RESUMO
BACKGROUND AND AIMS: Effervescent formulations of paracetamol containing sodium bicarbonate have been reported to associate with increased blood pressure and a higher risk of cardiovascular diseases and all-cause mortality. Given the major implications of these findings, the reported associations were re-examined. METHODS: Using linked electronic health records data, a cohort of 475 442 UK individuals with at least one prescription of paracetamol, aged between 60 and 90 years, was identified. Outcomes in patients taking sodium-based paracetamol were compared with those taking non-sodium-based formulations of the same. Using a deep learning approach, associations with systolic blood pressure (SBP), major cardiovascular events (myocardial infarction, heart failure, and stroke), and all-cause mortality within 1 year after baseline were investigated. RESULTS: A total of 460 980 and 14 462 patients were identified for the non-sodium-based and sodium-based paracetamol exposure groups, respectively (mean age: 74 years; 64% women). Analysis revealed no difference in SBP [mean difference -0.04 mmHg (95% confidence interval -0.51, 0.43)] and no association with major cardiovascular events [relative risk (RR) 1.03 (0.91, 1.16)]. Sodium-based paracetamol showed a positive association with all-cause mortality [RR 1.46 (1.40, 1.52)]. However, after further accounting of other sources of residual confounding, the observed association attenuated towards the null [RR 1.08 (1.01, 1.16)]. Exploratory analyses revealed dysphagia and related conditions as major sources of uncontrolled confounding by indication for this association. CONCLUSIONS: This study does not support previous suggestions of increased SBP and an elevated risk of cardiovascular events from short-term use of sodium bicarbonate paracetamol in routine clinical practice.
Assuntos
Doenças Cardiovasculares , Hipertensão , Infarto do Miocárdio , Humanos , Feminino , Pessoa de Meia-Idade , Idoso , Idoso de 80 Anos ou mais , Masculino , Pressão Sanguínea , Hipertensão/complicações , Acetaminofen/efeitos adversos , Anti-Hipertensivos/uso terapêutico , Sódio , Bicarbonato de Sódio/farmacologia , Infarto do Miocárdio/complicaçõesRESUMO
Dermal water content is an important biophysical parameter in preserving skin integrity and preventing skin damage. Traditional electrical-based and open-chamber evaporimeters have several well-known limitations. In particular, such devices are costly, sizeable, and only provide arbitrary outputs. They also do not permit continuous and non-invasive monitoring of dermal water content, which can be beneficial for various consumer, clinical, and cosmetic purposes. We report here on the design and development of a digital multi-wavelength optical sensor that performs continuous and non-invasive measurement of dermal water content. In silico investigation on porcine skin was carried out using the Monte Carlo modeling strategy to evaluate the feasibility and characterize the sensor. Subsequently, an in vitro experiment was carried out to evaluate the performance of the sensor and benchmark its accuracy against a high-end, broad band spectrophotometer. Reference measurements were made against gravimetric analysis. The results demonstrate that the developed sensor can deliver accurate, continuous, and non-invasive measurement of skin hydration through measurement of dermal water content. Remarkably, the novel design of the sensor exceeded the performance of the high-end spectrophotometer due to the important denoising effects of temporal averaging. The authors believe, in addition to wellbeing and skin health monitoring, the designed sensor can particularly facilitate disease management in patients presenting diabetes mellitus, hypothyroidism, malnutrition, and atopic dermatitis.
Assuntos
Pele , Água , Animais , Biofísica , Simulação por Computador , Humanos , Método de Monte Carlo , SuínosRESUMO
Near Infrared (800-2500 nm) spectroscopy has been extensively used in biomedical applications, as it offers rapid, in vivo, bed-side monitoring of important haemodynamic parameters, which is especially important in critical care settings. However, the choice of NIR spectrometer needs to be investigated for biomedical applications, as both the dual beam dispersive spectrophotomer and the FTNIR spectrometer have their own advantages and disadvantages. In this study, predictive analysis of lactate concentrations in whole blood were undertaken using multivariate techniques on spectra obtained from the two spectrometer types simultaneously and results were compared. Results showed significant improvement in predicting analyte concentration when analysis was performed on full range spectral data. This is in comparison to analysis of limited spectral regions or lactate signature peaks, which yielded poorer prediction models. Furthermore, for the same region, FTNIR showed 10% better predictive capability than the dual beam dispersive NIR spectrometer.
Assuntos
Ácido Láctico , Espectroscopia de Luz Próxima ao Infravermelho , Espectroscopia de Infravermelho com Transformada de FourierRESUMO
Uninterrupted monitoring of serum lactate levels is a prerequisite in the critical care of patients prone to sepsis, cardiogenic shock, cardiac arrest, or severe lung disease. Yet there exists no device to continuously measure blood lactate in clinical practice. Optical spectroscopy together with multivariate analysis is proposed as a viable noninvasive tool for estimation of lactate in blood. As an initial step towards this goal, we inspected the plausibility of predicting the concentration of sodium lactate (NaLac) from the UV/visible, near-infrared (NIR), and mid-infrared (MIR) spectra of 37 isotonic phosphate-buffered saline (PBS) samples containing NaLac ranging from 0 to 20 mmol/L. UV/visible (300-800 nm) and NIR (800-2600 nm) spectra of PBS samples were collected using the PerkinElmer Lambda 1050 dual-beam spectrophotometer, while MIR (4000-500 cm-1) spectra were collected using the Spectrum two FTIR spectrometer. Absorption bands in the spectra of all three regions were identified and functional groups were assigned. The concentration of lactate in samples was predicted using the Partial Least-Squares (PLS) regression analysis and leave-one-out cross-validation. The regression analysis showed a correlation coefficient (R2) of 0.926, 0.977, and 0.992 for UV/visible, NIR, and MIR spectra, respectively, between the predicted and reference samples. The RMSECV of UV/visible, NIR, and MIR spectra was 1.59, 0.89, and 0.49 mmol/L, respectively. The results indicate that optical spectroscopy together with multivariate models can achieve a superior technique in assessing lactate concentrations.
Assuntos
Ácido Láctico/sangue , Sepse/diagnóstico , Espectroscopia de Luz Próxima ao Infravermelho , Humanos , Análise dos Mínimos Quadrados , Análise MultivariadaRESUMO
Quantification of lactate/lactic acid in critical care environments is essential as lactate serves as an important biochemical marker for the adequacy of the haemodynamic circulation in shock and of cell respiration at the onset of sepsis/septic shock. Hence, in this study, ATR-FTIR was explored as a potential tool for lactate measurement, as the current techniques depend on sample preparation and fails to provide rapid response. Moreover, the effects of pH on PBS samples (7.4, 7, 6.5 and 6) and change in solution conditions (PBS to whole blood) on spectral features were also investigated. A total 189 spectra from five sets of lactate containing media were obtained. Results suggests that lactate could be measured with more than 90% accuracy in the wavenumber range of 1500-600 cm-1. The findings of this study further suggest that there exist no effects of change in pH or media, when estimating lactate concentration changes in this range of the Mid-IR spectral region.
Assuntos
Análise Química do Sangue/métodos , Ácido Láctico/sangue , Espectroscopia de Infravermelho com Transformada de Fourier , Métodos Analíticos de Preparação de Amostras , Animais , Concentração de Íons de Hidrogênio , Ovinos , Fatores de TempoRESUMO
The proliferation of street view images (SVIs) and the constant advancements in deep learning techniques have enabled urban analysts to extract and evaluate urban perceptions from large-scale urban streetscapes. However, many existing analytical frameworks have been found to lack interpretability due to their end-to-end structure and "black-box" nature, thereby limiting their value as a planning support tool. In this context, we propose a five-step machine learning framework for extracting neighborhood-level urban perceptions from panoramic SVIs, specifically emphasizing feature and result interpretability. By utilizing the MIT Place Pulse data, the developed framework can systematically extract six dimensions of urban perceptions from the given panoramas, including perceptions of wealth, boredom, depression, beauty, safety, and liveliness. The practical utility of this framework is demonstrated through its deployment in Inner London, where it was used to visualize urban perceptions at the Output Area (OA) level and to verify against real-world crime rate.
RESUMO
Background: The quality of evidence regarding the associations between road traffic noise and hypertension is low due to the limitations of cross-sectional study design, and the role of air pollution remains to be further clarified. Objectives: The purpose of this study was to evaluate the associations of long-term road traffic noise exposure with incident primary hypertension; we conducted a prospective population-based analysis in UK Biobank. Methods: Road traffic noise was estimated at baseline residential address using the common noise assessment method model. Incident hypertension was ascertained through linkage with medical records. Cox proportional hazard models were used to estimate hazard ratios (HRs) for association in an analytical sample size of over 240,000 participants free of hypertension at baseline, adjusting for covariates determined via directed acyclic graph. Results: During a median of 8.1 years follow-up, 21,140 incident primary hypertension (International Classification of Diseases-10th Revision [ICD-10]: I10) were ascertained. The HR for a 10 dB[A] increment in mean weighted average 24-hour road traffic noise level (L den ) exposure was 1.07 (95% CI: 1.02-1.13). A dose-response relationship was found, with HR of 1.13 (95% CI: 1.03-1.25) for L den >65 dB[A] vs ≤55 dB[A] (P for trend <0.05). The associations were all robust to adjustment for fine particles (PM2.5) and nitrogen dioxide (NO2). Furthermore, high exposure to both road traffic noise and air pollution was associated with the highest hypertension risk. Conclusions: Long-term exposure to road traffic noise was associated with increased incidence of primary hypertension, and the effect estimates were stronger in presence of higher air pollution.
RESUMO
OBJECTIVE: In individuals with complex underlying health problems, the association between systolic blood pressure (SBP) and cardiovascular disease is less well recognised. The association between SBP and risk of cardiovascular events in patients with chronic obstructive pulmonary disease (COPD) was investigated. METHODS AND ANALYSIS: In this cohort study, 39 602 individuals with a diagnosis of COPD aged 55-90 years between 1990 and 2009 were identified from validated electronic health records (EHR) in the UK. The association between SBP and risk of cardiovascular end points (composite of ischaemic heart disease, heart failure, stroke and cardiovascular death) was analysed using a deep learning approach. RESULTS: In the selected cohort (46.5% women, median age 69 years), 10 987 cardiovascular events were observed over a median follow-up period of 3.9 years. The association between SBP and risk of cardiovascular end points was found to be monotonic; the lowest SBP exposure group of <120 mm Hg presented nadir of risk. With respect to reference SBP (between 120 and 129 mm Hg), adjusted risk ratios for the primary outcome were 0.99 (95% CI 0.93 to 1.05) for SBP of <120 mm Hg, 1.02 (0.97 to 1.07) for SBP between 130 and 139 mm Hg, 1.07 (1.01 to 1.12) for SBP between 140 and 149 mm Hg, 1.11 (1.05 to 1.17) for SBP between 150 and 159 mm Hg and 1.16 (1.10 to 1.22) for SBP ≥160 mm Hg. CONCLUSION: Using deep learning for modelling EHR, we identified a monotonic association between SBP and risk of cardiovascular events in patients with COPD.
Assuntos
Doenças Cardiovasculares , Hipertensão , Doença Pulmonar Obstrutiva Crônica , Humanos , Feminino , Idoso , Masculino , Pressão Sanguínea/fisiologia , Doenças Cardiovasculares/diagnóstico , Doenças Cardiovasculares/epidemiologia , Doenças Cardiovasculares/etiologia , Hipertensão/diagnóstico , Estudos de Coortes , Fatores de Risco , Doença Pulmonar Obstrutiva Crônica/complicações , Doença Pulmonar Obstrutiva Crônica/diagnóstico , Doença Pulmonar Obstrutiva Crônica/epidemiologia , Fatores de Risco de Doenças Cardíacas , Anti-Hipertensivos/uso terapêuticoRESUMO
Aims: A diverse set of factors influence cardiovascular diseases (CVDs), but a systematic investigation of the interplay between these determinants and the contribution of each to CVD incidence prediction is largely missing from the literature. In this study, we leverage one of the most comprehensive biobanks worldwide, the UK Biobank, to investigate the contribution of different risk factor categories to more accurate incidence predictions in the overall population, by sex, different age groups, and ethnicity. Methods and results: The investigated categories include the history of medical events, behavioural factors, socioeconomic factors, environmental factors, and measurements. We included data from a cohort of 405 257 participants aged 37-73 years and trained various machine learning and deep learning models on different subsets of risk factors to predict CVD incidence. Each of the models was trained on the complete set of predictors and subsets where each category was excluded. The results were benchmarked against QRISK3. The findings highlight that (i) leveraging a more comprehensive medical history substantially improves model performance. Relative to QRISK3, the best performing models improved the discrimination by 3.78% and improved precision by 1.80%. (ii) Both model- and data-centric approaches are necessary to improve predictive performance. The benefits of using a comprehensive history of diseases were far more pronounced when a neural sequence model, BEHRT, was used. This highlights the importance of the temporality of medical events that existing clinical risk models fail to capture. (iii) Besides the history of diseases, socioeconomic factors and measurements had small but significant independent contributions to the predictive performance. Conclusion: These findings emphasize the need for considering broad determinants and novel modelling approaches to enhance CVD incidence prediction.
RESUMO
BACKGROUND: Whether the association between systolic blood pressure (SBP) and risk of cardiovascular disease is monotonic or whether there is a nadir of optimal blood pressure remains controversial. We investigated the association between SBP and cardiovascular events in patients with diabetes across the full spectrum of SBP. METHODS: A cohort of 49 000 individuals with diabetes aged 50 to 90 years between 1990 and 2005 was identified from linked electronic health records in the United Kingdom. Associations between SBP and cardiovascular outcomes (ischemic heart disease, heart failure, stroke, and cardiovascular death) were analyzed using a deep learning approach. RESULTS: Over a median follow-up of 7.3 years, 16 378 cardiovascular events were observed. The relationship between SBP and cardiovascular events followed a monotonic pattern, with the group with the lowest baseline SBP of <120 mm Hg exhibiting the lowest risk of cardiovascular events. In comparison to the reference group with the lowest SBP (<120 mm Hg), the adjusted risk ratio for cardiovascular disease was 1.03 (95% CI, 0.97-1.10) for SBP between 120 and 129 mm Hg, 1.05 (0.99-1.11) for SBP between 130 and 139 mm Hg, 1.08 (1.01-1.15) for SBP between 140 and 149 mm Hg, 1.12 (1.03-1.20) for SBP between 150 and 159 mm Hg, and 1.19 (1.09-1.28) for SBP ≥160 mm Hg. CONCLUSIONS: Using deep learning modeling, we found a monotonic relationship between SBP and risk of cardiovascular outcomes in patients with diabetes, without evidence of a J-shaped relationship.
Assuntos
Doenças Cardiovasculares , Diabetes Mellitus , Hipertensão , Humanos , Pressão Sanguínea , Doenças Cardiovasculares/epidemiologia , Doenças Cardiovasculares/etiologia , Hipertensão/epidemiologia , Estudos Prospectivos , Fatores de Risco , Diabetes Mellitus/epidemiologia , Fatores de Risco de Doenças CardíacasRESUMO
Electronic health records (EHR) represent a holistic overview of patients' trajectories. Their increasing availability has fueled new hopes to leverage them and develop accurate risk prediction models for a wide range of diseases. Given the complex interrelationships of medical records and patient outcomes, deep learning models have shown clear merits in achieving this goal. However, a key limitation of current study remains their capacity in processing long sequences, and long sequence modelling and its application in the context of healthcare and EHR remains unexplored. Capturing the whole history of medical encounters is expected to lead to more accurate predictions, but the inclusion of records collected for decades and from multiple resources can inevitably exceed the receptive field of the most existing deep learning architectures. This can result in missing crucial, long-term dependencies. To address this gap, we present Hi-BEHRT, a hierarchical Transformer-based model that can significantly expand the receptive field of Transformers and extract associations from much longer sequences. Using a multimodal large-scale linked longitudinal EHR, the Hi-BEHRT exceeds the state-of-the-art deep learning models 1% to 5% for area under the receiver operating characteristic (AUROC) curve and 1% to 8% for area under the precision recall (AUPRC) curve on average, and 2% to 8% (AUROC) and 2% to 11% (AUPRC) for patients with long medical history for 5-year heart failure, diabetes, chronic kidney disease, and stroke risk prediction. Additionally, because pretraining for hierarchical Transformer is not well-established, we provide an effective end-to-end contrastive pre-training strategy for Hi-BEHRT using EHR, improving its transferability on predicting clinical events with relatively small training dataset.
Assuntos
Registros Eletrônicos de Saúde , Insuficiência Cardíaca , Humanos , Área Sob a Curva , Fontes de Energia Elétrica , Curva ROCRESUMO
Diabetes is a heterogenous, multimorbid disorder with a large variation in manifestations, trajectories, and outcomes. The aim of this study is to validate a novel machine learning method for the phenotyping of diabetes in the context of comorbidities. Data from 9967 multimorbid patients with a new diagnosis of diabetes were extracted from Clinical Practice Research Datalink. First, using BEHRT (a transformer-based deep learning architecture), the embeddings corresponding to diabetes were learned. Next, topological data analysis (TDA) was carried out to test how different areas in high-dimensional manifold correspond to different risk profiles. The following endpoints were considered when profiling risk trajectories: major adverse cardiovascular events (MACE), coronary artery disease (CAD), stroke (CVA), heart failure (HF), renal failure (RF), diabetic neuropathy, peripheral arterial disease, reduced visual acuity and all-cause mortality. Kaplan Meier curves were plotted for each derived phenotype. Finally, we tested the performance of an established risk prediction model (QRISK) by adding TDA-derived features. We identified four subgroups of patients with diabetes and divergent comorbidity patterns differing in their risk of future cardiovascular, renal, and other microvascular outcomes. Phenotype 1 (young with chronic inflammatory conditions) and phenotype 2 (young with CAD) included relatively younger patients with diabetes compared to phenotypes 3 (older with hypertension and renal disease) and 4 (older with previous CVA), and those subgroups had a higher frequency of pre-existing cardio-renal diseases. Within ten years of follow-up, 2592 patients (26%) experienced MACE, 2515 patients (25%) died, and 2020 patients (20%) suffered RF. QRISK3 model's AUC was augmented from 67.26% (CI 67.25-67.28%) to 67.67% (CI 67.66-67.69%) by adding specific TDA-derived phenotype and the distances to both extremities of the TDA graph improving its performance in the prediction of CV outcomes. We confirmed the importance of accounting for multimorbidity when risk stratifying heterogenous cohort of patients with new diagnosis of diabetes. Our unsupervised machine learning method improved the prediction of clinical outcomes.
Assuntos
Diabetes Mellitus , Aprendizado de Máquina não Supervisionado , Humanos , Diabetes Mellitus/epidemiologia , Comorbidade , Análise de Dados , Doenças Cardiovasculares/epidemiologia , Medição de Risco , Nefropatias/epidemiologia , Masculino , Feminino , Pessoa de Meia-Idade , Idoso , FenótipoRESUMO
Multicollinearity refers to the presence of collinearity between multiple variables and renders the results of statistical inference erroneous (Type II error). This is particularly important in environmental health research where multicollinearity can hinder inference. To address this, correlated variables are often excluded from the analysis, limiting the discovery of new associations. An alternative approach to address this problem is the use of principal component analysis. This method, combines and projects a group of correlated variables onto a new orthogonal space. While this resolves the multicollinearity problem, it poses another challenge in relation to interpretability of results. Standard hypothesis testing methods can be used to evaluate the association of projected predictors, called principal components, with the outcomes of interest, however, there is no established way to trace the significance of principal components back to individual variables. To address this problem, we investigated the use of sparse principal component analysis which enforces a parsimonious projection. We hypothesise that this parsimony could facilitate the interpretability of findings. To this end, we investigated the association of 20 environmental predictors with all-cause mortality adjusting for demographic, socioeconomic, physiological, and behavioural factors. The study was conducted in a cohort of 379,690 individuals in the UK. During an average follow-up of 8.05 years (3,055,166 total person-years), 14,996 deaths were observed. We used Cox regression models to estimate the hazard ratio (HR) and 95% confidence intervals (CI). The Cox models were fitted to the standardised environmental predictors (a) without any transformation (b) transformed with PCA, and (c) transformed with SPCA. The comparison of findings underlined the potential of SPCA for conducting inference in scenarios where multicollinearity can increase the risk of Type II error. Our analysis unravelled a significant association between average noise pollution and increased risk of all-cause mortality. Specifically, those in the upper deciles of noise exposure have between 5 and 10% increased risk of all-cause mortality compared to the lowest decile.
Assuntos
Bancos de Espécimes Biológicos , Exposição Ambiental , Exposição Ambiental/efeitos adversos , Saúde Ambiental , Humanos , Análise de Componente Principal , Reino Unido/epidemiologiaRESUMO
Aims: Deep learning has dominated predictive modelling across different fields, but in medicine it has been met with mixed reception. In clinical practice, simple, statistical models and risk scores continue to inform cardiovascular disease risk predictions. This is due in part to the knowledge gap about how deep learning models perform in practice when they are subject to dynamic data shifts; a key criterion that common internal validation procedures do not address. We evaluated the performance of a novel deep learning model, BEHRT, under data shifts and compared it with several ML-based and established risk models. Methods and results: Using linked electronic health records of 1.1 million patients across England aged at least 35 years between 1985 and 2015, we replicated three established statistical models for predicting 5-year risk of incident heart failure, stroke, and coronary heart disease. The results were compared with a widely accepted machine learning model (random forests), and a novel deep learning model (BEHRT). In addition to internal validation, we investigated how data shifts affect model discrimination and calibration. To this end, we tested the models on cohorts from (i) distinct geographical regions; (ii) different periods. Using internal validation, the deep learning models substantially outperformed the best statistical models by 6%, 8%, and 11% in heart failure, stroke, and coronary heart disease, respectively, in terms of the area under the receiver operating characteristic curve. Conclusion: The performance of all models declined as a result of data shifts; despite this, the deep learning models maintained the best performance in all risk prediction tasks. Updating the model with the latest information can improve discrimination but if the prior distribution changes, the model may remain miscalibrated.
RESUMO
Observational causal inference is useful for decision-making in medicine when randomized clinical trials (RCTs) are infeasible or nongeneralizable. However, traditional approaches do not always deliver unconfounded causal conclusions in practice. The rise of "doubly robust" nonparametric tools coupled with the growth of deep learning for capturing rich representations of multimodal data offers a unique opportunity to develop and test such models for causal inference on comprehensive electronic health records (EHRs). In this article, we investigate causal modeling of an RCT-established causal association: the effect of classes of antihypertensive on incident cancer risk. We develop a transformer-based model, targeted bidirectional EHR transformer (T-BEHRT) coupled with doubly robust estimation to estimate average risk ratio (RR). We compare our model to benchmark statistical and deep learning models for causal inference in multiple experiments on semi-synthetic derivations of our dataset with various types and intensities of confounding. In order to further test the reliability of our approach, we test our model on situations of limited data. We find that our model provides more accurate estimates of relative risk least sum absolute error (SAE) from ground truth compared with benchmark estimations. Finally, our model provides an estimate of class-wise antihypertensive effect on cancer risk that is consistent with results derived from RCTs.
RESUMO
Ambient air pollution is projected to become a major environmental risk in sub-Saharan Africa (SSA). Research into its health impacts is hindered by limited data. We aimed to investigate the cross-sectional relationship between particulate matter with a diameter ≤ 2.5 µm (PM2.5) and prevalence of cough or acute lower respiratory infection (ALRI) among children under five in SSA. Data were collected from 31 Demographic and Health Surveys (DHS) in 21 SSA countries between 2005-2018. Prior-month average PM2.5 preceding the survey date was assessed based on satellite measurements and a chemical transport model. Cough and ALRI in the past two weeks were derived from questionnaires. Associations were analysed using conditional logistic regression within each survey cluster, adjusting for child's age, sex, birth size, household wealth, maternal education, maternal age and month of the interview. Survey-specific odds ratios (ORs) were pooled using random-effect meta-analysis. Included were 368,366 and 109,664 children for the analysis of cough and ALRI, respectively. On average, 20.5% children had reported a cough, 6.4% reported ALRI, and 32% of children lived in urban areas. Prior-month average PM2.5 ranged from 8.9 to 64.6 µg/m3. Pooling all surveys, no associations were observed with either outcome in the overall populations. Among countries with medium-to-high Human Development Index, positive associations were observed with both cough (pooled OR: 1.022, 95%CI: 0.982-1.064) and ALRI (pooled OR: 1.018, 95%CI: 0.975-1.064) for 1 µg/m3 higher of PM2.5. This explorative study found no associations between short-term ambient PM2.5 and respiratory health among young SSA children, necessitating future analyses using better-defined exposure and health metrics to study this important link.
Assuntos
Poluentes Atmosféricos , Poluição do Ar , África Subsaariana/epidemiologia , Poluentes Atmosféricos/análise , Poluentes Atmosféricos/toxicidade , Poluição do Ar/análise , Poluição do Ar/estatística & dados numéricos , Criança , Estudos Transversais , Exposição Ambiental/análise , Exposição Ambiental/estatística & dados numéricos , Humanos , Material Particulado/análise , Material Particulado/toxicidadeRESUMO
One major impediment to the wider use of deep learning for clinical decision making is the difficulty of assigning a level of confidence to model predictions. Currently, deep Bayesian neural networks and sparse Gaussian processes are the main two scalable uncertainty estimation methods. However, deep Bayesian neural networks suffer from lack of expressiveness, and more expressive models such as deep kernel learning, which is an extension of sparse Gaussian process, captures only the uncertainty from the higher-level latent space. Therefore, the deep learning model under it lacks interpretability and ignores uncertainty from the raw data. In this paper, we merge features of the deep Bayesian learning framework with deep kernel learning to leverage the strengths of both methods for a more comprehensive uncertainty estimation. Through a series of experiments on predicting the first incidence of heart failure, diabetes and depression applied to large-scale electronic medical records, we demonstrate that our method is better at capturing uncertainty than both Gaussian processes and deep Bayesian neural networks in terms of indicating data insufficiency and identifying misclassifications, with a comparable generalization performance. Furthermore, by assessing the accuracy and area under the receiver operating characteristic curve over the predictive probability, we show that our method is less susceptible to making overconfident predictions, especially for the minority class in imbalanced datasets. Finally, we demonstrate how uncertainty information derived by the model can inform risk factor analysis towards model interpretability.
RESUMO
Biochemical and medical literature establish lactate as a fundamental biomarker that can shed light on the energy consumption dynamics of the body at cellular and physiological levels. It is therefore, not surprising that it has been linked to many critical conditions ranging from the morbidity and mortality of critically ill patients to the diagnosis and prognosis of acute ischemic stroke, septic shock, lung injuries, insulin resistance in diabetic patients, and cancer. Currently, the gold standard for the measurement of lactate requires blood sampling. The invasive and costly nature of this procedure severely limits its application outside intensive care units. Optical sensors can provide a non-invasive, inexpensive, easy-to-use, continuous alternative to blood sampling. Previous efforts to achieve this have shown significant potential, but have been inconclusive. A measure that has been previously overlooked in this context, is the use of variable selection methods to identify regions of the optical spectrum that are most sensitive to and representative of the concentration of lactate. In this study, several wavelength selection methods are investigated and a new genetic algorithm-based wavelength selection method is proposed. This study shows that the development of more accurate and parsimonious models for optical estimation of lactate is possible. Unlike many existing methods, the proposed method does not impose additional locality constraints on the spectral features and therefore helps provide a much more granular interpretation of wavelength importance.