Search

By September 2022, more than 600 million cases of SARS-CoV-2 infection have been reported globally, resulting in over 6.5 million deaths. COVID-19 mortality risk estimators are often, however, developed with small unrepresentative samples and with methodological limitations. It is highly important to develop predictive tools for pulmonary embolism (PE) in COVID-19 patients as one of the most severe preventable complications of COVID-19. Early recognition can help provide life-saving targeted anti-coagulation therapy right at admission. Using a dataset of more than 800,000 COVID-19 patients from an international cohort, we propose a cost-sensitive gradient-boosted machine learning model that predicts occurrence of PE and death at admission. Logistic regression, Cox proportional hazards models, and Shapley values were used to identify key predictors for PE and death. Our prediction model had a test AUROC of 75.9% and 74.2%, and sensitivities of 67.5% and 72.7% for PE and all-cause mortality respectively on a highly diverse and held-out test set. The PE prediction model was also evaluated on patients in UK and Spain separately with test results of 74.5% AUROC, 63.5% sensitivity and 78.9% AUROC, 95.7% sensitivity. Age, sex, region of admission, comorbidities (chronic cardiac and pulmonary disease, dementia, diabetes, hypertension, cancer, obesity, smoking), and symptoms (any, confusion, chest pain, fatigue, headache, fever, muscle or joint pain, shortness of breath) were the most important clinical predictors at admission. Age, overall presence of symptoms, shortness of breath, and hypertension were found to be key predictors for PE using our extreme gradient boosted model. This analysis based on the, until now, largest global dataset for this set of problems can inform hospital prioritisation policy and guide long term clinical research and decision-making for COVID-19 patients globally. Our machine learning model developed from an international cohort can serve to better regulate hospital risk prioritisation of at-risk patients. [ABSTRACT FROM AUTHOR]

The rapid progress of industrial development, urbanization, and traffic has caused air quality degradation that negatively affects human health and environmental sustainability, especially in developed countries. However, due to the limited number of sensors available, the air quality index at many locations is not monitored. Therefore, many research, including statistical and machine learning approaches, have been proposed to tackle the problem of estimating air quality value at an arbitrary location. Most of the existing research perform interpolation process based on traditional techniques that leverage distance information. In this work, we propose a novel deep-learning-based model for air quality value estimation. This approach follows the encoder–decoder paradigm, with the encoder and decoder trained separately using different training mechanisms. In the encoder component, we proposed a new self-supervised graph representation learning approach for spatio-temporal data. For the decoder component, we designed a deep interpolation layer that employs two attention mechanisms and a fully connected layer using air quality data at known stations, distance information, and meteorology information at the target point to predict air quality at arbitrary locations. The experimental results demonstrate significant improvements in estimation accuracy achieved by our proposed model compared to state-of-the-art approaches. For the MAE indicator, our model enhances the estimation accuracy from 4.93% to 34.88% on the UK dataset, and from 6.89% to 31.94% regarding the Beijing dataset. In terms of the RMSE, the average improvements of our method on the two datasets are 13.33% and 14.37%, respectively. The statistics for MAPE are 36.05% and 13.25%, while for MDAPE, they are 24.48% and 36.33%, respectively. Furthermore, the value of R 2 score attained by our proposed model also shows considerable improvement, with increases of 5.39% and 32.58% compared to that of comparison benchmarks. Our source code and data are available at https://github.com/duclong1009/Unsupervised-Air-Quality-Estimation. [ABSTRACT FROM AUTHOR]

Across the globe, governments are developing policies and strategies to reduce carbon emissions to address climate change. Monitoring the impact of governments' carbon reduction policies can significantly enhance our ability to combat climate change and meet emissions reduction targets. One promising area in this regard is the role of artificial intelligence (AI) in carbon reduction policy and strategy monitoring. While researchers have explored applications of AI on data from various sources, including sensors, satellites, and social media, to identify areas for carbon emissions reduction, AI applications in tracking the effect of governments' carbon reduction plans have been limited. This study presents an AI framework based on long short-term memory (LSTM) and statistical process control (SPC) for the monitoring of variations in carbon emissions, using UK annual CO2 emission (per capita) data, covering a period between 1750 and 2021. This paper used LSTM to develop a surrogate model for the UK's carbon emissions characteristics and behaviours. As observed in our experiments, LSTM has better predictive abilities than ARIMA, Exponential Smoothing and feedforward artificial neural networks (ANN) in predicting CO2 emissions on a yearly prediction horizon. Using the deviation of the recorded emission data from the surrogate process, the variations and trends in these behaviours are then analysed using SPC, specifically Shewhart individual/moving range control charts. The result shows several assignable variations between the mid-1990s and 2021, which correlate with some notable UK government commitments to lower carbon emissions within this period. The framework presented in this paper can help identify periods of significant deviations from a country's normal CO2 emissions, which can potentially result from the government's carbon reduction policies or activities that can alter the amount of CO2 emissions. [ABSTRACT FROM AUTHOR]

Background: Cancer is a leading cause of morbidity and mortality worldwide, and better understanding of the risk factors could enhance prevention. Methods: We conducted a hypothesis‐free analysis combining machine learning and statistical approaches to identify cancer risk factors from 2828 potential predictors captured at baseline. There were 459,169 UK Biobank participants free from cancer at baseline and 48,671 new cancer cases during the 10‐year follow‐up. Logistic regression models adjusted for age, sex, ethnicity, education, material deprivation, smoking, alcohol intake, body mass index and skin colour (as a proxy for sun sensitivity) were used for obtaining adjusted odds ratios, with continuous predictors presented using quintiles (Q). Results: In addition to smoking, older age and male sex, positively associating features included several anthropometric characteristics, whole body water mass, pulse, hypertension and biomarkers such as urinary microalbumin (Q5 vs. Q1 OR 1.16, 95% CI = 1.13–1.19), C‐reactive protein (Q5 vs. Q1 OR 1.20, 95% CI = 1.16–1.24) and red blood cell distribution width (Q5 vs. Q1 OR 1.18, 95% CI = 1.14–1.21), among others. High‐density lipoprotein cholesterol (Q5 vs. Q1 OR 0.84, 95% CI = 0.81–0.87) and albumin (Q5 vs. Q1 OR 0.84, 95% CI = 0.81–0.87) were inversely associated with cancer. In sex‐stratified analyses, higher testosterone increased the risk in females but not in males (Q5 vs. Q1 ORfemales 1.23, 95% CI = 1.17–1.30). Phosphate was associated with a lower risk in females but a higher risk in males (Q5 vs. Q1 ORfemales 0.94, 95% CI = 0.90–0.99 vs. ORmales 1.09, 95% CI 1.04–1.15). Conclusions: This hypothesis‐free analysis suggests personal characteristics, metabolic biomarkers, physical measures and smoking as important predictors of cancer risk, with further studies needed to confirm causality and clinical relevance. [ABSTRACT FROM AUTHOR]

Quantifying an individual's risk for common diseases is an important goal of precision health. The polygenic risk score (PRS), which aggregates multiple risk alleles of candidate diseases, has emerged as a standard approach for identifying high-risk individuals. Although several studies have been performed to benchmark the PRS calculation tools and assess their potential to guide future clinical applications, some issues remain to be further investigated, such as lacking (i) various simulated data with different genetic effects; (ii) evaluation of machine learning models and (iii) evaluation on multiple ancestries studies. In this study, we systematically validated and compared 13 statistical methods, 5 machine learning models and 2 ensemble models using simulated data with additive and genetic interaction models, 22 common diseases with internal training sets, 4 common diseases with external summary statistics and 3 common diseases for trans -ancestry studies in UK Biobank. The statistical methods were better in simulated data from additive models and machine learning models have edges for data that include genetic interactions. Ensemble models are generally the best choice by integrating various statistical methods. LDpred2 outperformed the other standalone tools, whereas PRS-CS, lassosum and DBSLMM showed comparable performance. We also identified that disease heritability strongly affected the predictive performance of all methods. Both the number and effect sizes of risk SNPs are important; and sample size strongly influences the performance of all methods. For the trans -ancestry studies, we found that the performance of most methods became worse when training and testing sets were from different populations. [ABSTRACT FROM AUTHOR]

The identification and understanding of gene-environment interactions can provide insights into the pathways and mechanisms underlying complex diseases. However, testing for gene-environment interaction remains a challenge since a.) statistical power is often limited and b.) modeling of environmental effects is nontrivial and such model misspecifications can lead to false positive interaction findings. To address the lack of statistical power, recent methods aim to identify interactions on an aggregated level using, for example, polygenic risk scores. While this strategy can increase the power to detect interactions, identifying contributing genes and pathways is difficult based on these relatively global results. Here, we propose RITSS (Robust Interaction Testing using Sample Splitting), a gene-environment interaction testing framework for quantitative traits that is based on sample splitting and robust test statistics. RITSS can incorporate sets of genetic variants and/or multiple environmental factors. Based on the user's choice of statistical/machine learning approaches, a screening step selects and combines potential interactions into scores with improved interpretability. In the testing step, the application of robust statistics minimizes the susceptibility to main effect misspecifications. Using extensive simulation studies, we demonstrate that RITSS controls the type 1 error rate in a wide range of scenarios, and we show how the screening strategy influences statistical power. In an application to lung function phenotypes and human height in the UK Biobank, RITSS identified highly significant interactions based on subcomponents of genetic risk scores. While the contributing single variant interaction signals are weak, our results indicate interaction patterns that result in strong aggregated effects, providing potential insights into underlying gene-environment interaction mechanisms. Author summary: The understanding of gene-environment interactions provides potential insights into the pathways and mechanisms underlying complex diseases, but they are hard to detect since effect sizes are expected to be small. To facilitate the detection of such interactions in quantitative traits, we propose a robust and flexible approach called RITSS that can incorporate sets of genetic variants and multiple environmental factors. RITSS can utilize any suitable machine/statistical learning approach to screen for interactions and rigorously tests these aggregated signals using sample splitting and robust test statistics. We demonstrate the validity and power of our approach in extensive simulation studies. Furthermore, in an application to lung function and height data in the UK Biobank, RITSS discovers highly significant interactions based on subcomponents of genetic risk scores. [ABSTRACT FROM AUTHOR]

We present a simple and efficient hypothesis-free machine learning pipeline for risk factor discovery that accounts for non-linearity and interaction in large biomedical databases with minimal variable pre-processing. In this study, mortality models were built using gradient boosting decision trees (GBDT) and important predictors were identified using a Shapley values-based feature attribution method, SHAP values. Cox models controlled for false discovery rate were used for confounder adjustment, interpretability, and further validation. The pipeline was tested using information from 502,506 UK Biobank participants, aged 37–73 years at recruitment and followed over seven years for mortality registrations. From the 11,639 predictors included in GBDT, 193 potential risk factors had SHAP values ≥ 0.05, passed the correlation test, and were selected for further modelling. Of the total variable importance summed up, 60% was directly health related, and baseline characteristics, sociodemographics, and lifestyle factors each contributed about 10%. Cox models adjusted for baseline characteristics, showed evidence for an association with mortality for 166 out of the 193 predictors. These included mostly well-known risk factors (e.g., age, sex, ethnicity, education, material deprivation, smoking, physical activity, self-rated health, BMI, and many disease outcomes). For 19 predictors we saw evidence for an association in the unadjusted but not adjusted analyses, suggesting bias by confounding. Our GBDT-SHAP pipeline was able to identify relevant predictors 'hidden' within thousands of variables, providing an efficient and pragmatic solution for the first stage of hypothesis free risk factor identification. [ABSTRACT FROM AUTHOR]

Dynamic prediction models provide predicted survival probabilities that can be updated over time for an individual as new measurements become available. Two techniques for dynamic survival prediction with longitudinal data dominate the statistical literature: joint modelling and landmarking. There is substantial interest in the use of machine learning methods for prediction; however, their use in the context of dynamic survival prediction has been limited. We show how landmarking can be combined with a machine learning ensemble—the Super Learner. The ensemble combines predictions from different machine learning and statistical algorithms with the goal of achieving improved performance. The proposed approach exploits discrete time survival analysis techniques to enable the use of machine learning algorithms for binary outcomes. We discuss practical and statistical considerations involved in implementing the ensemble. The methods are illustrated and compared using longitudinal data from the UK Cystic Fibrosis Registry. Standard landmarking and the landmark Super Learner approach resulted in similar cross‐validated predictive performance, in this case, outperforming joint modelling. [ABSTRACT FROM AUTHOR]

All data processing, feature extraction, machine learning, and analysis code will be available at https://github.com/activityMonitoring." should read: "Upon publication, the summary variables that we have constructed are available as a part of the UK Biobank dataset at https://biobank.ndph.ox.ac.uk/ukb/dset.cgi?id=2242. All data processing, feature extraction, machine learning, and analysis code are available at https://github.com/activityMonitoring. [Extracted from the article]

• Both spatial and temporal transferability of real-time crash risk models are assessed. • Wasserstein Generative Adversarial Network (WGAN) is introduced to address the extreme data imbalance issue. • Deep Neural Networks (DNN) and Transfer Learning (TL) are applied for predicting crashes. • Transfer learning improves transferability where direct transfer is not feasible. Real-time crash prediction is a heavily studied area given their potential applications in proactive traffic safety management in which a plethora of statistical and machine learning (ML) models have been developed to predict traffic crashes in real-time. However, one of the fundamental issues relating to the application of these models is spatio-temporal transferability. The present paper attempts to address this gap of knowledge by combining Generative Adversarial Network (GAN) and transfer learning to examine the transferability of real-time crash prediction models under an extremely imbalanced data setting. Initially, a baseline model was developed using Deep Neural Network (DNN) with crash and microscopic traffic data collected from M1 Motorway in the UK in 2017. The dataset utilised in the baseline model is naturally imbalanced with 257 crash cases and 16,359,163 non-crash cases. To overcome data imbalance issue, Wasserstein GAN (WGAN) was utilised to generate synthetic crash data. Non-crash data were randomly undersampled due to computational limitations. The calibrated model was then applied to predict traffic crashes for five other datasets obtained from M1 (2018), M4 (2017 & 2018 separately) and M6 Motorway (2017 & 2018 separately) by using transfer learning. Model transferability was compared with standalone models and direct transfer from the baseline model. The study revealed that direct transfer is not feasible. However, models become transferable temporally, spatially, and spatio-temporally if transfer learning is applied. The predictability of the transferred models outperformed existing studies by achieving high Area Under Curve (AUC) values ranging between 0.69 and 0.95. The best transferred model can predict nearly 95% crashes with only a 5% false alarm rate by tuning thresholds. Furthermore, the performances of transferred models are on par with or better than the standalone model. The findings of this study proves that transfer learning can improve model transferability under extremely imbalanced settings which helps traffic engineers in developing highly transferable models in future. [ABSTRACT FROM AUTHOR]

• Statistical modelling of brain-heart interaction by including prior cardiac physiological knowledge. • Biophysical modelling constrained by a model of joint heart-brain variability. • The model identifies physiologically plausible mechanisms and significant differences associated to clinical conditions. • We present the first study on patient-specific modelling relating cardiac function and brain damage. • This study is based on the largest presented cohort of personalised subjects to date (3445 individuals). [Display omitted] Initiatives such as the UK Biobank provide joint cardiac and brain imaging information for thousands of individuals, representing a unique opportunity to study the relationship between heart and brain. Most of research on large multimodal databases has been focusing on studying the associations among the available measurements by means of univariate and multivariate association models. However, these approaches do not provide insights about the underlying mechanisms and are often hampered by the lack of prior knowledge on the physiological relationships between measurements. For instance, important indices of the cardiovascular function, such as cardiac contractility, cannot be measured in-vivo. While these non-observable parameters can be estimated by means of biophysical models, their personalisation is generally an ill-posed problem, often lacking critical data and only applied to small datasets. Therefore, to jointly study brain and heart, we propose an approach in which the parameter personalisation of a lumped cardiovascular model is constrained by the statistical relationships observed between model parameters and brain-volumetric indices extracted from imaging, i.e. ventricles or white matter hyperintensities volumes, and clinical information such as age or body surface area. We explored the plausibility of the learnt relationships by inferring the model parameters conditioned on the absence of part of the target clinical features, applying this framework in a cohort of more than 3 000 subjects and in a pathological subgroup of 59 subjects diagnosed with atrial fibrillation. Our results demonstrate the impact of such external features in the cardiovascular model personalisation by learning more informative parameter-space constraints. Moreover, physiologically plausible mechanisms are captured through these personalised models as well as significant differences associated to specific clinical conditions. [ABSTRACT FROM AUTHOR]

Background: In recent years, the availability of publicly available data related to public health has significantly increased. These data have substantial potential to develop public health policy; however, this requires meaningful and insightful analysis. Our aim is to demonstrate how data analysis techniques can be used to address the issues of data reduction, prediction and explanation using online available public health data, in order to provide a sound basis for informing public health policy.Methods: Observational suicide prevention data were analysed from an existing online United Kingdom national public health database. Multi-collinearity analysis and principal-component analysis were used to reduce correlated data, followed by regression analyses for prediction and explanation of suicide.Results: Multi-collinearity analysis was effective in reducing the indicator set of predictors by 30% and principal component analysis further reduced the set by 86%. Regression for prediction identified four significant indicator predictors of suicide behaviour (emergency hospital admissions for intentional self-harm, children leaving care, statutory homelessness and self-reported well-being/low happiness) and two main component predictors (relatedness dysfunction, and behavioural problems and mental illness). Regression for explanation identified significant moderation of a well-being predictor (low happiness) of suicide behaviour by a social factor (living alone), thereby supporting existing theory and providing insight beyond the results of regression for prediction. Two independent predictors capturing relatedness needs in social care service delivery were also identified.Conclusions: We demonstrate the effectiveness of regression techniques in the analysis of online public health data. Regression analysis for prediction and explanation can both be appropriate for public health data analysis for a better understanding of public health outcomes. It is therefore essential to clarify the aim of the analysis (prediction accuracy or theory development) as a basis for choosing the most appropriate model. We apply these techniques to the analysis of suicide data; however, we argue that the analysis presented in this study should be applied to datasets across public health in order to improve the quality of health policy recommendations. [ABSTRACT FROM AUTHOR]