Invisible ship tracks show large cloud sensitivity to aerosol.

Cloud reflectivity is sensitive to atmospheric aerosol concentrations because aerosols provide the condensation nuclei on which water condenses1. Increased aerosol concentrations due to human activity affect droplet number concentration, liquid water and cloud fraction2, but these changes are subject to large uncertainties3. Ship tracks, long lines of polluted clouds that are visible in satellite images, are one of the main tools for quantifying aerosol-cloud interactions4. However, only a small fraction of the clouds polluted by shipping show ship tracks5,6. Here we show that even when no ship tracks are visible in satellite images, aerosol emissions change cloud properties substantially. We develop a new method to quantify the effect of shipping on all clouds, showing a cloud droplet number increase and a more positive liquid water response when there are no visible tracks. We directly detect shipping-induced cloud property changes in the trade cumulus regions of the Atlantic, which are known to display almost no visible tracks. Our results indicate that previous studies of ship tracks were suffering from selection biases by focusing only on visible tracks from satellite imagery. The strong liquid water path response we find translates to a larger aerosol cooling effect on the climate, potentially masking a higher climate sensitivity than observed temperature trends would otherwise suggest.

Author Correction: Weak average liquid-cloud-water response to anthropogenic aerosols.

An Amendment to this paper has been published and can be accessed via a link at the top of the paper.

Weak average liquid-cloud-water response to anthropogenic aerosols.

The cooling of the Earth's climate through the effects of anthropogenic aerosols on clouds offsets an unknown fraction of greenhouse gas warming. An increase in the amount of water inside liquid-phase clouds induced by aerosols, through the suppression of rain formation, has been postulated to lead to substantial cooling, which would imply that the Earth's surface temperature is highly sensitive to anthropogenic forcing. Here we provide direct observational evidence that, instead of a strong increase, aerosols cause a relatively weak average decrease in the amount of water in liquid-phase clouds compared with unpolluted clouds. Measurements of polluted clouds downwind of various anthropogenic sources-such as oil refineries, smelters, coal-fired power plants, cities, wildfires and ships-reveal that aerosol-induced cloud-water increases, caused by suppressed rain formation, and decreases, caused by enhanced evaporation of cloud water, partially cancel each other out. We estimate that the observed decrease in cloud water offsets 23% of the global climate-cooling effect caused by aerosol-induced increases in the concentration of cloud droplets. These findings invalidate the hypothesis that increases in cloud water cause a substantial climate cooling effect and translate into reduced uncertainty in projections of future climate.

Shipping regulations lead to large reduction in cloud perturbations.

Global shipping accounts for 13% of global emissions of SO2, which, once oxidized to sulfate aerosol, acts to cool the planet both directly by scattering sunlight and indirectly by increasing the albedo of clouds. This cooling due to sulfate aerosol offsets some of the warming effect of greenhouse gasses and is the largest uncertainty in determining the change in the Earth's radiative balance by human activity. Ship tracks-the visible manifestation of the indirect of effect of ship emissions on clouds as quasi-linear features-have long provided an opportunity to quantify these effects. However, they have been arduous to catalog and typically studied only in particular regions for short periods of time. Using a machine-learning algorithm to automate their detection we catalog more than 1 million ship tracks to provide a global climatology. We use this to investigate the effect of stringent fuel regulations introduced by the International Maritime Organization in 2020 on their global prevalence since then, while accounting for the disruption in global commerce caused by COVID-19. We find a marked, but clearly nonlinear, decline in ship tracks globally: An 80% reduction in SO[Formula: see text] emissions causes only a 25% reduction in the number of tracks detected.

Aerosols enhance cloud lifetime and brightness along the stratus-to-cumulus transition.

Anthropogenic aerosols are hypothesized to enhance planetary albedo and offset some of the warming due to the buildup of greenhouse gases in Earth's atmosphere. Aerosols can enhance the coverage, reflectance, and lifetime of warm low-level clouds. However, the relationship between cloud lifetime and aerosol concentration has been challenging to measure from polar orbiting satellites. We estimate two timescales relating to the formation and persistence of low-level clouds over [Formula: see text] spatial domains using multiple years of geostationary satellite observations provided by the Clouds and Earth's Radiant Energy System (CERES) Synoptic (SYN) product. Lagrangian trajectories spanning several days along the classic stratus-to-cumulus transition zone are stratified by aerosol optical depth and meteorology. Clouds forming in relatively polluted trajectories tend to have lighter precipitation rates, longer average lifetime, and higher cloud albedo and cloud fraction compared with unpolluted trajectories. While liquid water path differences are found to be negligible, we find direct evidence of increased planetary albedo primarily through increased drop concentration ([Formula: see text]) and cloud fraction, with the caveat that the aerosol influence on cloud fraction is positive only for stable atmospheric conditions. While the increase in cloud fraction can be large typically in the beginning of trajectories, the Twomey effect accounts for the bulk (roughly 3/4) of the total aerosol indirect radiative forcing estimate.

Trends in Illness Severity, Hospitalization, and Mortality for Community-Onset Pneumonia at 118 US Veterans Affairs Medical Centers.

BACKGROUND: Deaths from pneumonia were decreasing globally prior to the COVID-19 pandemic, but it is unclear whether this was due to changes in patient populations, illness severity, diagnosis, hospitalization thresholds, or treatment. Using clinical data from the electronic health record among a national cohort of patients initially diagnosed with pneumonia, we examined temporal trends in severity of illness, hospitalization, and short- and long-term deaths. DESIGN: Retrospective cohort PARTICIPANTS: All patients >18 years presenting to emergency departments (EDs) at 118 VA Medical Centers between 1/1/2006 and 12/31/2016 with an initial clinical diagnosis of pneumonia and confirmed by chest imaging report. EXPOSURES: Year of encounter. MAIN MEASURES: Hospitalization and 30-day and 90-day mortality. Illness severity was defined as the probability of each outcome predicted by machine learning predictive models using age, sex, comorbidities, vital signs, and laboratory data from encounters during years 2006-2007, and similar models trained on encounters from years 2015 to 2016. We estimated the changes in hospitalizations and 30-day and 90-day mortality between the first and the last 2 years of the study period accounted for by illness severity using time covariate decompositions with model estimates. RESULTS: Among 196,899 encounters across the study period, hospitalization decreased from 71 to 63%, 30-day mortality 10 to 7%, 90-day mortality 16 to 12%, and 1-year mortality 29 to 24%. Comorbidity risk increased, but illness severity decreased. Decreases in illness severity accounted for 21-31% of the decrease in hospitalizations, and 45-47%, 32-24%, and 17-19% of the decrease in 30-day, 90-day, and 1-year mortality. Findings were similar among underrepresented patients and those with only hospital discharge diagnosis codes. CONCLUSIONS: Outcomes for community-onset pneumonia have improved across the VA healthcare system after accounting for illness severity, despite an increase in cases and comorbidity burden.

Simulated Adoption of 2019 Community-Acquired Pneumonia Guidelines Across 114 Veterans Affairs Medical Centers: Estimated Impact on Culturing and Antibiotic Selection in Hospitalized Patients.

BACKGROUND: The 2019 American Thoracic Society/Infectious Diseases Society of America guidelines for community-acquired pneumonia (CAP) revised recommendations for culturing and empiric broad-spectrum antibiotics. We simulated guideline adoption in Veterans Affairs (VA) inpatients. METHODS: For all VA acute hospitalizations for CAP from 2006-2016 nationwide, we compared observed with guideline-expected proportions of hospitalizations with initial blood and respiratory cultures obtained, empiric antibiotic therapy with activity against methicillin-resistant Staphylococcus aureus (anti-MRSA) or Pseudomonas aeruginosa (antipseudomonal), empiric "overcoverage" (receipt of anti-MRSA/antipseudomonal therapy without eventual detection of MRSA/P. aeruginosa on culture), and empiric "undercoverage" (lack of anti-MRSA/antipseudomonal therapy with eventual detection on culture). RESULTS: Of 115 036 CAP hospitalizations over 11 years, 17 877 (16%) were admitted to an intensive care unit (ICU). Guideline adoption would slightly increase respiratory culture (30% to 36%) and decrease blood culture proportions (93% to 36%) in hospital wards and increase both respiratory (40% to 100%) and blood (95% to 100%) cultures in ICUs. Adoption would decrease empiric selection of anti-MRSA (ward: 27% to 1%; ICU: 61% to 8%) and antipseudomonal (ward: 25% to 1%; ICU: 54% to 9%) therapies. This would correspond to greatly decreased MRSA overcoverage (ward: 27% to 1%; ICU: 56% to 8%), slightly increased MRSA undercoverage (ward: 0.6% to 1.3%; ICU: 0.5% to 3.3%), with similar findings for P. aeruginosa. For all comparisons, P < .001. CONCLUSIONS: Adoption of the 2019 CAP guidelines in this population would substantially change culturing and empiric antibiotic selection practices, with a decrease in overcoverage and slight increase in undercoverage for MRSA and P. aeruginosa.

Cardiorespiratory fitness, visceral fat, and body fat, but not dietary inflammatory index, are related to C-reactive protein in cancer survivors.

BACKGROUND: The control of chronic inflammation has emerged as a target for improving the health of cancer survivors (CS). AIM: To examine differences in fitness and dietary characteristics of CS when grouped by low vs. moderate to high serum C-reactive protein (CRP). METHODS: CS (N = 26, mean age = 68 ± 12 years) were evaluated for body mass index (BMI), body composition, cardiorespiratory fitness, dietary intake, dietary inflammatory index (DII), and serum CRP. Participants were assigned to one of two groups based on serum CRP concentrations: low CRP (≤1 mg/L) (LWC; n = 13) or moderate to high (CRP > 1 mg/L) (MHC; n = 13) and t-tests compared them. Data are presented as mean ± SD. RESULTS: LWC had higher VO2peak values (mL/kg/min) (p = 0.0003), and lower visceral fat area (cm2) (p = 0.02) and body fat mass (kg) (p = 0.04). Secondary analysis using Pearson's correlation coefficients, including all current study participant data, found significant negative relationships between CRP and total dietary fat intake (p = 0.02), saturated fat (p = 0.03), and polyunsaturated fat (p = 0.03). CONCLUSION: CS with moderate to high serum CRP concentrations had higher fat mass, visceral fat mass, and lower cardiorespiratory fitness. There was a significant negative relationship between dietary, fat, polyunsaturated and saturated fat, and CRP. However, these dietary fat related findings warrant further investigation. To summarize, improving cardiorespiratory fitness, maintaining lower body fat, may be helpful in altering chronic inflammation in CS.

Performance and Health-Related Characteristics of Physically Active Males Using Marijuana.

Lisano, JK, Smith, JD, Mathias, AB, Christensen, M, Smoak, P, Phillips, KT, Quinn, CJ, and Stewart, LK. Performance and health-related characteristics of physically active men using marijuana. J Strength Cond Res 33(6): 1659-1669, 2019-The influence of chronic marijuana use on the performance and health of physically active individuals has yet to be fully elucidated. The purpose of this study was to explore pulmonary function, aerobic and anaerobic fitness, strength, serum testosterone, cortisol, C-reactive protein (CRP), Δ-9-tetrahydrocannabinol (THC), 11-nor-9-carboxy-Δ-9-tetrahydrocannabinol (THC-COOH), and 11-hydroxy-Δ-9-tetrahydrocannabinol (THC-OH) concentrations in a physically active population either using or not using marijuana. Healthy, physically active males (N = 24) were compared based on their marijuana-use status: marijuana users (MU; n = 12) and nonusers (NU; n = 12). Statistical analysis (p = 0.05) revealed no difference between groups for age, body mass, body mass index, body fat, forced expiratory volume in 1 second percentage, VO2max, anaerobic power output, strength measures, testosterone, or cortisol concentrations. Although not statistically significant, MU showed a trend to fatigue to a greater percentage of absolute power output than NU from the beginning to the end of the Wingate Anaerobic Power Assessment (p = 0.08, effect size = 0.75). C-reactive protein in MU (1.76 ± 2.81 mg·L) and NU (0.86 ± 1.49 mg·L) was not significantly different (p = 0.60) but placed MU at moderate risk and NU at low risk for cardiovascular disease. Anaerobic fatigue was the only performance variable to show a trend for difference between groups. These results suggest that marijuana use in physically active males may not have significant effects on performance; however, it may be linked to elevated concentrations of CRP which place users at a higher risk for cardiovascular disease.

Remote Sensing of Droplet Number Concentration in Warm Clouds: A Review of the Current State of Knowledge and Perspectives.

The cloud droplet number concentration (N d) is of central interest to improve the understanding of cloud physics and for quantifying the effective radiative forcing by aerosol-cloud interactions. Current standard satellite retrievals do not operationally provide N d, but it can be inferred from retrievals of cloud optical depth (τ c) cloud droplet effective radius (r e) and cloud top temperature. This review summarizes issues with this approach and quantifies uncertainties. A total relative uncertainty of 78% is inferred for pixel-level retrievals for relatively homogeneous, optically thick and unobscured stratiform clouds with favorable viewing geometry. The uncertainty is even greater if these conditions are not met. For averages over 1° ×1° regions the uncertainty is reduced to 54% assuming random errors for instrument uncertainties. In contrast, the few evaluation studies against reference in situ observations suggest much better accuracy with little variability in the bias. More such studies are required for a better error characterization. N d uncertainty is dominated by errors in r e, and therefore, improvements in r e retrievals would greatly improve the quality of the N d retrievals. Recommendations are made for how this might be achieved. Some existing N d data sets are compared and discussed, and best practices for the use of N d data from current passive instruments (e.g., filtering criteria) are recommended. Emerging alternative N d estimates are also considered. First, new ideas to use additional information from existing and upcoming spaceborne instruments are discussed, and second, approaches using high-quality ground-based observations are examined.

Are three ports better than one? An evaluation of flow rates using all ports of a triple lumen central venous catheter in volume resuscitation.

BACKGROUND: Poiseuille's law states flow rates are directly proportional to the radius to the 4th power and indirectly proportional to the length of a tube. Because of this property, large bore catheters are commonly used in the resuscitation of the critically ill patient. However, there are no studies comparing simultaneous use of all three lumens of a triple lumen (TL) central venous catheter (CVC) with other catheter types. Our objective was to compare the flow rates of normal saline (NS) through various resuscitation catheters against a TL CVC using all 3 ports. METHODS: We performed a blinded prospective observational study of flow rates utilizing multiple resuscitation catheters. Each catheter type was attached to a 1l bag of NS using standard saline tubing and mean time to infuse 1l of normal saline was determined. Three trials each were completed with and without pressure bags. RESULTS: Simultaneous infusion of NS through all ports of a TL CVC demonstrated no statistically significant difference compared to the following catheters: 16ga peripheral venous catheter (PVC) and 6 Fr CVC with pressure bag. The 14 g PVC and 8.5Fr CVC had statistically significant faster flow rates than the TL CVC both with and without a pressure bag. The 6Fr CVC showed significantly faster flow rates than the TL CVC without a pressure bag. CONCLUSIONS: Simultaneous use of all 3 ports of a TL CVC generates flow rates comparable to many other commonly used resuscitation catheters.

Volcano and ship tracks indicate excessive aerosol-induced cloud water increases in a climate model.

Aerosol-cloud interaction is the most uncertain mechanism of anthropogenic radiative forcing of Earth's climate, and aerosol-induced cloud water changes are particularly poorly constrained in climate models. By combining satellite retrievals of volcano and ship tracks in stratocumulus clouds, we compile a unique observational dataset and confirm that liquid water path (LWP) responses to aerosols are bidirectional, and on average the increases in LWP are closely compensated by the decreases. Moreover, the meteorological parameters controlling the LWP responses are strikingly similar between the volcano and ship tracks. In stark contrast to observations, there are substantial unidirectional increases in LWP in the Hadley Centre climate model, because the model accounts only for the decreased precipitation efficiency and not for the enhanced entrainment drying. If the LWP increases in the model were compensated by the decreases as the observations suggest, its indirect aerosol radiative forcing in stratocumulus regions would decrease by 45%.

The histone deacetylase inhibitor Entinostat enhances polymer-mediated transgene expression in cancer cell lines.

Eukaryotic cells maintain an immense amount of genetic information by tightly wrapping their DNA around positively charged histones. While this strategy allows human cells to maintain more than 25,000 genes, histone binding can also block gene expression. Consequently, cells express histone acetyl transferases (HATs) to acetylate histone lysines and release DNA for transcription. Conversely, histone deacetylases (HDACs) are employed for restoring the positive charge on the histones, thereby silencing gene expression by increasing histone-DNA binding. It has previously been shown that histones bind and silence viral DNA, while hyperacetylation of histones via HDAC inhibition restores viral gene expression. In this study, we demonstrate that treatment with Entinostat, an HDAC inhibitor, enhances transgene (luciferase) expression by up to 25-fold in human prostate and murine bladder cancer cell lines when used with cationic polymers for plasmid DNA delivery. Entinostat treatment altered cell cycle progression, resulting in a significant increase in the fraction of cells present in the G0/G1 phase at low micromolar concentrations. While this moderate G0/G1 arrest disappeared at higher concentrations, a modest increase in the fraction of apoptotic cells and a decrease in cell proliferation were observed, consistent with the known anticancer effects of the drug. DNase accessibility studies revealed no significant change in plasmid transcriptional availability with Entinostat treatment. However, quantitative PCR studies indicated that Entinostat treatment, at the optimal dose for enhancing transgene expression, led to an increase in the amount of plasmid present in the nucleus in two cancer cell lines. Taken together, our results show that Entinostat enhances polymer- mediated transgene expression and can be useful in applications related to transient protein expression in mammalian cells. Biotechnol. Bioeng. 2016;113: 1345-1356. © 2015 Wiley Periodicals, Inc.

Innovative Power Wheelchair Control Interface: A Proof-of-Concept Study.

Some people without independent mobility are candidates for powered mobility but are unable to use a traditional power wheelchair joystick. This proof-of-concept study tested and further developed an innovative method of driving power wheelchairs for people whose impairments prevent them from operating commercial wheelchair controls. Our concept, Self-referenced Personal Orthotic Omni-purpose Control Interface (SPOOCI), is distinguished by referencing the control sensor not to the wheelchair frame but instead to the adjacent proximal lower-extremity segment via a custom-formed orthosis. Using a descriptive case-series design, we compared the pre-post functional power wheelchair driving skill data of 4 participants, measured by the Power Mobility Program, using descriptive analyses. The intervention consisted of standard-care power wheelchair training during 12 outpatient occupational or physical therapy sessions. All 4 participants who completed the 12-wk intervention improved their functional power wheelchair driving skills using SPOOCI, but only 3 were deemed safe to continue with power wheelchair driving.

Vision-language foundation model for echocardiogram interpretation.

The development of robust artificial intelligence models for echocardiography has been limited by the availability of annotated clinical data. Here, to address this challenge and improve the performance of cardiac imaging models, we developed EchoCLIP, a vision-language foundation model for echocardiography, that learns the relationship between cardiac ultrasound images and the interpretations of expert cardiologists across a wide range of patients and indications for imaging. After training on 1,032,975 cardiac ultrasound videos and corresponding expert text, EchoCLIP performs well on a diverse range of benchmarks for cardiac image interpretation, despite not having been explicitly trained for individual interpretation tasks. EchoCLIP can assess cardiac function (mean absolute error of 7.1% when predicting left ventricular ejection fraction in an external validation dataset) and identify implanted intracardiac devices (area under the curve (AUC) of 0.84, 0.92 and 0.97 for pacemakers, percutaneous mitral valve repair and artificial aortic valves, respectively). We also developed a long-context variant (EchoCLIP-R) using a custom tokenizer based on common echocardiography concepts. EchoCLIP-R accurately identified unique patients across multiple videos (AUC of 0.86), identified clinical transitions such as heart transplants (AUC of 0.79) and cardiac surgery (AUC 0.77) and enabled robust image-to-text search (mean cross-modal retrieval rank in the top 1% of candidate text reports). These capabilities represent a substantial step toward understanding and applying foundation models in cardiovascular imaging for preliminary interpretation of echocardiographic findings.

Deep learning for transesophageal echocardiography view classification.

Transesophageal echocardiography (TEE) imaging is a vital tool used in the evaluation of complex cardiac pathology and the management of cardiac surgery patients. A key limitation to the application of deep learning strategies to intraoperative and intraprocedural TEE data is the complexity and unstructured nature of these images. In the present study, we developed a deep learning-based, multi-category TEE view classification model that can be used to add structure to intraoperative and intraprocedural TEE imaging data. More specifically, we trained a convolutional neural network (CNN) to predict standardized TEE views using labeled intraoperative and intraprocedural TEE videos from Cedars-Sinai Medical Center (CSMC). We externally validated our model on intraoperative TEE videos from Stanford University Medical Center (SUMC). Accuracy of our model was high across all labeled views. The highest performance was achieved for the Trans-Gastric Left Ventricular Short Axis View (area under the receiver operating curve [AUC] = 0.971 at CSMC, 0.957 at SUMC), the Mid-Esophageal Long Axis View (AUC = 0.954 at CSMC, 0.905 at SUMC), the Mid-Esophageal Aortic Valve Short Axis View (AUC = 0.946 at CSMC, 0.898 at SUMC), and the Mid-Esophageal 4-Chamber View (AUC = 0.939 at CSMC, 0.902 at SUMC). Ultimately, we demonstrate that our deep learning model can accurately classify standardized TEE views, which will facilitate further downstream deep learning analyses for intraoperative and intraprocedural TEE imaging.

Physical science research needed to evaluate the viability and risks of marine cloud brightening.

Marine cloud brightening (MCB) is the deliberate injection of aerosol particles into shallow marine clouds to increase their reflection of solar radiation and reduce the amount of energy absorbed by the climate system. From the physical science perspective, the consensus of a broad international group of scientists is that the viability of MCB will ultimately depend on whether observations and models can robustly assess the scale-up of local-to-global brightening in today's climate and identify strategies that will ensure an equitable geographical distribution of the benefits and risks associated with projected regional changes in temperature and precipitation. To address the physical science knowledge gaps required to assess the societal implications of MCB, we propose a substantial and targeted program of research-field and laboratory experiments, monitoring, and numerical modeling across a range of scales.

Limitations in the use of automated mental status detection for clinical decision support.

BACKGROUND: Clinical decision support (CDS) tools improve adherence to evidence-based practices but are dependent upon data quality in the electronic health record (EHR). Mental status is an integral component of many risk stratification scores, but it is not known whether EHR-measures of altered mental status are reliable. The Glasgow Coma Scale (GCS) is a measure of altered mentation that is widely adopted and entered in the EHR in structured format. We sought to determine the accuracy GCS < 15 as an EHR-measure of altered mentation compared to ED provider documentation. METHODS: In patients presenting to an academic Emergency Department (ED) with pneumonia we abstracted GCS values entered by nurses during routine care and in a randomly selected subset manually reviewed provider documentation for evidence of altered mental status. We defined eConfusion as present if GCS < 15 at any point during the ED encounter. We then calculated the CURB-65 score and corresponding suggested disposition using each method. Performance of eConfusion and corresponding CURB-65 compared to manual versions was measured using agreement (Cohen's K), sensitivity, and specificity. RESULTS: Among 300 randomly selected encounters, 47 (16 %) had eConfusion present and 46 (15 %) had evidence of altered mental status in provider documentation with Cohen's K 0.73. eConfusion had 78 % sensitivity and 96 % specificity for provider documented altered mental status. When input into CURB-65 to recommend inpatient disposition, eConfusion had 95 % sensitivity, and recommended discordant disposition for 3 %. CONCLUSIONS: There was modest agreement between eConfusion and provider documentation of altered mental status. eConfusion had good specificity but low sensitivity which resulted in under-estimation of the CURB-65 score and occasional inappropriate disposition recommendations compared to provider documentation. These data do not support the use of GCS as a measure for altered mentation for use in CDS tools in the ED.

Deep learning-based electrocardiographic screening for chronic kidney disease.

BACKGROUND: Undiagnosed chronic kidney disease (CKD) is a common and usually asymptomatic disorder that causes a high burden of morbidity and early mortality worldwide. We developed a deep learning model for CKD screening from routinely acquired ECGs. METHODS: We collected data from a primary cohort with 111,370 patients which had 247,655 ECGs between 2005 and 2019. Using this data, we developed, trained, validated, and tested a deep learning model to predict whether an ECG was taken within one year of the patient receiving a CKD diagnosis. The model was additionally validated using an external cohort from another healthcare system which had 312,145 patients with 896,620 ECGs between 2005 and 2018. RESULTS: Using 12-lead ECG waveforms, our deep learning algorithm achieves discrimination for CKD of any stage with an AUC of 0.767 (95% CI 0.760-0.773) in a held-out test set and an AUC of 0.709 (0.708-0.710) in the external cohort. Our 12-lead ECG-based model performance is consistent across the severity of CKD, with an AUC of 0.753 (0.735-0.770) for mild CKD, AUC of 0.759 (0.750-0.767) for moderate-severe CKD, and an AUC of 0.783 (0.773-0.793) for ESRD. In patients under 60 years old, our model achieves high performance in detecting any stage CKD with both 12-lead (AUC 0.843 [0.836-0.852]) and 1-lead ECG waveform (0.824 [0.815-0.832]). CONCLUSIONS: Our deep learning algorithm is able to detect CKD using ECG waveforms, with stronger performance in younger patients and more severe CKD stages. This ECG algorithm has the potential to augment screening for CKD.

Chronic kidney disease (CKD) is a common condition involving loss of kidney function over time and results in a substantial number of deaths. However, CKD often has no symptoms during its early stages. To detect CKD earlier, we developed a computational approach for CKD screening using routinely acquired electrocardiograms (ECGs), a cheap, rapid, non-invasive, and commonly obtained test of the heart's electrical activity. Our model achieved good accuracy in identifying any stage of CKD, with especially high accuracy in younger patients and more severe stages of CKD. Given the high global burden of undiagnosed CKD, novel and accessible CKD screening strategies have the potential to help prevent disease progression and reduce premature deaths related to CKD.

Electronic surveillance criteria for non-ventilator-associated hospital-acquired pneumonia: Assessment of reliability and validity.

OBJECTIVE: Surveillance of non-ventilator-associated hospital-acquired pneumonia (NV-HAP) is complicated by subjectivity and variability in diagnosing pneumonia. We compared a fully automatable surveillance definition using routine electronic health record data to manual determinations of NV-HAP according to surveillance criteria and clinical diagnoses. METHODS: We retrospectively applied an electronic surveillance definition for NV-HAP to all adults admitted to Veterans' Affairs (VA) hospitals from January 1, 2015, to November 30, 2020. We randomly selected 250 hospitalizations meeting NV-HAP surveillance criteria for independent review by 2 clinicians and calculated the percent of hospitalizations with (1) clinical deterioration, (2) CDC National Healthcare Safety Network (CDC-NHSN) criteria, (3) NV-HAP according to a reviewer, (4) NV-HAP according to a treating clinician, (5) pneumonia diagnosis in discharge summary; and (6) discharge diagnosis codes for HAP. We assessed interrater reliability by calculating simple agreement and the Cohen κ (kappa). RESULTS: Among 3.1 million hospitalizations, 14,023 met NV-HAP electronic surveillance criteria. Among reviewed cases, 98% had a confirmed clinical deterioration; 67% met CDC-NHSN criteria; 71% had NV-HAP according to a reviewer; 60% had NV-HAP according to a treating clinician; 49% had a discharge summary diagnosis of pneumonia; and 82% had NV-HAP according to any definition according to at least 1 reviewer. Only 8% had diagnosis codes for HAP. Interrater agreement was 75% (κ = 0.50) for CDC-NHSN criteria and 78% (κ = 0.55) for reviewer diagnosis of NV-HAP. CONCLUSIONS: Electronic NV-HAP surveillance criteria correlated moderately with existing manual surveillance criteria. Reviewer variability for all manual assessments was high. Electronic surveillance using clinical data may therefore allow for more consistent and efficient surveillance with similar accuracy compared to manual assessments or diagnosis codes.