Deep learning prediction of hospital readmissions for asthma and COPD.

QUESTION: Severe asthma and COPD exacerbations requiring hospitalization are linked to increased disease morbidity and healthcare costs. We sought to identify Electronic Health Record (EHR) features of severe asthma and COPD exacerbations and evaluate the performance of four machine learning (ML) and one deep learning (DL) model in predicting readmissions using EHR data. STUDY DESIGN AND METHODS: Observational study between September 30, 2012, and December 31, 2017, of patients hospitalized with asthma and COPD exacerbations. RESULTS: This study included 5,794 patients, 1,893 with asthma and 3,901 with COPD. Patients with asthma were predominantly female (n = 1288 [68%]), 35% were Black (n = 669), and 25% (n = 479) were Hispanic. Black (44 vs. 33%, p = 0.01) and Hispanic patients (30 vs. 24%, p = 0.02) were more likely to be readmitted for asthma. Similarly, patients with COPD readmissions included a large percentage of Blacks (18 vs. 10%, p < 0.01) and Hispanics (8 vs. 5%, p < 0.01). To identify patients at high risk of readmission index hospitalization data of a subset of 2,682 patients, 777 with asthma and 1,905 with COPD, was analyzed with four ML models, and one DL model. We found that multilayer perceptron, the DL method, had the best sensitivity and specificity compared to the four ML methods implemented in the same dataset. INTERPRETATION: Multilayer perceptron, a deep learning method, had the best performance in predicting asthma and COPD readmissions, demonstrating that EHR and deep learning integration can improve high-risk patient detection.

Shared genetic architecture of blood eosinophil counts and asthma in UK Biobank.

Rationale: Asthma is a complex, heterogeneous disease strongly associated with type 2 inflammation, and blood eosinophil counts guide therapeutic interventions in moderate and severe asthma. Eosinophils are leukocytes involved in type 2 immune responses. Despite these critical associations between asthma and blood eosinophil counts, the shared genetic architecture of these two traits remains unknown. The objective of the present study was to characterise the genetic architecture of blood eosinophil counts and asthma in the UK Biobank. Methods: We performed genome-wide association studies (GWAS) of doctor-diagnosed asthma, blood eosinophil, neutrophil, lymphocyte and monocyte counts in the UK Biobank. Genetic correlation analysis was performed on GWAS results and validated in the Trans-National Asthma Genetic Consortium (TAGC) study of asthma. Results: GWAS of doctor-diagnosed asthma and blood eosinophil counts in the UK Biobank identified 585 and 3429 significant variants, respectively. STAT6, a transcription factor involved in interleukin-4 signalling, was a key shared pathway between asthma and blood eosinophil counts. Genetic correlation analysis demonstrated a positive correlation between doctor-diagnosed asthma and blood eosinophil counts (r=0.38±0.10, correlation±se; p=4.7×10-11). As a validation of this association, we found a similar correlation between TAGC and blood eosinophil counts in the UK Biobank (0.37±0.08, correlation±se; p=1.2×10-6). Conclusions: These findings define the shared genetic architecture between blood eosinophil counts and asthma risk in subjects of European ancestry and point to a genetic link to the STAT6 signalling pathway in these two traits.

Multiscale predictors of small tree survival across a heterogeneous tropical landscape.

Uncertainties about controls on tree mortality make forest responses to land-use and climate change difficult to predict. We tracked biomass of tree functional groups in tropical forest inventories across Puerto Rico and the U.S. Virgin Islands, and with random forests we ranked 86 potential predictors of small tree survival (young or mature stems 2.5-12.6 cm diameter at breast height). Forests span dry to cloud forests, range in age, geology and past land use and experienced severe drought and storms. When excluding species as a predictor, top predictors are tree crown ratio and height, two to three species traits and stand to regional factors reflecting local disturbance and the system state (widespread recovery, drought, hurricanes). Native species, and species with denser wood, taller maximum height, or medium typical height survive longer, but short trees and species survive hurricanes better. Trees survive longer in older stands and with less disturbed canopies, harsher geoclimates (dry, edaphically dry, e.g., serpentine substrates, and highest-elevation cloud forest), or in intervals removed from hurricanes. Satellite image phenology and bands, even from past decades, are top predictors, being sensitive to vegetation type and disturbance. Covariation between stand-level species traits and geoclimate, disturbance and neighboring species types may explain why most neighbor variables, including introduced vs. native species, had low or no importance, despite univariate correlations with survival. As forests recovered from a hurricane in 1998 and earlier deforestation, small trees of introduced species, which on average have lighter wood, died at twice the rate of natives. After hurricanes in 2017, the total biomass of trees ≥12.7 cm dbh of the introduced species Spathodea campanulata spiked, suggesting that more frequent hurricanes might perpetuate this light-wooded species commonness. If hurricane recovery favors light-wooded species while drought favors others, climate change influences on forest composition and ecosystem services may depend on the frequency and severity of extreme climate events.

An Overview of Sex and Gender in Pulmonary and Critical Care Medicine.

Disparities in health care have risen to the forefront of medicine in the past several years. One of the most notable disparities in the research and delivery of health care relates to sex and gender. Sex and gender affect the epidemiology, pathophysiology, and outcomes of disease and social determinants of health and access to medical care. This article discusses some of the history of considering sex as a biologic variable in medical research and clinical care. It also clarifies the definitions and terminology necessary for understanding the biologic and social underpinnings of sex and gender.

Endemic Jeffrey Pine Beetle Associates: Beetle/Mite Fungal Dissemination Strategies and Interactions That May Influence Beetle Population Levels.

Fungal and mite associates may drive changes in bark beetle populations, and mechanisms constraining beetle irruptions may be hidden in endemic populations. We characterized common fungi of endemic-level Jeffrey pine beetle (JPB) in western USA and analyzed their dissemination by JPB (maxillae and fecal pellet) and fungivorous mites to identify if endogenous regulation drove the population. We hypothesized that: (1) as in near-endemic mountain pine beetle populations, JPB's mutualistic fungus would either be less abundant in endemic than in non-endemic populations or that another fungus may be more prevalent; (2) JPB primarily transports its mutualistic fungus, while its fungivorous mites primarily transport another fungus, and (3) based on the prevalence of yeasts in bark beetle symbioses, that a mutualistic interaction with blue-stain fungi present in that system may exist. Grosmannia clavigera was the most frequent JPB symbiont; however, the new here reported antagonist, Ophiostoma minus, was second in frequency. As hypothesized, JPB mostly carried its mutualist fungus while another fungus (i.e., antagonistic) was mainly carried by mites, but no fungal transport was obligate. Furthermore, we found a novel mutualistic interaction between the yeast Kuraishia molischiana and G. clavigera which fostered a growth advantage at temperatures associated with beetle colonization.

The Patient Care Ownership Scale: External Validation of an Instrument that Measures Patient Care Ownership Among Internal Medicine Trainees-a Multi-Institutional Study.

BACKGROUND: Patient care ownership improves accountability, clinical skills, and quality of patient care among resident physicians, but appears to be gradually eroding. Research is limited by the lack of a reliable, objective measure of ownership. OBJECTIVE: To validate the Patient Care Ownership Scale, an instrument that measures decision ownership among internal medicine residents. DESIGN: Multi-institutional, cross-sectional study using a 66-item, online survey that queried residents on ownership's key constructs (advocacy, responsibility, accountability, follow-through, knowledge, communication, initiative, continuity of care, autonomy, self-efficacy, and perceived ownership) as well as mood and burnout. PARTICIPANTS: Internal medicine residents in five geographically diverse residency programs completing an inpatient rotation. MAIN MEASURES: We performed exploratory and confirmatory factor analysis in two randomly split groups to evaluate for subscales and inform item reduction. We conducted reliability testing with Cronbach's α. We performed bivariate analyses to examine construct validity and identify correlates of ownership. KEY RESULTS: Of the 785 eligible residents, 625 completed the survey (80% response rate); we included responses from 563 in the analysis. We identified three factors corresponding to assertiveness, conscientiousness, and confidence or perceived competence. After iterative item reduction, the 13-item ownership scale demonstrated good reliability (Cronbach's α = 0.82). Convergent validity was supported by a significant association with perceived ownership (eliminated from the final scale) (r = 0.67, p < 0.001). There was a positive association between ownership and training level (p < 0.01) and prior experience in the intensive care unit (p < 0.001). There were significant, inverse relationships between ownership and self-defined burnout (r = - 0.24, p < 0.001), depression (r = - 0.22, p < 0.001), detachment (r = - 0.26, p < 0.001), and frustration (r = - 0.15, p = 0.02), and significant positive associations between ownership and feeling energetic (r = 0.29, p < 0.001), happy (r = 0.33, p < 0.001), and fulfilled (r = 0.34, p < 0.001). CONCLUSIONS: The Patient Care Ownership Scale is valid in diverse residency program settings. Medical educators and investigators can use our scale to assess interventions aimed at fostering ownership.

Factors and costs associated with removal of a newly established population of invasive wild pigs in Northern U.S.

The human-mediated spread of exotic and invasive species often leads to unintentional and harmful consequences. Invasive wild pigs (Sus scrofa) are one such species that have been repeatedly translocated throughout the United States and cause extensive damage to natural ecosystems, threatened and endangered species, agricultural resources, and private lands. In 2005, a newly established population of wild pigs was confirmed in Fulton County, Illinois, U.S. In 2011, a state-wide wild pig damage management program involving federal, state, and local government authorities directed a concerted effort to remove wild pigs from the county until the last wild pig (of 376 total) was successfully removed in 2016. We examined surveillance data from camera traps at bait sites and records of wild pig removals during this elimination program to identify environmental and anthropogenic factors that optimized removal of this population. Our results revealed that wild pigs used bait sites most during evening and nocturnal periods and on days with lower daily maximum temperatures. Increased removals of wild pigs coincided with periods of cold weather. We also identified that fidelity and time spent at bait sites by wild pigs was not influenced by increasing removals of wild pigs. Finally, the costs to remove wild pigs averaged $50 per wild pig (6.8 effort hours per wild pig) for removing the first 99% of the animals. Cost for removing the last 1% increased 84-fold, and averaged 122.8 effort hours per wild pig removed. Our results demonstrated that increased effort in removing wild pigs using bait sites should be focused during periods of environmental stress to maximize removal efficiency. These results inform elimination programs attempting to remove newly established populations of wild pigs, and ultimately prevent population and geographic expansion.

Factors affecting lifetime reproduction, long-term territory-specific reproduction, and estimation of habitat quality in northern goshawks.

One measure of habitat quality is a species' demographic performance in a habitat and the gold standard metric of performance is reproduction. Such a measure, however, may be misleading if individual quality is a fitness determinant. We report on factors affecting lifetime reproduction (LR), the total number of lifetime fledglings produced by an individual, and long-term territory-specific reproduction in a multi-generational study of northern goshawks (Accipiter gentilis). LR increased with longer lifespans and more breeding attempts and was strongly correlated with the number of recruits in two filial generations indicating that LR was a good fitness predictor. Extensive differences in LR attested to heterogeneity in individual quality, a requisite for the ideal pre-emptive distribution model (IPD) of habitat settling wherein high quality individuals get the best habitats forcing lower quality individuals into poorer habitats with lower reproduction. In response to 7‒9-year prey abundance cycles, annual frequency of territory occupancy by breeders was highly variable and low overall with monotonic increases in vacancies through low prey years. Occupancy of territories by breeders differed from random; some appeared preferred while others were avoided, producing a right-skewed distribution of total territory-specific fledgling production. However, mean fledglings per nest attempt was only slightly lower in less versus more productive territories, and, contrary to IPD predictions of increases in annual territory-specific coefficients of variation (CV) in reproduction as breeder densities increase, the CV of production decreased as density increased. Rather than habitat quality per se, conspecific attraction elicited territory selection by prospecting goshawks as 70% of settlers comprised turnovers on territories, resulting in occupancy continuity and increased territory-specific reproduction. Top-producing territories had as few as 2 long-lived (high LR) and up to 6 short-lived (low LR) sequential breeders. While individual quality appeared to effect territory-specific heterogeneity in reproductive performance, our data suggests that differences in individual quality may be washed-out by a random settling of prospectors in response to conspecific attraction.

Quantifying drivers of wild pig movement across multiple spatial and temporal scales.

BACKGROUND: The movement behavior of an animal is determined by extrinsic and intrinsic factors that operate at multiple spatio-temporal scales, yet much of our knowledge of animal movement comes from studies that examine only one or two scales concurrently. Understanding the drivers of animal movement across multiple scales is crucial for understanding the fundamentals of movement ecology, predicting changes in distribution, describing disease dynamics, and identifying efficient methods of wildlife conservation and management. METHODS: We obtained over 400,000 GPS locations of wild pigs from 13 different studies spanning six states in southern U.S.A., and quantified movement rates and home range size within a single analytical framework. We used a generalized additive mixed model framework to quantify the effects of five broad predictor categories on movement: individual-level attributes, geographic factors, landscape attributes, meteorological conditions, and temporal variables. We examined effects of predictors across three temporal scales: daily, monthly, and using all data during the study period. We considered both local environmental factors such as daily weather data and distance to various resources on the landscape, as well as factors acting at a broader spatial scale such as ecoregion and season. RESULTS: We found meteorological variables (temperature and pressure), landscape features (distance to water sources), a broad-scale geographic factor (ecoregion), and individual-level characteristics (sex-age class), drove wild pig movement across all scales, but both the magnitude and shape of covariate relationships to movement differed across temporal scales. CONCLUSIONS: The analytical framework we present can be used to assess movement patterns arising from multiple data sources for a range of species while accounting for spatio-temporal correlations. Our analyses show the magnitude by which reaction norms can change based on the temporal scale of response data, illustrating the importance of appropriately defining temporal scales of both the movement response and covariates depending on the intended implications of research (e.g., predicting effects of movement due to climate change versus planning local-scale management). We argue that consideration of multiple spatial scales within the same framework (rather than comparing across separate studies post-hoc) gives a more accurate quantification of cross-scale spatial effects by appropriately accounting for error correlation.

Inferring infection hazard in wildlife populations by linking data across individual and population scales.

Our ability to infer unobservable disease-dynamic processes such as force of infection (infection hazard for susceptible hosts) has transformed our understanding of disease transmission mechanisms and capacity to predict disease dynamics. Conventional methods for inferring FOI estimate a time-averaged value and are based on population-level processes. Because many pathogens exhibit epidemic cycling and FOI is the result of processes acting across the scales of individuals and populations, a flexible framework that extends to epidemic dynamics and links within-host processes to FOI is needed. Specifically, within-host antibody kinetics in wildlife hosts can be short-lived and produce patterns that are repeatable across individuals, suggesting individual-level antibody concentrations could be used to infer time since infection and hence FOI. Using simulations and case studies (influenza A in lesser snow geese and Yersinia pestis in coyotes), we argue that with careful experimental and surveillance design, the population-level FOI signal can be recovered from individual-level antibody kinetics, despite substantial individual-level variation. In addition to improving inference, the cross-scale quantitative antibody approach we describe can reveal insights into drivers of individual-based variation in disease response, and the role of poorly understood processes such as secondary infections, in population-level dynamics of disease.

The basis function approach for modeling autocorrelation in ecological data.

Analyzing ecological data often requires modeling the autocorrelation created by spatial and temporal processes. Many seemingly disparate statistical methods used to account for autocorrelation can be expressed as regression models that include basis functions. Basis functions also enable ecologists to modify a wide range of existing ecological models in order to account for autocorrelation, which can improve inference and predictive accuracy. Furthermore, understanding the properties of basis functions is essential for evaluating the fit of spatial or time-series models, detecting a hidden form of collinearity, and analyzing large data sets. We present important concepts and properties related to basis functions and illustrate several tools and techniques ecologists can use when modeling autocorrelation in ecological data.

Evaluating wildlife-cattle contact rates to improve the understanding of dynamics of bovine tuberculosis transmission in Michigan, USA.

Direct and indirect contacts among individuals drive transmission of infectious disease. When multiple interacting species are susceptible to the same pathogen, risk assessment must include all potential host species. Bovine tuberculosis (bTB) is an example of a disease that can be transmitted among several wildlife species and to cattle, although the potential role of several wildlife species in spillback to cattle remains unclear. To better understand the complex network of contacts and factors driving disease transmission, we fitted proximity logger collars to beef and dairy cattle (n=37), white-tailed deer (Odocoileus virginianus; n=29), raccoon (Procyon lotor; n=53), and Virginia opossum (Didelphis virginiana; n=79) for 16 months in Michigan's Lower Peninsula, USA. We determined inter- and intra-species direct and indirect contact rates. Data on indirect contact was calculated when collared animals visited stationary proximity loggers placed at cattle feed and water resources. Most contact between wildlife species and cattle was indirect, with the highest contact rates occurring between raccoons and cattle during summer and fall. Nearly all visits (>99%) to cattle feed and water sources were by cattle, whereas visitation to stored cattle feed was dominated by deer and raccoon (46% and 38%, respectively). Our results suggest that indirect contact resulting from wildlife species visiting cattle-related resources could pose a risk of disease transmission to cattle and deserves continued attention with active mitigation.