Showing papers in "PLOS ONE in 2019"

Machine learning algorithm validation with a limited sample size

Abstract: Advances in neuroimaging, genomic, motion tracking, eye-tracking and many other technology-based data collection methods have led to a torrent of high dimensional datasets, which commonly have a small number of samples because of the intrinsic high cost of data collection involving human participants. High dimensional data with a small number of samples is of critical importance for identifying biomarkers and conducting feasibility and pilot work, however it can lead to biased machine learning (ML) performance estimates. Our review of studies which have applied ML to predict autistic from non-autistic individuals showed that small sample size is associated with higher reported classification accuracy. Thus, we have investigated whether this bias could be caused by the use of validation methods which do not sufficiently control overfitting. Our simulations show that K-fold Cross-Validation (CV) produces strongly biased performance estimates with small sample sizes, and the bias is still evident with sample size of 1000. Nested CV and train/test split approaches produce robust and unbiased performance estimates regardless of sample size. We also show that feature selection if performed on pooled training and testing data is contributing to bias considerably more than parameter tuning. In addition, the contribution to bias by data dimensionality, hyper-parameter space and number of CV folds was explored, and validation methods were compared with discriminable data. The results suggest how to design robust testing methodologies when working with small datasets and how to interpret the results of other studies based on what validation method was used.

Biased efficacy estimates in phase-III dengue vaccine trials due to heterogeneous exposure and differential detectability of primary infections across trial arms.

Abstract: Vaccine efficacy (VE) estimates are crucial for assessing the suitability of dengue vaccine candidates for public health implementation, but efficacy trials are subject to a known bias to estimate VE toward the null if heterogeneous exposure is not accounted for in the analysis of trial data. In light of many well-characterized sources of heterogeneity in dengue virus (DENV) transmission, our goal was to estimate the potential magnitude of this bias in VE estimates for a hypothetical dengue vaccine. To ensure that we realistically modeled heterogeneous exposure, we simulated city-wide DENV transmission and vaccine trial protocols using an agent-based model calibrated with entomological and epidemiological data from long-term field studies in Iquitos, Peru. By simulating a vaccine with a true VE of 0.8 in 1,000 replicate trials each designed to attain 90% power, we found that conventional methods underestimated VE by as much as 21% due to heterogeneous exposure. Accounting for the number of exposures in the vaccine and placebo arms eliminated this bias completely, and the more realistic option of including a frailty term to model exposure as a random effect reduced this bias partially. We also discovered a distinct bias in VE estimates away from the null due to lower detectability of primary DENV infections among seronegative individuals in the vaccinated group. This difference in detectability resulted from our assumption that primary infections in vaccinees who are seronegative at baseline resemble secondary infections, which experience a shorter window of detectable viremia due to a quicker immune response. This resulted in an artefactual finding that VE estimates for the seronegative group were approximately 1% greater than for the seropositive group. Simulation models of vaccine trials that account for these factors can be used to anticipate the extent of bias in field trials and to aid in their interpretation.

Visualizing a field of research: A methodology of systematic scientometric reviews.

Abstract: Systematic scientometric reviews, empowered by computational and visual analytic approaches, offer opportunities to improve the timeliness, accessibility, and reproducibility of studies of the literature of a field of research. On the other hand, effectively and adequately identifying the most representative body of scholarly publications as the basis of subsequent analyses remains a common bottleneck in the current practice. What can we do to reduce the risk of missing something potentially significant? How can we compare different search strategies in terms of the relevance and specificity of topical areas covered? In this study, we introduce a flexible and generic methodology based on a significant extension of the general conceptual framework of citation indexing for delineating the literature of a research field. The method, through cascading citation expansion, provides a practical connection between studies of science from local and global perspectives. We demonstrate an application of the methodology to the research of literature-based discovery (LBD) and compare five datasets constructed based on three use scenarios and corresponding retrieval strategies, namely a query-based lexical search (one dataset), forward expansions starting from a groundbreaking article of LBD (two datasets), and backward expansions starting from a recently published review article by a prominent expert in LBD (two datasets). We particularly discuss the relevance of areas captured by expansion processes with reference to the query-based scientometric visualization. The method used in this study for comparing bibliometric datasets is applicable to comparative studies of search strategies.

Climate change has likely already affected global food production.

Abstract: Crop yields are projected to decrease under future climate conditions, and recent research suggests that yields have already been impacted. However, current impacts on a diversity of crops subnationally and implications for food security remains unclear. Here, we constructed linear regression relationships using weather and reported crop data to assess the potential impact of observed climate change on the yields of the top ten global crops-barley, cassava, maize, oil palm, rapeseed, rice, sorghum, soybean, sugarcane and wheat at ~20,000 political units. We find that the impact of global climate change on yields of different crops from climate trends ranged from -13.4% (oil palm) to 3.5% (soybean). Our results show that impacts are mostly negative in Europe, Southern Africa and Australia but generally positive in Latin America. Impacts in Asia and Northern and Central America are mixed. This has likely led to ~1% average reduction (-3.5 X 1013 kcal/year) in consumable food calories in these ten crops. In nearly half of food insecure countries, estimated caloric availability decreased. Our results suggest that climate change has already affected global food production.

Hate speech detection: Challenges and solutions.

Abstract: As online content continues to grow, so does the spread of hate speech. We identify and examine challenges faced by online automatic approaches for hate speech detection in text. Among these difficulties are subtleties in language, differing definitions on what constitutes hate speech, and limitations of data availability for training and testing of these systems. Furthermore, many recent approaches suffer from an interpretability problem-that is, it can be difficult to understand why the systems make the decisions that they do. We propose a multi-view SVM approach that achieves near state-of-the-art performance, while being simpler and producing more easily interpretable decisions than neural methods. We also discuss both technical and practical challenges that remain for this task.

Intraclass correlation - A discussion and demonstration of basic features.

Abstract: A re-analysis of intraclass correlation (ICC) theory is presented together with Monte Carlo simulations of ICC probability distributions. A partly revised and simplified theory of the single-score ICC is obtained, together with an alternative and simple recipe for its use in reliability studies. Our main, practical conclusion is that in the analysis of a reliability study it is neither necessary nor convenient to start from an initial choice of a specified statistical model. Rather, one may impartially use all three single-score ICC formulas. A near equality of the three ICC values indicates the absence of bias (systematic error), in which case the classical (one-way random) ICC may be used. A consistency ICC larger than absolute agreement ICC indicates the presence of non-negligible bias; if so, classical ICC is invalid and misleading. An F-test may be used to confirm whether biases are present. From the resulting model (without or with bias) variances and confidence intervals may then be calculated. In presence of bias, both absolute agreement ICC and consistency ICC should be reported, since they give different and complementary information about the reliability of the method. A clinical example with data from the literature is given.

Applications of artificial neural networks in health care organizational decision-making: A scoping review

Abstract: Health care organizations are leveraging machine-learning techniques, such as artificial neural networks (ANN), to improve delivery of care at a reduced cost. Applications of ANN to diagnosis are well-known; however, ANN are increasingly used to inform health care management decisions. We provide a seminal review of the applications of ANN to health care organizational decision-making. We screened 3,397 articles from six databases with coverage of Health Administration, Computer Science and Business Administration. We extracted study characteristics, aim, methodology and context (including level of analysis) from 80 articles meeting inclusion criteria. Articles were published from 1997–2018 and originated from 24 countries, with a plurality of papers (26 articles) published by authors from the United States. Types of ANN used included ANN (36 articles), feed-forward networks (25 articles), or hybrid models (23 articles); reported accuracy varied from 50% to 100%. The majority of ANN informed decision-making at the micro level (61 articles), between patients and health care providers. Fewer ANN were deployed for intra-organizational (meso- level, 29 articles) and system, policy or inter-organizational (macro- level, 10 articles) decision-making. Our review identifies key characteristics and drivers for market uptake of ANN for health care organizational decision-making to guide further adoption of this technique.

Clustering algorithms: A comparative approach

Abstract: Many real-world systems can be studied in terms of pattern recognition tasks, so that proper use (and understanding) of machine learning methods in practical applications becomes essential. While many classification methods have been proposed, there is no consensus on which methods are more suitable for a given dataset. As a consequence, it is important to comprehensively compare methods in many possible scenarios. In this context, we performed a systematic comparison of 9 well-known clustering methods available in the R language assuming normally distributed data. In order to account for the many possible variations of data, we considered artificial datasets with several tunable properties (number of classes, separation between classes, etc). In addition, we also evaluated the sensitivity of the clustering methods with regard to their parameters configuration. The results revealed that, when considering the default configurations of the adopted methods, the spectral approach tended to present particularly good performance. We also found that the default configuration of the adopted implementations was not always accurate. In these cases, a simple approach based on random selection of parameters values proved to be a good alternative to improve the performance. All in all, the reported approach provides subsidies guiding the choice of clustering algorithms.

Systematic review of patient and caregivers' satisfaction with telehealth videoconferencing as a mode of service delivery in managing patients' health.

Abstract: Telehealth is an alternative method of delivering health care to people required to travel long distances for routine health care. The aim of this systematic review was to examine whether patients and their caregivers living in rural and remote areas are satisfied with telehealth videoconferencing as a mode of service delivery in managing their health. A protocol was registered with PROSPERO international prospective register of systematic reviews (#CRD42017083597) and conducted in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. A systematic search of Ovid Medline, Embase, CINAHL, ProQuest Health Research Premium Collection, Joanna Briggs Institute and the Cochrane Library was conducted. Studies of people living in rural and remote areas who attended outpatient appointments for a health condition via videoconference were included if the studies measured patient and/or caregivers' satisfaction with telehealth. Data on satisfaction was extracted and descriptively synthesised. Methodological quality of the included studies was assessed using a modified version of the McMaster Critical Review Forms for Quantitative or Qualitative Studies. Thirty-six studies of varying study design and quality met the inclusion criteria. The outcomes of satisfaction with telehealth were categorised into system experience, information sharing, consumer focus and overall satisfaction. There were high levels of satisfaction across all these dimensions. Despite these positive findings, the current evidence base lacks clarity in terms of how satisfaction is defined and measured. People living in rural and remote areas are generally satisfied with telehealth as a mode of service delivery as it may improve access to health care and avoid the inconvenience of travel.

Lenvatinib plus anti-PD-1 antibody combination treatment activates CD8+ T cells through reduction of tumor-associated macrophage and activation of the interferon pathway.

Abstract: Lenvatinib is a multiple receptor tyrosine kinase inhibitor targeting mainly vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) receptors. We investigated the immunomodulatory activities of lenvatinib in the tumor microenvironment and its mechanisms of enhanced antitumor activity when combined with a programmed cell death-1 (PD-1) blockade. Antitumor activity was examined in immunodeficient and immunocompetent mouse tumor models. Single-cell analysis, flow cytometric analysis, and immunohistochemistry were used to analyze immune cell populations and their activation. Gene co-expression network analysis and pathway analysis using RNA sequencing data were used to identify lenvatinib-driven combined activity with anti-PD-1 antibody (anti-PD-1). Lenvatinib showed potent antitumor activity in the immunocompetent tumor microenvironment compared with the immunodeficient tumor microenvironment. Antitumor activity of lenvatinib plus anti-PD-1 was greater than that of either single treatment. Flow cytometric analysis revealed that lenvatinib reduced tumor-associated macrophages (TAMs) and increased the percentage of activated CD8+ T cells secreting interferon (IFN)-γ+ and granzyme B (GzmB). Combination treatment further increased the percentage of T cells, especially CD8+ T cells, among CD45+ cells and increased IFN-γ+ and GzmB+ CD8+ T cells. Transcriptome analyses of tumors resected from treated mice showed that genes specifically regulated by the combination were significantly enriched for type-I IFN signaling. Pretreatment with lenvatinib followed by anti-PD-1 treatment induced significant antitumor activity compared with anti-PD-1 treatment alone. Our findings show that lenvatinib modulates cancer immunity in the tumor microenvironment by reducing TAMs and, when combined with PD-1 blockade, shows enhanced antitumor activity via the IFN signaling pathway. These findings provide a scientific rationale for combination therapy of lenvatinib with PD-1 blockade to improve cancer immunotherapy.

Suicide among physicians and health-care workers: A systematic review and meta-analysis.

Abstract: Background : Medical-related professions are at high suicide risk. However, data are contradictory and comparisons were not made between gender, occupation and specialties, epochs of times. Thus, we conducted a systematic review and meta-analysis on suicide risk among healthcare workers. Method : The PubMed, Cochrane Library, Science Direct and Embase databases were searched without language restriction on April 2019, with the following keywords: suicide* AND (« health care worker* » OR physician* OR nurse*). When possible, we stratified results by gender, countries, time, and specialties. Estimates were pooled using random-effect metaanalysis. Differences by study-level characteristics were estimated using stratified metaanalysis and meta-regression. Suicides, suicidal attempts, and suicidal ideation were retrieved from national or local specific registers or case records. In addition, suicide attempts and suicidal ideation were also retrieved from questionnaires (paper or internet). Results : The overall SMR for suicide in physicians was 1.44 (95CI 1.16, 1.72) with an important heterogeneity (I2 = 93.9%, p<0.001). Female were at higher risk (SMR = 1.9; 95CI 1.49, 2.58; and ES = 0.67; 95CI 0.19, 1.14; p<0.001 compared to male). US physicians were at higher risk (ES = 1.34; 95CI 1.28, 1.55; p <0.001 vs Rest of the world). Suicide decreased over time, especially in Europe (ES = -0.18; 95CI -0.37, -0.01; p = 0.044). Some specialties might be at higher risk such as anesthesiologists, psychiatrists, general practitioners and general surgeons. There were 1.0% (95CI 1.0, 2.0; p<0.001) of suicide attempts and 17% (95CI 12, 21; p<0.001) of suicidal ideation in physicians. Insufficient data precluded meta-analysis on other health-care workers. Conclusion : Physicians are an at-risk profession of suicide, with women particularly at risk. The rate of suicide in physicians decreased over time, especially in Europe. The high prevalence of physicians who committed suicide attempt as well as those with suicidal ideation should benefits for preventive strategies at the workplace. Finally, the lack of data on other health-care workers suggest to implement studies investigating those occupations.

Cardiovascular disease risk prediction using automated machine learning: A prospective study of 423,604 UK Biobank participants.

Abstract: Background Identifying people at risk of cardiovascular diseases (CVD) is a cornerstone of preventative cardiology. Risk prediction models currently recommended by clinical guidelines are typically based on a limited number of predictors with sub-optimal performance across all patient groups. Data-driven techniques based on machine learning (ML) might improve the performance of risk predictions by agnostically discovering novel risk predictors and learning the complex interactions between them. We tested (1) whether ML techniques based on a state-of-the-art automated ML framework (AutoPrognosis) could improve CVD risk prediction compared to traditional approaches, and (2) whether considering non-traditional variables could increase the accuracy of CVD risk predictions. Methods and findings Using data on 423,604 participants without CVD at baseline in UK Biobank, we developed a ML-based model for predicting CVD risk based on 473 available variables. Our ML-based model was derived using AutoPrognosis, an algorithmic tool that automatically selects and tunes ensembles of ML modeling pipelines (comprising data imputation, feature processing, classification and calibration algorithms). We compared our model with a well-established risk prediction algorithm based on conventional CVD risk factors (Framingham score), a Cox proportional hazards (PH) model based on familiar risk factors (i.e, age, gender, smoking status, systolic blood pressure, history of diabetes, reception of treatments for hypertension and body mass index), and a Cox PH model based on all of the 473 available variables. Predictive performances were assessed using area under the receiver operating characteristic curve (AUC-ROC). Overall, our AutoPrognosis model improved risk prediction (AUC-ROC: 0.774, 95% CI: 0.768-0.780) compared to Framingham score (AUC-ROC: 0.724, 95% CI: 0.720-0.728, p < 0.001), Cox PH model with conventional risk factors (AUC-ROC: 0.734, 95% CI: 0.729-0.739, p < 0.001), and Cox PH model with all UK Biobank variables (AUC-ROC: 0.758, 95% CI: 0.753-0.763, p < 0.001). Out of 4,801 CVD cases recorded within 5 years of baseline, AutoPrognosis was able to correctly predict 368 more cases compared to the Framingham score. Our AutoPrognosis model included predictors that are not usually considered in existing risk prediction models, such as the individuals’ usual walking pace and their self-reported overall health rating. Furthermore, our model improved risk prediction in potentially relevant sub-populations, such as in individuals with history of diabetes. We also highlight the relative benefits accrued from including more information into a predictive model (information gain) as compared to the benefits of using more complex models (modeling gain). Conclusions Our AutoPrognosis model improves the accuracy of CVD risk prediction in the UK Biobank population. This approach performs well in traditionally poorly served patient subgroups. Additionally, AutoPrognosis uncovered novel predictors for CVD disease that may now be tested in prospective studies. We found that the “information gain” achieved by considering more risk factors in the predictive model was significantly higher than the “modeling gain” achieved by adopting complex predictive models.

Reproduction study using public data of: Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs.

Abstract: We have attempted to reproduce the results in Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs, published in JAMA 2016; 316(22), using publicly available data sets. We re-implemented the main method in the original study since the source code is not available. The original study used non-public fundus images from EyePACS and three hospitals in India for training. We used a different EyePACS data set from Kaggle. The original study used the benchmark data set Messidor-2 to evaluate the algorithm's performance. We used another distribution of the Messidor-2 data set, since the original data set is no longer available. In the original study, ophthalmologists re-graded all images for diabetic retinopathy, macular edema, and image gradability. We have one diabetic retinopathy grade per image for our data sets, and we assessed image gradability ourselves. We were not able to reproduce the original study's results with publicly available data. Our algorithm's area under the receiver operating characteristic curve (AUC) of 0.951 (95% CI, 0.947-0.956) on the Kaggle EyePACS test set and 0.853 (95% CI, 0.835-0.871) on Messidor-2 did not come close to the reported AUC of 0.99 on both test sets in the original study. This may be caused by the use of a single grade per image, or different data. This study shows the challenges of reproducing deep learning method results, and the need for more replication and reproduction studies to validate deep learning methods, especially for medical image analysis. Our source code and instructions are available at: https://github.com/mikevoets/jama16-retina-replication.

Global burden of maternal and congenital syphilis and associated adverse birth outcomes—Estimates for 2016 and progress since 2012

Abstract: Background In 2007 the World Health Organization (WHO) launched the global initiative to eliminate mother-to-child transmission of syphilis (congenital syphilis, or CS). To assess progress towards the goal of <50 CS cases per 100,000 live births, we generated regional and global estimates of maternal and congenital syphilis for 2016 and updated the 2012 estimates. Methods Maternal syphilis estimates were generated using the Spectrum-STI model, fitted to sentinel surveys and routine testing of pregnant women during antenatal care (ANC) and other representative population data. Global and regional estimates of CS used the same approach as previous WHO estimates. Results The estimated global maternal syphilis prevalence in 2016 was 0.69% (95% confidence interval: 0.57–0.81%) resulting in a global CS rate of 473 (385–561) per 100,000 live births and 661,000 (538,000–784,000) total CS cases, including 355,000 (290,000–419,000) adverse birth outcomes (ABO) and 306,000 (249,000–363,000) non-clinical CS cases (infants without clinical signs born to un-treated mothers). The ABOs included 143,000 early fetal deaths and stillbirths, 61,000 neonatal deaths, 41,000 preterm or low-birth weight births, and 109,000 infants with clinical CS. Of these ABOs– 203,000 (57%) occurred in pregnant women attending ANC but not screened for syphilis; 74,000 (21%) in mothers not enrolled in ANC, 55,000 (16%) in mothers screened but not treated, and 23,000 (6%) in mothers enrolled, screened and treated. The revised 2012 estimates were 0.70% (95% CI: 0.63–0.77%) maternal prevalence, and 748,000 CS cases (539 per 100,000 live births) including 397,000 (361,000–432,000) ABOs. The estimated decrease in CS case rates between 2012 and 2016 reflected increased access to ANC and to syphilis screening and treatment. Conclusions Congenital syphilis decreased worldwide between 2012 and 2016, although maternal prevalence was stable. Achieving global CS elimination, however, will require improving access to early syphilis screening and treatment in ANC, clinically monitoring all women diagnosed with syphilis and their infants, improving partner management, and reducing syphilis prevalence in the general population by expanding testing, treatment and partner referral beyond ANC.

Classification of skin lesions using transfer learning and augmentation with Alex-net

Abstract: Skin cancer is one of most deadly diseases in humans. According to the high similarity between melanoma and nevus lesions, physicians take much more time to investigate these lesions. The automated classification of skin lesions will save effort, time and human life. The purpose of this paper is to present an automatic skin lesions classification system with higher classification rate using the theory of transfer learning and the pre-trained deep neural network. The transfer learning has been applied to the Alex-net in different ways, including fine-tuning the weights of the architecture, replacing the classification layer with a softmax layer that works with two or three kinds of skin lesions, and augmenting dataset by fixed and random rotation angles. The new softmax layer has the ability to classify the segmented color image lesions into melanoma and nevus or into melanoma, seborrheic keratosis, and nevus. The three well-known datasets, MED-NODE, Derm (IS & Quest) and ISIC, are used in testing and verifying the proposed method. The proposed DCNN weights have been fine-tuned using the training and testing dataset from ISIC in addition to 10-fold cross validation for MED-NODE and DermIS—DermQuest. The accuracy, sensitivity, specificity, and precision measures are used to evaluate the performance of the proposed method and the existing methods. For the datasets, MED-NODE, Derm (IS & Quest) and ISIC, the proposed method has achieved accuracy percentages of 96.86%, 97.70%, and 95.91% respectively. The performance of the proposed method has outperformed the performance of the existing classification methods of skin cancer.

SleepEEGNet: Automated sleep stage scoring with sequence to sequence deep learning approach.

Abstract: Electroencephalogram (EEG) is a common base signal used to monitor brain activities and diagnose sleep disorders. Manual sleep stage scoring is a time-consuming task for sleep experts and is limited by inter-rater reliability. In this paper, we propose an automatic sleep stage annotation method called SleepEEGNet using a single-channel EEG signal. The SleepEEGNet is composed of deep convolutional neural networks (CNNs) to extract time-invariant features, frequency information, and a sequence to sequence model to capture the complex and long short-term context dependencies between sleep epochs and scores. In addition, to reduce the effect of the class imbalance problem presented in the available sleep datasets, we applied novel loss functions to have an equal misclassified error for each sleep stage while training the network. We evaluated the performance of the proposed method on different single-EEG channels (i.e., Fpz-Cz and Pz-Oz EEG channels) from the Physionet Sleep-EDF datasets published in 2013 and 2018. The evaluation results demonstrate that the proposed method achieved the best annotation performance compared to current literature, with an overall accuracy of 84.26%, a macro F1-score of 79.66% and κ = 0.79. Our developed model can be applied to other sleep EEG signals and aid the sleep specialists to arrive at an accurate diagnosis. The source code is available at https://github.com/SajadMo/SleepEEGNet.

Breast cancer histopathological image classification using convolutional neural networks with small SE-ResNet module.

Abstract: Although successful detection of malignant tumors from histopathological images largely depends on the long-term experience of radiologists, experts sometimes disagree with their decisions. Computer-aided diagnosis provides a second option for image diagnosis, which can improve the reliability of experts’ decision-making. Automatic and precision classification for breast cancer histopathological image is of great importance in clinical application for identifying malignant tumors from histopathological images. Advanced convolution neural network technology has achieved great success in natural image classification, and it has been used widely in biomedical image processing. In this paper, we design a novel convolutional neural network, which includes a convolutional layer, small SE-ResNet module, and fully connected layer. We propose a small SE-ResNet module which is an improvement on the combination of residual module and Squeeze-and-Excitation block, and achieves the similar performance with fewer parameters. In addition, we propose a new learning rate scheduler which can get excellent performance without complicatedly fine-tuning the learning rate. We use our model for the automatic classification of breast cancer histology images (BreakHis dataset) into benign and malignant and eight subtypes. The results show that our model achieves the accuracy between 98.87% and 99.34% for the binary classification and achieve the accuracy between 90.66% and 93.81% for the multi-class classification.

The 2017 human monkeypox outbreak in Nigeria—Report of outbreak experience and response in the Niger Delta University Teaching Hospital, Bayelsa State, Nigeria

Abstract: Background In September 2017, Nigeria experienced a large outbreak of human monkeypox (HMPX). In this study, we report the outbreak experience and response in the Niger Delta University Teaching Hospital (NDUTH), Bayelsa state, where the index case and majority of suspected cases were reported. Methods In a cross-sectional study between September 25th and 31st December 2017, we reviewed the clinical and laboratory characteristics of all suspected and confirmed cases of HMPX seen at the NDUTH and appraised the plans, activities and challenges of the hospital in response to the outbreak based on documented observations of the hospital's infection control committee (IPC). Monkeypox cases were defined using the interim national guidelines as provided by the Nigerian Centre for Disease Control (NCDC). Results Of 38 suspected cases of HMPX, 18(47.4%) were laboratory confirmed, 3(7.9%) were probable, while 17 (18.4%) did not fit the case definition for HMPX. Majority of the confirmed/probable cases were adults (80.9%) and males (80.9%). There was concomitant chicken pox, syphilis and HIV-1 infections in two confirmed cases and a case of nosocomial infection in one healthcare worker (HCW). The hospital established a make-shift isolation ward for case management, constituted a HMPX response team and provided IPC resources. At the outset, some HCWs were reluctant to participate in the outbreak and others avoided suspected patients. Some patients and their family members experienced stigma and discrimination and there were cases of refusal of isolation. Repeated trainings and collaborative efforts by all stakeholders addressed some of these challenges and eventually led to successful containment of the outbreak. Conclusion While the 2017 outbreak of human monkeypox in Nigeria was contained, our report reveals gaps in outbreak response that could serve as lessons to other hospitals to strengthen epidemic preparedness and response activities in the hospital setting.

Forecasting stock prices with a feature fusion LSTM-CNN model using different representations of the same data.

Abstract: Forecasting stock prices plays an important role in setting a trading strategy or determining the appropriate timing for buying or selling a stock. We propose a model, called the feature fusion long short-term memory-convolutional neural network (LSTM-CNN) model, that combines features learned from different representations of the same data, namely, stock time series and stock chart images, to predict stock prices. The proposed model is composed of LSTM and a CNN, which are utilized for extracting temporal features and image features. We measure the performance of the proposed model relative to those of single models (CNN and LSTM) using SPDR S&P 500 ETF data. Our feature fusion LSTM-CNN model outperforms the single models in predicting stock prices. In addition, we discover that a candlestick chart is the most appropriate stock chart image to use to forecast stock prices. Thus, this study shows that prediction error can be efficiently reduced by using a combination of temporal and image features from the same data rather than using these features separately.

Early mobilization of critically ill patients in the intensive care unit: A systematic review and meta-analysis.

Abstract: Background Physical therapy can prevent functional impairments and improve the quality of life of patients after hospital discharge. However, the effect of early mobilization on patients with a critical illness remains unclear. This study was performed to assess the evidence available regarding the effect of early mobilization on critically ill patients in the intensive care unit (ICU). Methods Electronic databases were searched from their inception to March 21, 2019. Randomized controlled trials (RCTs) comprising critically ill patients who received early mobilization were included. The methodological quality and risk of bias of each eligible trial were assessed using the Cochrane Collaboration tool. Data were extracted using a standard collection form each included study, and processed using the Mantel-Haenszel (M-H) or inverse-variance (I-V) test in the STATA v12.0 statistical software. Results A total of 1,898 records were screened. Twenty-three RCTs comprising 2,308 critically ill patients were ultimately included. Early mobilization decreased the incidence of ICU-acquired weakness (ICU-AW) at hospital discharge (three studies, 190 patients, relative risk (RR): 0.60, 95% confidence interval (CI) [0.40, 0.90]; p = 0.013, I2 = 0.0%), increased the number of patients who were able to stand (one study, 50 patients, 90% vs. 62%, p = 0.02), increased the number of ventilator-free days (six studies, 745 patients, standardized mean difference (SMD): 0.17, 95% CI [0.02, 0.31]; p = 0.023, I2 = 35.5%) during hospitalization, increased the distance the patient was able to walk unassisted (one study, 104 patients, 33.4 (0-91.4) meters vs. 0 (0-30.4) meters, p = 0.004) at hospital discharge, and increased the discharged-to-home rate (seven studies, 793 patients, RR: 1.16, 95% CI [1.00, 1.34]; p = 0.046). The mortality (28-day, ICU and hospital) and adverse event rates were moderately increased by early mobilization, but the differences were statistically non-significant. However, due to the substantial heterogeneity among the included studies, and the low quality of the evidence, the results of this study should be interpreted with caution. Publication bias was not identified. Conclusions Early mobilization appears to decrease the incidence of ICU-AW, improve the functional capacity, and increase the number of ventilator-free days and the discharged-to-home rate for patients with a critical illness in the ICU setting.

Socioeconomic status, stressful life situations and mental health problems in children and adolescents: Results of the German BELLA cohort-study.

Abstract: Aim Children and adolescents with low socioeconomic status (SES) suffer from mental health problems more often than their peers with high SES. The aim of the current study was to investigate the direct and interactive association between commonly used indicators of SES and the exposure to stressful life situations in relation to children’s mental health problems. Methods The prospective BELLA cohort study is the mental health module of the representative, population-based German National Health Interview and Examination Survey for children and adolescents (KiGGS). Sample data include 2,111 participants (aged 7–17 years at baseline) from the first three measurement points (2003–2006, 2004–2007 and 2005–2008). Hierarchical multiple linear regression models were conducted to analyze associations among the SES indicators household income, parental education and parental unemployment (assessed at baseline), number of stressful life situations (e.g., parental accident, mental illness or severe financial crises; 1- and 2-year follow-ups) and parent-reported mental health problems (Strength and Difficulties Questionnaire; 2-year follow-up). Results All indicators of SES separately predicted mental health problems in children and adolescents at the 2-year follow-up. Stressful life situations (between baseline and 2-year follow-up) and the interaction of parental education and the number of stressful life situations remained significant in predicting children’s mental health problems after adjustment for control variables. Thereby, children with higher educated parents showed fewer mental health problems in a stressful life situation. No moderating effect was found for household income and parental employment. Overall, the detected effect sizes were small. Mental health problems at baseline were the best predictor for mental health problems two years later. Conclusions Children and adolescents with a low SES suffer from multiple stressful life situations and are exposed to a higher risk of developing mental health problems. The findings suggest that the reduction of socioeconomic inequalities and interventions for families with low parental education might help to reduce children’s mental health problems.

Time to initial cancer treatment in the United States and association with survival over time: An observational study.

Abstract: Background Delays in time to treatment initiation (TTI) for new cancer diagnoses cause patient distress and may adversely affect outcomes. We investigated trends in TTI for common solid tumors treated with curative intent, determinants of increased TTI and association with overall survival. Methods and findings We utilized prospective data from the National Cancer Database for newly diagnosed United States patients with early-stage breast, prostate, lung, colorectal, renal and pancreas cancers from 2004–13. TTI was defined as days from diagnosis to first treatment (surgery, systemic or radiation therapy). Negative binomial regression and Cox proportional hazard models were used for analysis. The study population of 3,672,561 patients included breast (N = 1,368,024), prostate (N = 944,246), colorectal (N = 662,094), non-small cell lung (N = 363,863), renal (N = 262,915) and pancreas (N = 71,419) cancers. Median TTI increased from 21 to 29 days (P<0.001). Aside from year of diagnosis, determinants of increased TTI included care at academic center, race, education, prior history of cancer, transfer of facility, comorbidities and age. Increased TTI was associated with worsened survival for stages I and II breast, lung, renal and pancreas cancers, and stage I colorectal cancers, with hazard ratios ranging from 1.005 (95% confidence intervals [CI] 1.002–1.008) to 1.030 (95% CI 1.025–1.035) per week of increased TTI. Conclusions TTI has lengthened significantly and is associated with absolute increased risk of mortality ranging from 1.2–3.2% per week in curative settings such as early-stage breast, lung, renal and pancreas cancers. Studies of interventions to ease navigation and reduce barriers are warranted to diminish potential harm to patients.

Instruments for the detection of frailty syndrome in older adults: A systematic review

Abstract: Frailty is a dynamic process in which there is a reduction in the physical, psychological and/or social function associated with aging. The aim of this study was to identify instruments for the detection of frailty in older adults, characterizing their components, application scenarios, ability to identify pre-frailty and clinimetric properties evaluated. The study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), under registration number CRD42017039318. A total of 14 electronic sources were searched to identify studies that investigated instruments for the detection of frailty or that presented the construction and/or clinimetric evaluation of the instrument, according to criteria established by the COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN). 96 studies were included in the qualitative synthesis: 51 instruments for the detection of frailty were identified, with predominantly physical domains; 40 were constructed and/or validated for use in the older adult community population, 28 only highlighted the distinction between frail and non-frail individuals and 23 presented three or more levels of frailty. The FRAGIRE, FRAIL Scale, Edmonton Frail Scale and IVCF-20 instruments were the most frequently analyzed in relation to clinimetric properties. It was concluded that: (I) there is a large number of instruments for measuring the same construct, which makes it difficult for researchers and clinicians to choose the most appropriate; (II) the FRAGIRE and CFAI stand out due to their multidimensional aspects, including an environmental assessment; however, (III) the need for standardization of the scales was identified, since the use of different instruments in clinical trials may prevent the comparability of the results in systematic reviews and; (IV) considering the different instruments identified in this review, the choice of researchers/clinicians should be guided by the issues related to the translation and validation for their location and the suitability for their context.

Helping patients help themselves: A systematic review of self-management support strategies in primary health care practice

Abstract: Background Primary health professionals are well positioned to support the delivery of patient self-management in an evidence-based, structured capacity. A need exists to better understand the active components required for effective self-management support, how these might be delivered within primary care, and the training and system changes that would subsequently be needed. Objectives (1) To examine self-management support interventions in primary care on health outcomes for a wide range of diseases compared to usual standard of care; and (2) To identify the effective strategies that facilitate positive clinical and humanistic outcomes in this setting. Method A systematic review of randomized controlled trials evaluating self-management support interventions was conducted following the Cochrane handbook & PRISMA guidelines. Published literature was systematically searched from inception to June 2019 in PubMed, Scopus and Web of Science. Eligible studies assessed the effectiveness of individualized interventions with follow-up, delivered face-to-face to adult patients with any condition in primary care, compared with usual standard of care. Matrices were developed that mapped the evidence and components for each intervention. The methodological quality of included studies were appraised. Results 6,510 records were retrieved. 58 studies were included in the final qualitative synthesis. Findings reveal a structured patient-provider exchange is required in primary care (including a one-on-one patient-provider consultation, ongoing follow up and provision of self-help materials). Interventions should be tailored to patient needs and may include combinations of strategies to improve a patient's disease or treatment knowledge; independent monitoring of symptoms, encouraging self-treatment through a personalized action plan in response worsening symptoms or exacerbations, psychological coping and stress management strategies, and enhancing responsibility in medication adherence and lifestyle choices. Follow-up may include tailored feedback, monitoring of progress with respect to patient set healthcare goals, or honing problem-solving and decision-making skills. Theoretical models provided a strong base for effective SMS interventions. Positive outcomes for effective SMS included improvements in clinical indicators, health-related quality of life, self-efficacy (confidence to self-manage), disease knowledge or control. An SMS model has been developed which sets the foundation for the design and evaluation of practical strategies for the construct of self-management support interventions in primary healthcare practice. Conclusions These findings provide primary care professionals with evidence-based strategies and structure to deliver SMS in practice. For this collaborative partnership approach to be more widely applied, future research should build on these findings for optimal SMS service design and upskilling healthcare providers to effectively support patients in this collaborative process.

Dose-response relationship between exercise and cognitive function in older adults with and without cognitive impairment: A systematic review and meta-analysis

Abstract: This systematic review and meta-analysis examined the dose-response relationship between exercise and cognitive function in older adults with and without cognitive impairments. We included single-modality randomized controlled aerobic, anaerobic, multicomponent or psychomotor exercise trials that quantified training frequency, session and program duration and specified intensity quantitatively or qualitatively. We defined total exercise duration in minutes as the product of program duration, session duration, and frequency. For each study, we grouped test-specific Hedges' d (n = 163) and Cohen's d (n = 23) effect sizes in the domains Global cognition, Executive function and Memory. We used multilevel mixed-effects models to investigate dose-related predictors of exercise effects. In healthy older adults (n = 23 studies), there was a small positive effect of exercise on executive function (d = 0.27) and memory (d = 0.24), but dose-parameters did not predict the magnitude of effect sizes. In older adults with cognitive impairments (n = 13 studies), exercise had a moderate positive effect on global cognition (d = 0.37). For older adults with cognitive impairments, we found evidence for exercise programs with a short session duration and high frequency to predict higher effect sizes (d = 0.43-0.50). In healthy older adults, dose-parameters did not predict the magnitude of exercise effects on cognition. For older adults with cognitive impairments, exercise programs with shorter session duration and higher frequency may generate the best cognitive results. Studies are needed in which different exercise doses are directly compared among randomized subjects or conditions.

The effects of chemical and organic fertilizer usage on rhizosphere soil in tea orchards.

Abstract: Sustainable agriculture is an important global issue. The use of organic fertilizers can enhance crop yield and soil properties while restraining pests and diseases. The objective of this study was to assess the effects of long-term use of chemical and organic fertilizers on tea and rhizosphere soil properties in tea orchards. Inductively coupled plasma mass spectrometry (ICP-MS) and high-throughput sequencing technology analyses were used to investigate heavy metals content and bacterial composition in rhizosphere soils. Our results indicated that organic fertilizer treatment significantly decreased Cu, Pb and Cd contents in rhizosphere soil sample. The results also showed that treatment with organic fertilizer significantly decreased the contents of Cd, Pb and As in tea leaves. Furthermore, organic fertilizer significantly increased the amino acids content of tea and the pH of the soil. The use of organic fertilizer significantly increased in the relative abundance of Burkholderiales, Myxococcales, Streptomycetales, Nitrospirales, Ktedonobacterales, Acidobacteriales, Gemmatimonadales, and Solibacterales, and decreased the abundance of Pseudonocardiales, Frankiales, Rhizobiales, and Xanthomonadales. In conclusion, organic fertilizer can help to shape the microbial composition and recruit beneficial bacteria into the rhizosphere of tea, leading to improved tea quality and reduced heavy metals content in rhizosphere soil and tea leaves.

Consistency and differences between centrality measures across distinct classes of networks.

Abstract: The roles of different nodes within a network are often understood through centrality analysis, which aims to quantify the capacity of a node to influence, or be influenced by, other nodes via its connection topology. Many different centrality measures have been proposed, but the degree to which they offer unique information, and whether it is advantageous to use multiple centrality measures to define node roles, is unclear. Here we calculate correlations between 17 different centrality measures across 212 diverse real-world networks, examine how these correlations relate to variations in network density and global topology, and investigate whether nodes can be clustered into distinct classes according to their centrality profiles. We find that centrality measures are generally positively correlated to each other, the strength of these correlations varies across networks, and network modularity plays a key role in driving these cross-network variations. Data-driven clustering of nodes based on centrality profiles can distinguish different roles, including topological cores of highly central nodes and peripheries of less central nodes. Our findings illustrate how network topology shapes the pattern of correlations between centrality measures and demonstrate how a comparative approach to network centrality can inform the interpretation of nodal roles in complex networks.

The composition of bacterial communities associated with plastic biofilms differs between different polymers and stages of biofilm succession

Abstract: Once in the ocean, plastics are rapidly colonized by complex microbial communities. Factors affecting the development and composition of these communities are still poorly understood. Additionally, whether there are plastic-type specific communities developing on different plastics remains enigmatic. We determined the development and succession of bacterial communities on different plastics under ambient and dim light conditions in the coastal Northern Adriatic over the course of two months using scanning electron microscopy and 16S rRNA gene analyses. Plastics used were low- and high-density polyethylene (LDPE and HDPE, respectively), polypropylene (PP) and polyvinyl chloride with two typical additives (PVC DEHP and PVC DINP). The bacterial communities developing on the plastics clustered in two groups; one group was found on PVC and the other group on all the other plastics and on glass, which was used as an inert control. Specific bacterial taxa were found on specific surfaces in essentially all stages of biofilm development and in both ambient and dim light conditions. Differences in bacterial community composition between the different plastics and light exposures were stronger after an incubation period of one week than at the later stages of the incubation. Under both ambient and dim light conditions, one part of the bacterial community was common on all plastic types, especially in later stages of the biofilm development, with families such as Flavobacteriaceae, Rhodobacteraceae, Planctomycetaceae and Phyllobacteriaceae presenting relatively high relative abundances on all surfaces. Another part of the bacterial community was plastic-type specific. The plastic-type specific fraction was variable among the different plastic types and was more abundant after one week of incubation than at later stages of the succession.

Why Cohen's Kappa should be avoided as performance measure in classification

Abstract: We show that Cohen’s Kappa and Matthews Correlation Coefficient (MCC), both extended and contrasted measures of performance in multi-class classification, are correlated in most situations, albeit can differ in others. Indeed, although in the symmetric case both match, we consider different unbalanced situations in which Kappa exhibits an undesired behaviour, i.e. a worse classifier gets higher Kappa score, differing qualitatively from that of MCC. The debate about the incoherence in the behaviour of Kappa revolves around the convenience, or not, of using a relative metric, which makes the interpretation of its values difficult. We extend these concerns by showing that its pitfalls can go even further. Through experimentation, we present a novel approach to this topic. We carry on a comprehensive study that identifies an scenario in which the contradictory behaviour among MCC and Kappa emerges. Specifically, we find out that when there is a decrease to zero of the entropy of the elements out of the diagonal of the confusion matrix associated to a classifier, the discrepancy between Kappa and MCC rise, pointing to an anomalous performance of the former. We believe that this finding disables Kappa to be used in general as a performance measure to compare classifiers.

Community participation in health services development, implementation, and evaluation: A systematic review of empowerment, health, community, and process outcomes.

Abstract: Background Community participation is widely believed to be beneficial to the development, implementation and evaluation of health services. However, many challenges to successful and sustainable community involvement remain. Importantly, there is little evidence on the effect of community participation in terms of outcomes at both the community and individual level. Our systematic review seeks to examine the evidence on outcomes of community participation in high and upper-middle income countries. Methods and findings This review was developed according to PRISMA guidelines. Eligible studies included those that involved the community, service users, consumers, households, patients, public and their representatives in the development, implementation, and evaluation of health services, policy or interventions. We searched the following databases from January 2000 to September 2016: Medline, Embase, Global Health, Scopus, and LILACs. We independently screened articles for inclusion, conducted data extraction, and assessed studies for risk of bias. No language restrictions were made. 27,232 records were identified, with 23,468 after removal of duplicates. Following titles and abstracts screening, 49 met the inclusion criteria for this review. A narrative synthesis of the findings was conducted. Outcomes were categorised as process outcomes, community outcomes, health outcomes, empowerment and stakeholder perspectives. Our review reports a breadth of evidence that community involvement has a positive impact on health, particularly when substantiated by strong organisational and community processes. This is in line with the notion that participatory approaches and positive outcomes including community empowerment and health improvements do not occur in a linear progression, but instead consists of complex processes influenced by an array of social and cultural factors. Conclusion This review adds to the evidence base supporting the effectiveness of community participation in yielding positive outcomes at the organizational, community and individual level. Trial registration Prospero record number: CRD42016048244.