Refine
Document Type
- Article (3)
- Part of Periodical (1)
Is part of the Bibliography
- yes (4)
Keywords
- Aphasie (1)
- COVID-19 (1)
- EHR (1)
- Lese-Rechtschreibschwierigkeiten (1)
- Patholinguistik (1)
- Schluckstörungen (1)
- Spracherwerb (1)
- Sprachtherapie (1)
- TRIPOD (1)
- ahasia (1)
Background:
COVID-19 has infected millions of people worldwide and is responsible for several hundred thousand fatalities. The COVID-19 pandemic has necessitated thoughtful resource allocation and early identification of high-risk patients. However, effective methods to meet these needs are lacking.
Objective:
The aims of this study were to analyze the electronic health records (EHRs) of patients who tested positive for COVID-19 and were admitted to hospitals in the Mount Sinai Health System in New York City; to develop machine learning models for making predictions about the hospital course of the patients over clinically meaningful time horizons based on patient characteristics at admission; and to assess the performance of these models at multiple hospitals and time points.
Methods:
We used Extreme Gradient Boosting (XGBoost) and baseline comparator models to predict in-hospital mortality and critical events at time windows of 3, 5, 7, and 10 days from admission. Our study population included harmonized EHR data from five hospitals in New York City for 4098 COVID-19-positive patients admitted from March 15 to May 22, 2020. The models were first trained on patients from a single hospital (n=1514) before or on May 1, externally validated on patients from four other hospitals (n=2201) before or on May 1, and prospectively validated on all patients after May 1 (n=383). Finally, we established model interpretability to identify and rank variables that drive model predictions.
Results:
Upon cross-validation, the XGBoost classifier outperformed baseline models, with an area under the receiver operating characteristic curve (AUC-ROC) for mortality of 0.89 at 3 days, 0.85 at 5 and 7 days, and 0.84 at 10 days. XGBoost also performed well for critical event prediction, with an AUC-ROC of 0.80 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. In external validation, XGBoost achieved an AUC-ROC of 0.88 at 3 days, 0.86 at 5 days, 0.86 at 7 days, and 0.84 at 10 days for mortality prediction. Similarly, the unimputed XGBoost model achieved an AUC-ROC of 0.78 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. Trends in performance on prospective validation sets were similar. At 7 days, acute kidney injury on admission, elevated LDH, tachypnea, and hyperglycemia were the strongest drivers of critical event prediction, while higher age, anion gap, and C-reactive protein were the strongest drivers of mortality prediction.
Conclusions:
We externally and prospectively trained and validated machine learning models for mortality and critical events for patients with COVID-19 at different time horizons. These models identified at-risk patients and uncovered underlying relationships that predicted outcomes.
Das 10. Herbsttreffen Patholinguistik mit dem Schwerpunktthema »Panorama Patholinguistik: Sprachwissenschaft trifft Sprachtherapie« fand am 19.11.2016 in Potsdam statt. Das Herbsttreffen wird seit 2007 jährlich vom Verband für Patholinguistik e.V. (vpl) durchgeführt. Der vorliegende Tagungsband beinhaltet die vier Hauptvorträge zum Schwerpunktthema sowie Beiträge zu den Kurzvorträgen »Patholinguistik im Fokus« und der Posterpräsentationen zu weiteren Themen aus der sprachtherapeutischen Forschung und Praxis.
The near-infrared is an important part of the spectrum in astronomy, especially in cosmology because the light from objects in the early universe is redshifted to these wavelengths. However, deep near-infrared observations are extremely difficult to make from ground-based telescopes due to the bright background from the atmosphere. Nearly all of this background comes from the bright and narrow emission lines of atmospheric hydroxyl (OH) molecules. The atmospheric background cannot be easily removed from data because the brightness fluctuates unpredictably on short timescales. The sensitivity of ground-based optical astronomy far exceeds that of near-infrared astronomy because of this long-standing problem. GNOSIS is a prototype astrophotonic instrument that utilizes "OH suppression fibers" consisting of fiber Bragg gratings and photonic lanterns to suppress the 103 brightest atmospheric emission doublets between 1.47 and 1.7 mu m. GNOSIS was commissioned at the 3.9 m Anglo-Australian Telescope with the IRIS2 spectrograph to demonstrate the potential of OH suppression fibers, but may be potentially used with any telescope and spectrograph combination. Unlike previous atmospheric suppression techniques GNOSIS suppresses the lines before dispersion and in a manner that depends purely on wavelength. We present the instrument design and report the results of laboratory and on-sky tests from commissioning. While these tests demonstrated high throughput (approximate to 60%) and excellent suppression of the skylines by the OH suppression fibers, surprisingly GNOSIS produced no significant reduction in the interline background and the sensitivity of GNOSIS+IRIS2 is about the same as IRIS2. It is unclear whether the lack of reduction in the interline background is due to physical sources or systematic errors as the observations are detector noise dominated. OH suppression fibers could potentially impact ground-based astronomy at the level of adaptive optics or greater. However, until a clear reduction in the interline background and the corresponding increasing in sensitivity is demonstrated optimized OH suppression fibers paired with a fiber-fed spectrograph will at least provide a real benefit at low resolving powers.
1. Global pressures on freshwater ecosystems are high and rising. Viewed primarily as a resource for humans, current practices of water use have led to catastrophic declines in freshwater species and the degradation of freshwater ecosystems, including their genetic and functional diversity. Approximately three-quarters of the world's inland wetlands have been lost, one-third of the 28 000 freshwater species assessed for the International Union for Conservation of Nature (IUCN) Red List are threatened with extinction, and freshwater vertebrate populations are undergoing declines that are more rapid than those of terrestrial and marine species. This global loss continues unchecked, despite the importance of freshwater ecosystems as a source of clean water, food, livelihoods, recreation, and inspiration.
2. The causes of these declines include hydrological alterations, habitat degradation and loss, overexploitation, invasive species, pollution, and the multiple impacts of climate change. Although there are policy initiatives that aim to protect freshwater life, these are rarely implemented with sufficient conviction and enforcement. Policies that focus on the development and management of fresh waters as a resource for people almost universally neglect the biodiversity that they contain.
3. Here we introduce the Alliance for Freshwater Life, a global initiative, uniting specialists in research, data synthesis, conservation, education and outreach, and policymaking. This expert network aims to provide the critical mass required for the effective representation of freshwater biodiversity at policy meetings, to develop solutions balancing the needs of development and conservation, and to better convey the important role freshwater ecosystems play in human well-being. Through this united effort we hope to reverse this tide of loss and decline in freshwater biodiversity. We introduce several short- and medium-term actions as examples for making positive change, and invite individuals, organizations, authorities, and governments to join the Alliance for Freshwater Life.