups. Results: Over the 10-year study period, 26,848 veterans had 30,740 urine cultures positive for GBS. Applying the definitions above, there were 2,807 cases of infection, 8,789 cases of colonization, and 15,252 cases that were unclassifiable. Patients with a GBS UTI were slightly older compared to those who were colonized, with a higher Charlson comorbidity index and greater burden of chronic renal disease (Table 1). Individuals with infection versus colonization had 30-day mortality rates of 1% and 0%, respectively, and 1-year mortality rates of 9% and 4%, respectively (Figure 1). Conclusions: The association of a greater burden of illness among veterans who met our definition of UTI compared to colonization might be more reflective of providers’ responses to patients with chronic medical conditions rather than a difference in GBS as a cause of UTI. Overall, the prospect of a urine culture that grows GBS does not appear to be associated with adverse long-term outcomes.

Funding: No

Disclosures: None

Table 1.

Figure 1.

Background: The influence of increased use of telehealth during the emergence of COVID-19 on antibiotic prescriptions in outpatient settings is unknown. The VA Northeast Ohio Healthcare System has 13 community-based outpatient clinics (CBOCs) that provide primary and preventive care. We assessed changes in antibiotic prescriptions that occurred as care shifted from in-person to telehealth visits. Methods: Using VHA administrative databases, we identified all primary care CBOC visits between January 1, 2019, and December 31, 2020, that included a diagnosis for an acute respiratory infection (ARI), a urinary tract infection (UTI), or a skin or soft-tissue infection (SSTI), excluding visits with >1 of these diagnoses or with additional infectious diagnoses (eg, pneumonia, influenza). We summarized the proportion of telehealth visits and the proportion of patients prescribed antibiotics at quarterly intervals. We specifically assessed outpatient visits from April to December 2019 compared to the same months in 2020 to account for seasonality while analyzing diagnosis and antibiotic trends in the emergence of the COVID-19 pandemic. Results: The patients receiving care in April–December 2019 compared to April–December 2020 were similar (Table 1). From April through December 2019, 90% of CBOC primary care visits with a diagnosis for ARI, UTI, or SSTI were in-person, and antibiotics were prescribed at 63%, 46%, and 65% of visits in either modality, respectively (Figure 1). From April through December 2020, only 33% of CBOC primary care visits for ARI, UTI, and SSTI were in person, and antibiotics were prescribed at 46%, 38%, and 47% of visits in either modality, respectively. Comparing April–December in 2019 and 2020, the number of CBOC visits for ARI fell by 76% (2,152 visits to 509 visits), with a more modest decline of 20% and 35% observed for UTI and SSTI visits. In-person visits for ARIs and SSTIs were more likely than telehealth visits to result in an antibiotic prescription (Figure 2). Conclusions: Among the CBOCs at our healthcare system, an increase in the proportion of telehealth visits and a reduction in ARI diagnoses occurred after the emergence of COVID-19. In this setting, we observed a reduction in the proportion of visits for ARIs, UTIs, and SSTIs that included an antibiotic prescription.

Funding: Merck

Disclosures: None

Table 1.

Figure 1.

Figure 2.

Background: Outcomes among nursing home residents with asymptomatic compared to symptomatic COVID-19 are not well characterized. We assessed all-cause mortality among Veterans’ Affairs (VA) community living center (CLC) residents; we compared those residents with a negative SARS-CoV-2 test to residents with symptomatic, presymptomatic, and asymptomatic SARS-CoV-2 infections. Methods: We conducted a national retrospective cohort study of CLC residents tested for COVID-19 between March 1 and July 31, 2020, based on data compiled through the VA COVID-19 shared data resource. Among those with a positive SARS-CoV-2 test, residents were considered symptomatic if they had experienced COVID-19 symptoms in the 30 days prior to the test. Residents were considered presymptomatic if they did not experience symptoms in the 30 days prior to testing and developed a fever (>38°C) or required supplemental oxygen within 14 and 60 days, respectively, following the test. Residents were considered asymptomatic in the absence of these pre- and posttest symptoms. Results: From March 1 to July 31, 2020, of 9,052 CLC residents screened for COVID-19, 8,325 (92%) tested negative (Table 1). Among 727 residents with positive tests, 467 (64%) were symptomatic, 88 (12%) were presymptomatic, and 172 (24%) remained asymptomatic. We observed significant differences in the racial makeup of these disease groups. Among CLC residents who were symptomatic or presymptomatic, 176 (32%) of 555 were black compared to 39 (23%) of 172 who were asymptomatic and 1,810 (22%) of 8,325 who tested negative for SAR-CoV-2. All-cause 30-day mortality rates for symptomatic and presymptomatic residents were 25% and 34%, respectively, which exceeded the all-cause 30-day mortality of asymptomatic residents (12%) and residents with a negative test (6%) (Figure 1). Conclusions: More than one-third of CLC residents with COVID-19 were asymptomatic at the time of testing. This finding highlights the importance of vigilant infection prevention and control measures. Our finding that mortality among presymptomatic residents exceeded that of symptomatic residents raises consideration for enhancing supportive care measures, such as supplemental oxygen and mitigation of inflammatory reactions, as a means to reduce mortality among nursing home residents with presymptomatic SARS-CoV-2 infections.

Funding: No

Disclosures: None

Table 1.

Figure 1.

This article considers the origins, both scholarly and personal, of The American Renaissance 1876–1917 exhibition (1979) and the accompanying book catalogue, setting them in the context of architectural and cultural developments in the United States from the 1940s to the 1970s. It traces how the exhibition came about, what it was trying to achieve and how it was received, both at the time and subsequently. It shows that the exhibition was not conceived as an attack on modernism as such or as a work of architectural conservatism. Rather, it was an attempt to rescue from obscurity an entire chapter of American architectural history that had been excluded by the modernising narratives of Henry-Russell Hitchcock, Vincent Scully and others, and to reassess what this architecture might contribute to the present.

Background: Rates of invasive infections caused by caused group B Streptococcus (GBS) are increasing among adults. The burden of noninvasive GBS infections, including pneumonia, has not been well characterized. Here, we compare comorbidities and mortality associated with invasive and noninvasive pneumonia caused by GBS. Methods: Using the Veterans’ Health Administration national data warehouse, we studied a retrospective cohort review of veterans diagnosed with GBS pneumonia between 2008 and 2017. Invasive pneumonia was defined as blood cultures positive for GBS associated with an order for a chest x-ray and an International Classification of Disease (ICD) code for pneumonia. Noninvasive pneumonia was defined as a respiratory culture positive for GBS associated with both an order for a chest x-ray and an ICD code for pneumonia among patients with negative or without blood cultures. Patients with respiratory cultures positive for GBS without either an associated chest x-ray or ICD code for pneumonia were considered colonized. We compared demographics, comorbid conditions, and mortality among patients with invasive and noninvasive GBS pneumonia. Results: Between 2008 and 2017, we detected 706 cases of invasive GBS pneumonia, 1,244 cases of noninvasive GBS pneumonia, and 1,470 cases of respiratory colonization with GBS. Most patients were male (97%), with an average age of 69.0 years (SD, 12.0 years). The prevalence of several comorbid conditions differed between those with invasive and noninvasive disease: diabetes mellitus (61% and 46%, respectively); chronic pulmonary diseases (53% and 65%, respectively); chronic heart disease (58% and 44%, respectively), chronic kidney disease (43% and 27%, respectively). Mortality was similar among those with invasive and noninvasive GBS pneumonia at 30 days (17% and 18%, respectively) and at 1 year (38% and 43%, respectively) (Fig. 1). Conclusions: We identified important differences in underlying comorbid conditions between patients with invasive and noninvasive GBS pneumonia, which may give rise to differences in their clinical presentation. Overall mortality, however, was similar: more than one-third of patients with GBS pneumonia died within 1 year. These findings indicate that noninvasive GBS pneumonia is an important clinical entity.

Funding: This study was supported by Pfizer.

Disclosures: None