Estimation of the area under the ROC curve

Forest fire susceptibility assessment under small sample scenario: A semi-supervised learning approach using transductive support vector machine

Forest fires threaten global ecosystems, socio-economic structures, and public safety. Accurately assessing forest fire susceptibility is critical for effective environmental management. Supervised learning methods dominate this assessment, relying on a substantial dataset of forest fire occurrences for model training. However, obtaining precise forest fire location data remains challenging. To address this issue, semi-supervised learning emerges as a viable solution, leveraging both a limited set of collected samples and unlabeled data containing environmental factors for training. Our study employed the transductive support vector machine (TSVM), a key semi-supervised learning method, to assess forest fire susceptibility in scenarios with limited samples. We conducted a comparative analysis, evaluating its performance against widely used supervised learning methods. The assessment area for forest fire susceptibility lies in Dayu County, Jiangxi Province, China, renowned for its vast forest cover and frequent fire incidents. We analyzed and generated maps depicting forest fire susceptibility, evaluating prediction accuracies for both supervised and semi-supervised learning methods across various small sample scenarios (e.g., 4, 8, 12, 16, 20, 24, 28, and 32 samples). Our findings indicate that TSVM exhibits superior prediction accuracy compared to supervised learning with limited samples, yielding more plausible forest fire susceptibility maps. For instance, at sample sizes of 4, 16, and 28, TSVM achieves prediction accuracies of approximately 0.8037, 0.9257, and 0.9583, respectively. In contrast, random forests, the top performers in supervised learning, demonstrate accuracies of approximately 0.7424, 0.8916, and 0.9431, respectively, for the same small sample sizes. Additionally, we discussed three key aspects: TSVM parameter configuration, the impact of unlabeled sample size, and performance within typical sample sizes. Our findings support semi-supervised learning as a promising approach compared to supervised learning for forest fire susceptibility assessment and mapping, particularly in scenarios with small sample sizes.

Rapid and non-invasive detection of cystic echinococcosis in sheep based on serum fluorescence spectrum combined with machine learning algorithms

Cystic echinococcosis (CE) is a grievous zoonotic parasitic disease. Currently, the traditional technology of screening CE is laborious and expensive, developing an innovative technology is urgent. In this study, we combined serum fluorescence spectroscopy with machine learning algorithms to develop an innovative screening technique to diagnose CE in sheep. Serum fluorescence spectra of Echinococcus granulosus sensu stricto-infected group (n = 63) and uninfected E. granulosus s.s. group (n = 60) under excitation at 405 nm were recorded. The linear support vector machine (Linear SVM), Quadratic SVM, medium radial basis function (RBF) SVM, K-nearest neighbor (KNN), and principal component analysis-linear discriminant analysis (PCA-LDA) were used to analyze the spectra data. The results showed that Quadratic SVM had the great classification capacity, its sensitivity, specificity, and accuracy were 85.0%, 93.8%, and 88.9%, respectively. In short, serum fluorescence spectroscopy combined with Quadratic SVM algorithm has great potential in the innovative diagnosis of CE in sheep.

Space-time data-driven modeling of precipitation-induced shallow landslides in South Tyrol, Italy

Shallow landslides represent potentially damaging processes in mountain areas worldwide. These geomorphic processes are usually caused by an interplay of predisposing, preparatory, and triggering environmental factors. At regional scales, data-driven methods have been used to model shallow landslides by addressing the spatial and temporal components separately. So far, few studies have explored the integration of space and time for landslide prediction. This research leverages generalized additive mixed models to develop an integrated approach to model shallow landslides in space and time. We built upon data on precipitation-induced landslide records from 2000 to 2020 in South Tyrol, Italy (7400 km2). The slope unit-based model predicts landslide occurrence as a function of static and dynamic factors while seasonal effects are incorporated. The model also accounts for spatial and temporal biases inherent in the underlying landslide data. We validated the resulting predictions through a suite of cross-validation techniques, obtaining consistent performance scores above 0.85. The analyses revealed that the best-performing model combines static ground conditions and two precipitation time windows: a short-term cumulative precipitation of 2 days before the landslide event and a medium-term cumulative precipitation of 14 days. We demonstrated the model's predictive capabilities by predicting the dynamic landslide probabilities over historical data associated with a heavy precipitation event on August 4th and August 5th, 2016, and hypothetical non-spatially explicit precipitation (what-if) scenarios. The novel approach shows the potential to integrate static and dynamic landslide factors for large areas, accounting for the underlying data structure and data limitations.

Establishing the validity of a diagnostic questionnaire for childbirth-related posttraumatic stress disorder

BACKGROUND

Labor and delivery can entail complications and severe maternal morbidities that threaten a woman's life or cause her to believe that her life is in danger. Women with these experiences are at risk for developing posttraumatic stress disorder. Postpartum posttraumatic stress disorder, or childbirth-related posttraumatic stress disorder, can become an enduring and debilitating condition. At present, validated tools for a rapid and efficient screen for childbirth-related posttraumatic stress disorder are lacking.

OBJECTIVE

We examined the diagnostic validity of the Posttraumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, for detecting posttraumatic stress disorder among women who have had a traumatic childbirth. This Checklist assesses the 20 Diagnostic and Statistical Manual of Mental Disorders, posttraumatic stress disorder symptoms and is a commonly used patient-administrated screening instrument. Its diagnostic accuracy for detecting childbirth-related posttraumatic stress disorder is unknown.

STUDY DESIGN

The sample included 59 patients who reported a traumatic childbirth experience determined in accordance with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, posttraumatic stress disorder criterion A for exposure involving a threat or potential threat to the life of the mother or infant, experienced or perceived, or physical injury. The majority (66%) of the participants were less than 1 year postpartum (for full sample: median, 4.67 months; mean, 1.5 years) and were recruited via the Mass General Brigham's online platform, during the postpartum unit hospitalization or after discharge. Patients were instructed to complete the Posttraumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, concerning posttraumatic stress disorder symptoms related to childbirth. Other comorbid conditions (ie, depression and anxiety) were also assessed. They also underwent a clinician interview for posttraumatic stress disorder using the gold-standard Clinician-Administered PTSD Scale for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. A second administration of the checklist was performed in a subgroup (n=43), altogether allowing an assessment of internal consistency, test-retest reliability, and convergent and diagnostic validity of the Checklist. The diagnostic accuracy of the Posttraumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, in reference to the Clinician-Administered PTSD Scale for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, was determined using the area under the receiver operating characteristic curve; an optimal cutoff score was identified using the Youden's J index.

RESULTS

One-third of the sample (35.59%) met the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, criteria for a posttraumatic stress disorder diagnosis stemming from childbirth. The Posttraumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, symptom severity score was strongly correlated with the Clinician-Administered PTSD Scale for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, total score (ρ=0.82; P<.001). The area under the receiver operating characteristic curve was 0.93 (95% confidence interval, 0.87-0.99), indicating excellent diagnostic performance of the Posttraumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. A cutoff value of 28 maximized the sensitivity (0.81) and specificity (0.90) and correctly diagnosed 86% of women. A higher value (32) identified individuals with more severe posttraumatic stress disorder symptoms (specificity, 0.95), but with lower sensitivity (0.62). Checklist scores were also stable over time (intraclass correlation coefficient, 0.73), indicating good test-retest reliability. Posttraumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, scores were moderately correlated with the depression and anxiety symptom scores (Edinburgh Postnatal Depression Scale: ρ=0.58; P<.001 and the Brief Symptom Inventory, anxiety subscale: ρ=0.51; P<.001).

CONCLUSION

This study demonstrates the validity of the Posttraumatic Stress Disorder Checklist for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, as a screening tool for posttraumatic stress disorder among women who had a traumatic childbirth experience. The instrument may facilitate screening for childbirth-related posttraumatic stress disorder on a large scale and help identify women who might benefit from further diagnostics and services. Replication of the findings in larger, postpartum samples is needed.

Detection of Pedestrians in Reverse Camera Using Multimodal Convolutional Neural Networks

In recent years, the application of artificial intelligence (AI) in the automotive industry has led to the development of intelligent systems focused on road safety, aiming to improve protection for drivers and pedestrians worldwide to reduce the number of accidents yearly. One of the most critical functions of these systems is pedestrian detection, as it is crucial for the safety of everyone involved in road traffic. However, pedestrian detection goes beyond the front of the vehicle; it is also essential to consider the vehicle's rear since pedestrian collisions occur when the car is in reverse drive. To contribute to the solution of this problem, this research proposes a model based on convolutional neural networks (CNN) using a proposed one-dimensional architecture and the Inception V3 architecture to fuse the information from the backup camera and the distance measured by the ultrasonic sensors, to detect pedestrians when the vehicle is reversing. In addition, specific data collection was performed to build a database for the research. The proposed model showed outstanding results with 99.85% accuracy and 99.86% correct classification performance, demonstrating that it is possible to achieve the goal of pedestrian detection using CNN by fusing two types of data.

Normal tissue complication probability of acute eyelids erythema following radiotherapy of head and neck cancers and skull-base tumors

PURPOSE

Radiation therapy is broadly used as one of the main treatment methods for patients with head and neck cancers and skull base tumors. However, it can lead to normal tissue complications. Therefore, this study aimed to model normal tissue complication probability (NTCP) of eyelid skin erythema after radiation therapy.

METHODS

The dataset of 45 patients with head and neck and skull base tumors was prospectively collected from their dose-volume histograms (DVHs). Grade 1 + eyelid skin erythema based on the Common Terminology Criteria for Adverse Events (CTCAE 4.0) was evaluated as the endpoint after a three-month follow-up. The Lyman-Kutcher-Burman (LKB) radiobiological model was developed based on generalized equivalent uniform dose (gEUD). Model parameters were calculated by maximum likelihood estimation. Model performance was evaluated by ROC-AUC, Brier score and Hosmer-Lemeshow test.

RESULTS

After three months of follow-up, 13.33% of patients experienced eyelids skin erythema grade 1 or more. The parameters of the LKB model were: TD50 = 30 Gy, m = 0.14, and n = 0.10. The model showed good predictive performance with ROC-AUC = 0.80 (CI:0.66-0.94) and a Brier score of 0.20.

CONCLUSIONS

In this study, NTCP of eyelid skin erythema was modeled based on the LKB radiobiological model with good predictive performance.

Sleep Loss the night before surgery and incidence of postoperative delirium in adults 65-95 years of age

STUDY OBJECTIVES

To describe the association between preoperative sleep disruption and postoperative delirium.

METHODS

Prospective cohort study with six time points (3 nights pre-hospitalization and 3 nights post-surgery). The sample included 180 English-speaking patients ≥65 years old scheduled for major non-cardiac surgery and anticipated minimum hospital stay of 3 days. Six days of wrist actigraphy recorded continuous movement to estimate wake and sleep minutes during the night from 22:00 to 05:59. Postoperative delirium was measured by a structured interview using the Confusion Assessment Method. Sleep variables for patients with (n = 32) and without (n = 148) postoperative delirium were compared using multivariate logistic regression.

RESULTS

Participants had a mean age of 72 ± 5 years (range 65-95 years). The incidence of postoperative delirium during any of the three postoperative days was 17.8%. Postoperative delirium was significantly associated with surgery duration (OR = 1.49, 95% CI 1.24-1.83) and sleep loss >15% on the night before surgery (OR = 2.64, 95% CI 1.10-6.62). Preoperative symptoms of pain, anxiety and depression were unrelated to preoperative sleep loss.

CONCLUSIONS

In this study of adults ≥65 years of age, short sleep duration was more severe preoperatively in the patients who experienced postoperative delirium as evidenced by sleep loss >15% of their normal night's sleep. However, we were unable to identify potential reasons for this sleep loss. Further investigation should include additional factors that may be associated with preoperative sleep loss to inform potential intervention strategies to mitigate preoperative sleep loss and reduce risk of postoperative delirium.

A classification model for detection of ductal carcinoma in situ by Fourier transform infrared spectroscopy based on deep structured semantic model

At present, deep learning is widely used in spectral data processing. Deep learning requires a large amount of data for training, while the collection of biological serum spectra is limited by sample numbers and labor costs, so it is impractical to obtain a large amount of serum spectral data for disease detection. In this study, we propose a spectral classification model based on the deep structured semantic model (DSSM) and successfully apply it to Fourier Transform Infrared (FT-IR) spectroscopy for ductal carcinoma in situ (DCIS) detection. Compared with the traditional deep learning model, we match the spectral data into positive and negative pairs according to whether the spectra are from the same category. The DSSM structure is constructed by extracting features according to the spectral similarity of spectra pairs. This new construction model increases the data amount used for model training and reduces the dimension of spectral data. Firstly, the FT-IR spectra are paired. The spectra pairs are labeled as positive pairs if they come from the same category, and the spectra pairs are labeled as negative pairs if they come from different categories. Secondly, two spectra in each spectra pair are put into two deep neural networks of the DSSM structure separately. Then the spectral similarity between the output feature maps of two deep neural networks is calculated. The DSSM structure is trained by maximizing the conditional likelihood of the spectra pairs from the same category. Thirdly, after the training of DSSM is done, the training set and testing set are input into two deep neural networks separately. The output feature maps of the training set are put into the reference library. Lastly, the k-nearest neighbor (KNN) model is used for classification according to Euclidean distances between the output feature map of each unknown sample to the reference library. The category of the unknown sample is judged according to the categories of k nearest samples. We also use principal component analysis (PCA) to reduce dimension for comparison. The accuracies of the KNN model, principal component analysis-k nearest neighbor (PCA-KNN) model, and deep structured semantic model-k nearest neighbor (DSSM-KNN) model are 78.8%, 72.7%, and 97.0%, which proves that our proposed model has higher accuracy.

Rapid Diagnosis of Ductal Carcinoma In Situ and Breast Cancer Based on Raman Spectroscopy of Serum Combined with Convolutional Neural Network

Ductal carcinoma in situ (DCIS) and breast cancer are common female breast diseases and pose a serious health threat to women. Early diagnosis of breast cancer and DCIS can help to develop targeted treatment plans in time. In this paper, we investigated the feasibility of using Raman spectroscopy combined with convolutional neural network (CNN) to discriminate between healthy volunteers, breast cancer and DCIS patients. Raman spectra were collected from the sera of 241 healthy volunteers, 463 breast cancer and 100 DCIS patients, and a total of 804 spectra were recorded. The pre-processed Raman spectra were used as the input of CNN to establish a model to classify the three different spectra. After using cross-validation to optimize its hyperparameters, the model's final classification performance was assessed using an unknown test set. For comparison with other machine learning algorithms, we additionally built models using support vector machine (SVM), random forest (RF) and k-nearest neighbor (KNN) methods. The final accuracies for CNN, SVM, RF and KNN were 98.76%, 94.63%, 80.99% and 78.93%, respectively. The values for area under curve (AUC) were 0.999, 0.994, 0.931 and 0.900, respectively. Therefore, our study results demonstrate that CNN outperforms three traditional algorithms in terms of classification performance for Raman spectral data and can be a useful auxiliary diagnostic tool of breast cancer and DCIS.

A network approach to compute hypervolume under receiver operating characteristic manifold for multi-class biomarkers

Computation of hypervolume under ROC manifold (HUM) is necessary to evaluate biomarkers for their capability to discriminate among multiple disease types or diagnostic groups. However the original definition of HUM involves multiple integration and thus a medical investigation for multi-class receiver operating characteristic (ROC) analysis could suffer from huge computational cost when the formula is implemented naively. We introduce a novel graph-based approach to compute HUM efficiently in this article. The computational method avoids the time-consuming multiple summation when sample size or the number of categories is large. We conduct extensive simulation studies to demonstrate the improvement of our method over existing R packages. We apply our method to two real biomedical data sets to illustrate its application.

Are better existing WASH practices in urban slums associated with a lower long-term risk of severe cholera? A prospective cohort study with 4 years of follow-up in Mirpur, Bangladesh

OBJECTIVE

To investigate the association between existing household water quality, sanitation and hygiene (WASH) practices and severe cholera risk in a dense urban slum where cholera is highly endemic.

DESIGN, SETTING AND PARTICIPANTS

We assembled a large prospective cohort within a cluster randomised trial evaluating the effectiveness of oral cholera vaccine. Our dynamic cohort population (n=193 576) comprised individuals living in the 'non-intervention' clusters of the trial, and were followed over 4 years. This study was conducted in a dense urban slum community of Dhaka, Bangladesh and cholera surveillance was undertaken in 12 hospitals serving the study area.

PRIMARY OUTCOME MEASURE

First severe cholera episode detected during follow-up period.

METHODS

We applied a machine learning algorithm on a training subpopulation (n=96 943) to develop a binary ('better', 'not better') composite WASH variable predictive of severe cholera. The WASH rule was evaluated for performance in a separate validation subpopulation (n=96 633). Afterwards, we used Cox regression models to evaluate the association between 'better' WASH households and severe cholera risk over 4 years in the entire study population.

RESULTS

The 'better' WASH rule found that water quality and access were the most significant factors associated with severe cholera risk. Members of 'better' WASH households, constituting one-third of the population, had a 47% reduced risk of severe cholera (95% CI: 29 to 69; p<0.001), after adjusting for covariates. The protective association between living in a 'better' WASH household and severe cholera persisted in all age groups.

CONCLUSIONS

Salutary existing household WASH practices were associated with a significantly reduced long-term risk of severe cholera in an urban slum of Dhaka. These findings suggest that WASH adaptations already practised in the community may be important for developing and implementing effective and sustainable cholera control programmes in similar settings.

TRIAL REGISTRATION NUMBER

This article is a re-analysis of data from a cluster randomized trial; can be found on ClinicalTrials.gov NCT01339845.

On optimal biomarker cutoffs accounting for misclassification costs in diagnostic trilemmas with applications to pancreatic cancer

Pancreatic ductal adenocarcinoma (PDAC) is the most deadly cancer and currently there is strong clinical interest in novel biomarkers that contribute to its early detection. Assessing appropriately the accuracy of such biomarkers is a crucial issue and often one needs to take into account that many assays include biospecimens of individuals coming from three groups: healthy, chronic pancreatitis, and PDAC. The ROC surface is an appropriate tool for assessing the overall accuracy of a marker employed under such trichotomous settings. A decision/classification rule is often based on the so-called Youden index and its three-dimensional generalization. However, both the clinical and the statistical literature have not paid the necessary attention to the underlying false classification (FC) rates that are of equal or even greater importance. In this article we provide a framework to make inferences around all classification rates as well as comparisons. We explore the trinormal model, flexible models based on power transformations, and robust non-parametric alternatives. We provide a full framework for the construction of confidence intervals, regions, and spaces for joint inferences or for clinically meaningful points of interest. We further discuss the implications of costs related to different FCs. We evaluate our approaches through extensive simulations and illustrate them using data from a recent PDAC study conducted at the MD Anderson Cancer Center.

Rapidly detecting fennel origin of the near-infrared spectroscopy based on extreme learning machine

Fennel contains many antioxidant and antibacterial substances, and it has very important applications in food flavoring and other fields. The kinds and contents of chemical substances in fennel vary from region to region, which can affect the taste and efficacy of the fennel and its derivatives. Therefore, it is of great significance to accurately classify the origin of the fennel. Recently, origin detection methods based on deep networks have shown promising results. However, the existing methods spend a relatively large time cost, a drawback that is fatal for large amounts of data in practical application scenarios. To overcome this limitation, we explore an origin detection method that guarantees faster detection with classification accuracy. This research is the first to use the machine learning algorithm combined with the Fourier transform-near infrared (FT-NIR) spectroscopy to realize the classification and identification of the origin of the fennel. In this experiment, we used Rubberband baseline correction on the FT-NIR spectral data of fennel (Yumen, Gansu and Turpan, Xinjiang), using principal component analysis (PCA) for data dimensionality reduction, and selecting extreme learning machine (ELM), Convolutional Neural Network (CNN), recurrent neural network (RNN), Transformer, generative adversarial networks (GAN) and back propagation neural network (BPNN) classification model of the company realizes the classification of the sample origin. The experimental results show that the classification accuracy of ELM, RNN, Transformer, GAN and BPNN models are above 96%, and the ELM model using the hardlim as the activation function has the best classification effect, with an average accuracy of 100% and a fast classification speed. The average time of 30 experiments is 0.05 s. This research shows the potential of the machine learning algorithm combined with the FT-NIR spectra in the field of food production area classification, and provides an effective means for realizing rapid detection of the food production area, so as to merchants from selling shoddy products as good ones and seeking illegal profits.

Machine Learning for Predicting Discharge Disposition After Traumatic Brain Injury

BACKGROUND

Current traumatic brain injury (TBI) prognostic calculators are commonly used to predict the mortality and Glasgow Outcome Scale, but these outcomes are most relevant for severe TBI. Because mild and moderate TBI rarely reaches severe outcomes, there is a need for novel prognostic endpoints.

OBJECTIVE

To generate machine learning (ML) models with a strong predictive capacity for trichotomized discharge disposition, an outcome not previously used in TBI prognostic models. The outcome can serve as a proxy for patients' functional status, even in mild and moderate patients with TBI.

METHODS

Using a large data set (n = 5292) of patients with TBI from a quaternary care center and 84 predictors, including vitals, demographics, mechanism of injury, initial Glasgow Coma Scale, and comorbidities, we trained 6 different ML algorithms using a nested-stratified-cross-validation protocol. After optimizing hyperparameters and performing model selection, isotonic regression was applied to calibrate models.

RESULTS

When maximizing the microaveraged area under the receiver operating characteristic curve during hyperparameter optimization, a random forest model exhibited top performance. A random forest model was also selected when maximizing the microaveraged area under the precision-recall curve. For both models, the weighted average area under the receiver operating characteristic curves was 0.84 (95% CI 0.81-0.87) and the weighted average area under the precision-recall curves was 0.85 (95% CI 0.82-0.88).

CONCLUSION

Our group presents high-performing ML models to predict trichotomized discharge disposition. These models can assist in optimization of patient triage and treatment, especially in cases of mild and moderate TBI.

Predicting Student Outcomes in Online Courses Using Machine Learning Techniques: A Review

Recent years have witnessed an increased interest in online education, both massive open online courses (MOOCs) and small private online courses (SPOCs). This significant interest in online education has raised many challenges related to student engagement, performance, and retention assessments. With the increased demands and challenges in online education, several researchers have investigated ways to predict student outcomes, such as performance and dropout in online courses. This paper presents a comprehensive review of state-of-the-art studies that examine online learners’ data to predict their outcomes using machine and deep learning techniques. The contribution of this study is to identify and categorize the features of online courses used for learners’ outcome prediction, determine the prediction outputs, determine the strategies and feature extraction methodologies used to predict the outcomes, describe the metrics used for evaluation, provide a taxonomy to analyze related studies, and provide a summary of the challenges and limitations in the field.

On estimating the area under the ROC curve in ranked set sampling

In medical research, the receiver operating characteristic curve is widely used to evaluate accuracy of a continuous biomarker. The area under this curve is known as an index for overall performance of the biomarker. This article develops three new estimators of the area under the receiver operating characteristic curve in ranked set sampling. The first estimator is obtained under normality assumption. The two other estimators are constructed by applying a Box-Cox transformation on data, and then using either a parametric estimator or a kernel-density-based estimator. A simulation study is carried out to compare the proposed estimators with those available in the literature. It emerges that the new estimators offer some advantages in specific situations. Application of the methods is demonstrated using real data in the context of medicine.

Neural mechanisms of the mood effects on third‐party responses to injustice after unfair experiences

Behavioral decision theory argues that humans can adjust their third‐party responses (e.g., punishment and compensation) to injustice by integrating unfair experiences. Typically, the mood plays an important role in such a decision‐making process. However, the underlying neurocognitive bases remain largely unclear. We first employ a modified third‐party justice game in which an allocator split an amount of money between oneself and a receiver. The participants can reapportion the money as observers by choosing from the following three costly options: compensate the receiver, accept the current allocation, or punish the allocator. Then, a second‐party pseudo interaction is conducted where participants receive more (i.e., advantageous unfair experience) or less (i.e., disadvantageous unfair experience) than others. Finally, participants perform the third‐party justice game again after unfair experiences. Here, we use functional near‐infrared spectroscopy (fNIRS) to measure participants' brain activities during third‐party responses to injustice. We find participants compensate more to the receiver after advantageous unfair experience, which involved enhanced positive emotion, weakened sense of unfairness, and is linked with increased activity in the right dorsolateral prefrontal cortex (rDLPFC). In contrast, participants punish more on the allocator after disadvantageous unfair experience, which might primarily stem from their negative emotional responses, strong sense of unfairness, and is associated with significantly decreased activity in the rDLPFC. Our results suggest that third‐party compensation and punishment involved differential psychological and neural bases. Our findings highlight the crucial roles of second‐party unfair experiences and the corresponding mood responses in third‐party responses to unfairness, and unravel the intermediate neural architecture.

A Stepwise Algorithm for Linearly Combining Biomarkers under Youden Index Maximization

Combining multiple biomarkers to provide predictive models with a greater discriminatory ability is a discipline that has received attention in recent years. Choosing the probability threshold that corresponds to the highest combined marker accuracy is key in disease diagnosis. The Youden index is a statistical metric that provides an appropriate synthetic index for diagnostic accuracy and a good criterion for choosing a cut-off point to dichotomize a biomarker. In this study, we present a new stepwise algorithm for linearly combining continuous biomarkers to maximize the Youden index. To investigate the performance of our algorithm, we analyzed a wide range of simulated scenarios and compared its performance with that of five other linear combination methods in the literature (a stepwise approach introduced by Yin and Tian, the min-max approach, logistic regression, a parametric approach under multivariate normality and a non-parametric kernel smoothing approach). The obtained results show that our proposed stepwise approach showed similar results to other algorithms in normal simulated scenarios and outperforms all other algorithms in non-normal simulated scenarios. In scenarios of biomarkers with the same means and a different covariance matrix for the diseased and non-diseased population, the min-max approach outperforms the rest. The methods were also applied on two real datasets (to discriminate Duchenne muscular dystrophy and prostate cancer), whose results also showed a higher predictive ability in our algorithm in the prostate cancer database.

A Novel Smartphone-Based Color Test for Detection of Color Vision Defects in Age Related Macular Degeneration

Purpose

To evaluate the efficacy of the smartphone-based K-color test to detect color defects in patients with Age-related Macular Degeneration (AMD).

Methods

88 patients (n = 135 eyes) with AMD and 28 controls (n = 53 eyes) underwent color testing with the Hardy–Rand–Rittler (H-R-R), the K-color test, and the Ishihara test. The K-color test presents randomized colored shapes in decreasing steps of intensity, providing also a record system for result tele-transmission. Sensitivity, specificity, and reliability were examined to investigate the validity of the novel test. 26 participants with AMD also completed a questionnaire regarding the feasibility of the test.

Results

Linear mixed-effects models indicated a significant difference (p < 0.001) between AMD and normal eyes. The areas under the curve (AUC) were estimated to be 0.897 [95% CI: 0.841–0.952], 0.943 [95% CI: 0.901–0.984], and 0.931 [95% CI: 0.886–0.977] for the red, green, and blue color, respectively. Based on the H-R-R, the sensitivity of the test was 0.79, 0.90, and 0.95 for the red, green, and blue colors, respectively, and specificity was 0.88 for all colors. The new test recognized more abnormal cases than the Ishihara (sensitivity of 0.98 and 1.0 and specificity of 0.48 and 0.38 for red and green colors, respectively). Test-retest reliability was found to be high for the red [ICC = 0.996 (0.990–0.999)], green [ICC = 0.974 (0.929–0.990)], and blue [ICC = 0.992 (0.981–0.997)] colors. The majority of the asked participants stated that they could easily perform the test.

Conclusion

The K-color test was found to be sensitive and specific in detecting color defects in AMD patients. The K-color test may serve as a useful tool both for patients and their physicians.

Application of the skew exponential power distribution to ROC curves

The bi-Normal ROC model and corresponding metrics are commonly used in medical studies to evaluate the discriminatory ability of a biomarker. However, in practice, many clinical biomarkers tend to have skewed or other non-Normal distributions. And while the bi-Normal ROC model's AUC tends to be unbiased in this setting, providing a reasonable measure of global performance, the corresponding decision thresholds tend to be biased. To correct this bias, we propose using an ROC model based on the skew exponential power (SEP) distribution, whose additional parameters can accommodate skewed, heavy tailed, or other non-Normal distributions. Additionally, the SEP distribution can be used to evaluate whether the bi-Normal model would be appropriate. The performance of these ROC models and the non-parametric approach are evaluated via a simulation study and applied to a real data set involving infections from Klebsiella pneumoniae. The SEP based ROC-model provides some efficiency gains with respect to estimation of the AUC and provides cut-points with improved classification rates. As such, in the presence non-Normal data, we suggest using the proposed SEP ROC model.

The Risk of Salt Reduction in Dry-Cured Sausage Assessed by the Influence on Water Activity and the Survival of Salmonella

Water activity (aw) is the main hurdle for microbial control in dry-cured sausages. The aw can be influenced by drying or adding electrolytes or humectants. Dry-cured meat products are partially dried, which, together with added salt, results in safe aw values. Currently, there is a trend to reduce salt in meat products, which can compromise the preservation process. The present work aims to evaluate the influences of added salt levels (1% or 3%) and the use or omission of phosphates and wine on the aw of a dry-cured sausage, and to evaluate the possibility of estimating the aw from the moisture loss and the behavior of Salmonella during dry-cured sausage (chouriço) processing. There was a strong relationship between moisture and aw, regardless of the salt level and the presence of phosphates or wine. Predicting aw from moisture loss is possible using the Boltzmann sigmoid function. The salt level strongly influences Salmonella behavior, mainly through aw reduction. An increase in aw by 0.01 units reduced the odds of achieving a 5-log reduction in Salmonella counts to half. Increasing added salt from 1% to 3% increased the odds of achieving a 5-log Salmonella reduction 7.5-fold. The current trend to reduce salt in foods must be carefully approached if applied to cured meat products, as it has substantial consequences on aw evolution and Salmonella survival.

Inference on the overlap coefficient: The binormal approach and alternatives

The overlap coefficient (OVL) measures the similarity between two distributions through the overlapping area of their distribution functions. Given its intuitive description and ease of visual representation by the straightforward depiction of the amount of overlap between the two corresponding histograms based on samples of measurements from each one of the two distributions, the development of accurate methods for confidence interval construction can be useful for applied researchers. The overlap coefficient has received scant attention in the literature since it lacks readily available software for its implementation, while inferential procedures that can cover the whole range of distributional scenarios for the two underlying distributions are missing. Such methods, both parametric and non-parametric are developed in this article, while R-code is provided for their implementation. Parametric approaches based on the binormal model show better performance and are appropriate for use in a wide range of distributional scenarios. Methods are assessed through a large simulation study and are illustrated using a dataset from a study on human immunodeficiency virus-related cognitive function assessment.

Melanoma Recognition by Fusing Convolutional Blocks and Dynamic Routing between Capsules

Simple Summary

The early treatment of skin cancer can effectively reduce mortality rates. Recently, automatic melanoma diagnosis from skin images has gained attention, which was mainly encouraged by the well-known challenge developed by the International Skin Imaging Collaboration project. The majority of contestant submitted Convolutional Neural Network based solutions. However, this type of model presents disadvantages. As a consequence, Dynamic Routing between Capsules has been proposed to overcome such limitations. The aim of our proposal was to assess the advantages of combining both architectures. An extensive experimental study showed the proposal significantly outperformed state-of-the-art models, achieving 166% higher predictive performance compared to ResNet in non-dermoscopic images. In addition, the pixels activated during prediction were shown, which allows to assess the rationale to give such a conclusion. Finally, more research should be conducted in order to demonstrate the potential of this neural network architecture in other areas.

Abstract

Skin cancer is one of the most common types of cancers in the world, with melanoma being the most lethal form. Automatic melanoma diagnosis from skin images has recently gained attention within the machine learning community, due to the complexity involved. In the past few years, convolutional neural network models have been commonly used to approach this issue. This type of model, however, presents disadvantages that sometimes hamper its application in real-world situations, e.g., the construction of transformation-invariant models and their inability to consider spatial hierarchies between entities within an image. Recently, Dynamic Routing between Capsules architecture (CapsNet) has been proposed to overcome such limitations. This work is aimed at proposing a new architecture which combines convolutional blocks with a customized CapsNet architecture, allowing for the extraction of richer abstract features. This architecture uses high-quality 299×299×3 skin lesion images, and a hyper-tuning of the main parameters is performed in order to ensure effective learning under limited training data. An extensive experimental study on eleven image datasets was conducted where the proposal significantly outperformed several state-of-the-art models. Finally, predictions made by the model were validated through the application of two modern model-agnostic interpretation tools.

Predicting Unplanned Health Care Utilization and Cost: Comparing Patient-reported Outcomes Measurement Information System and Claims

OBJECTIVES

There is little literature describing if and how payers are utilizing patient-reported outcomes to predict future costs. This study assessed if Patient-reported Outcomes Measurement Information System (PROMIS) domain scores, collected in routine practice at neurology clinics, improved payer predictive models for unplanned care utilization and cost.

STUDY DESIGN

Retrospective cohort analysis of private Health Plan-insured patients with visits at 18 Health Plan-affiliated neurology clinics.

METHODS

PROMIS domains (Anxiety v1.0, Cognitive Function Abilities v2.0, Depression v1.0, Fatigue v1.0, Pain Interference v1.0, Physical Function v2.0, Sleep Disturbance v1.0, and Ability to Participate in Social Roles and Activities v2.0) are collected as part of routine care. Data from patients' first PROMIS measures between June 27, 2018 and April 16, 2019 were extracted and combined with claims data. Using (1) claims data alone and (2) PROMIS and claims data, we examined the association of covariates to utilization (using a logit model) and cost (using a generalized linear model). We evaluated model fit using area under the receiver operating characteristic curve (for unplanned care utilization), akaike information criterion (for unplanned care costs), and sensitivity and specificity in predicting top 15% of unplanned care costs.

RESULTS

Area under the receiver operating curve values were slightly higher, and akaike information criterion values were similar, for PROMIS plus claims covariates compared with claims alone. The PROMIS plus claims model had slightly higher sensitivity and equivalent specificity compared with claims-only models.

CONCLUSION

One-time PROMIS measure data combined with claims data slightly improved predictive model performance compared with claims alone, but likely not to an extent that indicates improved practical utility for payers.

Accuracy of Dual-Energy Computed Tomography Techniques for Fat Quantification in Comparison With Magnetic Resonance Proton Density Fat Fraction and Single-Energy Computed Tomography in an Anthropomorphic Phantom Environment

OBJECTIVE

To investigate in an anthropomorphic phantom study the accuracy of dual-energy computed tomography (DECT) techniques for fat quantification in comparison with magnetic resonance (MR) proton density fat fraction (PDFF) and single-energy computed tomography (SECT), using known fat content as reference standard.

METHODS

Between August 2018 and November 2020, organic material-based cylinders, composed of mixtures of lean and fat tissues mimics, iodine, and iron, were constructed to simulate varying fat content levels (0%, 10%, 15%, 25%, 50%, 75%, and 100%) in a parenchymal organ and were embedded into an anthropomorphic phantom simulating 3 patient sizes (circumference, 91, 126, and 161 cm). The phantom was imaged with multiecho MR, DECT, and SECT. Magnetic resonance PDFF, DECT fat fraction, and computed tomography (CT) numbers (SECT polychromatic and DECT monochromatic data, virtual unenhanced images) were estimated. Performances of MR PDFF and CT techniques to detect differences in fat content were measured using the area under the curve (AUC). Noninferiority of each CT technique relative to MR PDFF was tested using a noninferiority margin of -0.1.

RESULTS

MR PDFF, DECT 140 keV monochromatic data, and fat fraction most closely correlated with known fat content (R2 = 0.98, 0.98, and 0.96, respectively). Unlike SECT and all other DECT techniques, DECT fat fraction was not affected by presence of iodine (mean difference, 0.3%; 95% confidence interval [CI], -0.9% to 1.5%). Dual-energy computed tomography fat fraction showed noninferiority to MR PDFF in detecting differences of 5% in fat content in medium-sized phantoms (ΔAUC, -0.05; 95% CI, -0.08 to -0.01), and 7% in large (ΔAUC, -0.04; 95% CI, -0.0 to 0.00) or extralarge sized phantoms (ΔAUC, -0.02; 95% CI, -0.07 to 0.00).

CONCLUSIONS

Dual-energy computed tomography fat fraction shows linear correlation with true fat content in the range up to 50% fat fraction. Dual-energy computed tomography fat fraction has comparable estimation error and shows noninferiority to MR PDFF in detecting small differences in fat content across different body sizes.

Statistical inference for the difference between two maximized Youden indices obtained from correlated biomarkers

Currently, there is global interest in deriving new promising cancer biomarkers that could complement or substitute the conventional ones. Clinical decisions can often be based on the cutoff that corresponds to the maximized Youden index when maximum accuracy drives decisions. When more than one classification criteria are measured within the same individuals, correlated measurements arise. In this work, we propose hypothesis tests and confidence intervals for the comparison of two correlated receiver operating characteristic (ROC) curves in terms of their corresponding maximized Youden indices. We explore delta-based techniques under parametric assumptions, or power transformations. Nonparametric kernel-based methods are also examined. We evaluate our approaches through simulations and illustrate them using data from a metabolomic study referring to the detection of pancreatic cancer.

A multi-hazard map-based flooding, gully erosion, forest fires, and earthquakes in Iran

We used three state-of-the-art machine learning techniques (boosted regression tree, random forest, and support vector machine) to produce a multi-hazard (MHR) map illustrating areas susceptible to flooding, gully erosion, forest fires, and earthquakes in Kohgiluyeh and Boyer-Ahmad Province, Iran. The earthquake hazard map was derived from a probabilistic seismic hazard analysis. The mean decrease Gini (MDG) method was implemented to determine the relative importance of effective factors on the spatial occurrence of each of the four hazards. Area under the curve (AUC) plots, based on a validation dataset, were created for the maps generated using the three algorithms to compare the results. The random forest model had the highest predictive accuracy, with AUC values of 0.994, 0.982, and 0.885 for gully erosion, flooding, and forest fires, respectively. Approximately 41%, 40%, 28%, and 3% of the study area are at risk of forest fires, earthquakes, floods, and gully erosion, respectively.

Can Existing Improvements of Water, Sanitation, and Hygiene (WASH) in Urban Slums Reduce the Burden of Typhoid Fever in These Settings?

BACKGROUND

Sustained investments in water, sanitation, and hygiene (WASH) have lagged in resource-poor settings; incremental WASH improvements may, nonetheless, prevent diseases such as typhoid in disease-endemic populations.

METHODS

Using prospective data from a large cohort in urban Kolkata, India, we evaluated whether baseline WASH variables predicted typhoid risk in a training subpopulation (n = 28 470). We applied a machine learning algorithm to the training subset to create a composite, dichotomous (good, not good) WASH variable based on 4 variables, and evaluated sensitivity and specificity of this variable in a validation subset (n = 28 470). We evaluated in Cox regression models whether residents of "good" WASH households experienced a lower typhoid risk after controlling for potential confounders. We constructed virtual clusters (radius 50 m) surrounding each household to evaluate whether a prevalence of good WASH practices modified the typhoid risk in central household members.

RESULTS

Good WASH practices were associated with protection in analyses of all households (hazard ratio [HR] = 0.57; 95% confidence interval [CI], .37-.90; P = .015). This protection was evident in persons ≥5 years old at baseline (HR = 0.47; 95% CI, .34-.93; P = .005) and was suggestive, though not statistically significant, in younger age groups (HR = 0.61; 95% CI, .27-1.38; P = .235). The level of surrounding household good WASH coverage was also associated with protection (HR = 0.988; 95% CI, .979-.996; P = .004, for each percent coverage increase). However, collinearity between household WASH and WASH coverage prevented an assessment of their independent predictive contributions.

CONCLUSIONS

In this typhoid-endemic setting, natural variation in household WASH was associated with typhoid risk. If replicated elsewhere, these findings suggest that WASH improvements may enhance typhoid control, short of major infrastructural investments.

Joint application of biochemical markers and imaging techniques in the accurate and early detection of glioblastoma

Glioblastoma is a primary brain tumor with the most metastatic effect in adults. Despite the wide range of multidimensional treatments, tumor heterogeneity is one of the main causes of tumor spread and gives great complexity to diagnostic and therapeutic methods. Therefore, featuring noble noninvasive prognostic methods that are focused on glioblastoma heterogeneity is perceived as an urgent need. Imaging neuro-oncological biomarkers including MGMT (O⁶-methylguanine-DNA methyltransferase) promoter methylation status, tumor grade along with other tumor characteristics and demographic features (e.g., age) are commonly referred to during diagnostic, therapeutic and prognostic processes. Therefore, the use of new noninvasive prognostic methods focused on glioblastoma heterogeneity is considered an urgent need. Some neuronal biomarkers, including the promoter methylation status of the promoter MGMT, the characteristics and grade of the tumor, along with the patient's demographics (such as age and sex) are involved in diagnosis, treatment, and prognosis. Among the wide array of imaging techniques, magnetic resonance imaging combined with the more physiologically detailed technique of H-magnetic resonance spectroscopy can be useful in diagnosing neurological cancer patients. In addition, intracranial tumor qualitative analysis and sometimes tumor biopsies help in accurate diagnosis. This review summarizes the evidence for biochemical biomarkers being a reliable biomarker in the early detection and disease management in GBM. Moreover, we highlight the correlation between Imaging techniques and biochemical biomarkers and ask whether they can be combined.

Confirming the factor structure and improving the screening function of the Medical Fear Survey - short in a Hungarian community sample

Medical Fear Survey (MFS) was designed to assess the fear of medical treatments and related factors through five dimensions including fears of injections and blood draws, sharp objects, blood, mutilation, medical examination and physical symptoms. We analysed the factor structure and validity of a short version of MFS in Hungarian, on a large sample (2631 participants; 558 men, 2067 women) focusing on possible age and gender differences (aged Mean = 30.4, SD = 13.4), which were not reported for the original version. Furthermore, using discriminant analysis, potential screening function of the MFS-short was examined. Results supported construct and convergent validity and scale-reliability for the five-factor structure of the MFS-short. Further analyses demonstrated excellent discriminatory power for four subscales, while one subscale had acceptable power. Our findings provide implications for the utility of MFS-short as a screening measure in assessing the severity of medical fears controlling for gender differences and age-biases.

SARS-CoV-2 antibody seroprevalence in NHS healthcare workers in a large double-sited UK hospital

We determined the seroprevalence of SARS-CoV-2 antibodies in NHS healthcare workers (HCWs) in a cross-sectional study from a large general hospital located in a double-sited rural and semi-rural area. The sample size of 3,119 HCWs (mean age 43±13) consisted of 75.2% women, 61.1% White individuals and predominantly (62.4%) asymptomatic individuals. Seroprevalence of SARS-CoV-2 antibodies was 19.7%. Determinants of seropositivity were preceding symptomatic infection and non-White ethnicity. Regardless of staff role or sex, multivariate regression analysis revealed that non-White HCWs were three times (odds ratio [OR] 3.12, 95% confidence interval [CI] 2.53-3.86, P<0.001) more likely to have antibodies than White staff, and seven times (OR 7.10, 95% CI 5.72-8.87, P<0.001) more likely if there was a history of preceding symptoms. We report relatively high rates of seropositivity in all NHS healthcare workers. Non-White symptomatic HCWs were significantly more likely to be seropositive than their colleagues, independent of age, sex or staff role.

A probabilistic modeling approach for interpretable data inference and classification

In this paper, we propose a new probabilistic modeling approach for interpretable inference and classification using the maximum likelihood evidential reasoning (MAKER) framework. This approach integrates statistical analysis, hybrid evidence combination and belief rule-based (BRB) inference, and machine learning. Statistical analysis is used to acquire evidence from data. The BRB inference is applied to analyze the relationship between system inputs and outputs. An interdependence index is used to quantify the interdependence between input variables. An adapted genetic algorithm is applied to train the models. The model established by the approach features a unique strong interpretability, which is reflected in three aspects: (1) interpretable evidence acquisition, (2) interpretable inference mechanism, and (3) interpretable parameters determination. The MAKER-based model is shown to be a competitive classifier for the Banana, Haberman’s survival, and Iris data set, and generally performs better than other interpretable classifiers, e.g., complex tree, logistic regression, and naive Bayes.

Time to positivity of Klebsiella pneumoniae in blood culture as prognostic indicator for pediatric bloodstream infections

The aim of this study is to explore the prognostic values and optimal cutoff point of time to positivity (TTP) of blood culture in children with Klebsiella pneumoniae (K. pneumoniae) bloodstream infection. Ninety-four children with K. pneumoniae bloodstream infection hospitalized in Children's Hospital of Chongqing Medical University from April 2014 to January 2019 were enrolled retrospectively. TTP and risk factors were determined and analyzed by receiver operating characteristic (ROC) analysis and logistic regression analysis. The standard cutoff point of TTP was 13 h. Patients in early TTP (≤ 13 h) group had significantly higher in-hospital mortality (37.93% vs 6.15%, P = 0.000), higher incidence of septic shock (44.83% vs 6.15%, P = 0.000), higher proportion of PRISM III scores ≥ 10 (48.28% vs 20.00%, P = 0.005), and higher proportion of hypoalbuminemia on admission (44.83% vs 18.46%, P = 0.008). Multivariate analysis indicated PRISM III scores ≥ 10, early TTP, and hypoalbuminemia on admission were independent risk factors of in-hospital mortality (OR 8.36, 95% CI 1.80-38.92, P = 0.007; OR 5.85, 95% CI 1.33-25.61, P = 0.019; OR 5.73, 95% CI 1.30-25.22, P = 0.021, respectively) and septic shock (OR 14.04, 95% CI 2.63-75.38, P = 0.002; OR 11.26, 95% CI 2.10-60.22, P = 0.005; OR 10.27, 95% CI 2.01-52.35, P = 0.005, respectively).Conclusion: Early TTP (TTP ≤ 13 h), PRISM III scores ≥ 10, and hypoalbuminemia on admission appeared to be associated with worse outcomes for K. pneumoniae bloodstream infection children. What is Known: • Klebsiella pneumoniae bloodstream infection is an important cause of infectious disease morbidity and mortality worldwide in children. • Short duration of time to positivity indicated poor clinical outcomes. What is New: • Time to positivity ≤ 13 h, along with PRISM III scores ≥ 10 and hypoalbuminemia on admission, indicated higher in-hospital mortality and incidence of septic shock in Klebsiella pneumoniae bloodstream infection children. • The cut-off point of TTP in this pediatric study was much longer than that reported in adult patients.

An efficient variance estimator of AUC and its applications to binary classification

The area under the ROC (receiver operating characteristic) curve, AUC, is one of the most commonly used measures to evaluate the performance of a binary classifier. Due to sampling variation, the model with the largest observed AUC score is not necessarily optimal, so it is crucial to assess the variation of AUC estimate. We extend the proposal by Wang and Lindsay and devise an unbiased variance estimator of AUC estimate that is of a two-sample U-statistic form. The proposal can be easily generalized to estimate the variance of a K-sample U-statistic (K ≥ 2). To make our developed variance estimator more applicable, we employ a partition-resampling scheme that is computationally efficient. Simulation studies suggest that the developed AUC variance estimator yields much better or comparable performance to jackknife and bootstrap variance estimators, and computational times that are about 10 to 30 times faster than the times of its counterparts. In practice, the proposal can be used in the one-standard-error rule for model selection, or to construct an asymptotic confidence interval of AUC in binary classification. In addition to conducting simulation studies, we illustrate its practical applications using two real datasets in medical sciences.

Prognostic role of time to positivity of blood culture in children with Pseudomonas aeruginosa bacteremia

BACKGROUND

Pseudomonas aeruginosa (P. aeruginosa) is a major Gram-negative pathogen, which has been reported to result in high mortality. We aim to investigate the prognostic value and optimum cut-off point of time-to-positivity (TTP) of blood culture in children with P. aeruginosa bacteremia.

METHODS

From August 2014 to November 2018, we enrolled the inpatients with P. aeruginosa bacteremia in a 1500-bed tertiary teaching hospital in Chongqing, China retrospectively. Receiver operating characteristic (ROC) analysis was used to determine the optimum cut-off point of TTP, and logistic regression were employed to explore the risk factors for in-hospital mortality and septic shock.

RESULTS

Totally, 52 children with P. aeruginosa bacteremia were enrolled. The standard cut-off point of TTP was18 h. Early TTP (≤18 h) group patients had remarkably higher in-hospital mortality (42.9% vs 9.7%, P = 0.014), higher incidence of septic shock (52.4% vs12.9%, P = 0.06), higher Pitt bacteremia scores [3.00 (1.00-5.00) vs 1.00 (1.00-4.00), P = 0.046] and more intensive care unit admission (61.9% vs 22.6%, P = 0.008) when compared with late TTP (> 18 h) groups. Multivariate analysis indicated TTP ≤18 h, Pitt bacteremia scores ≥4 were the independent risk factors for in-hospital mortality (OR 5.88, 95%CI 1.21-21.96, P = 0.035; OR 4.95, 95%CI 1.26-27.50, P = 0.024; respectively). The independent risk factors for septic shock were as follows: TTP ≤18 h, Pitt bacteremia scores ≥4 and hypoalbuminemia (OR 6.30, 95%CI 1.18-33.77, P = 0.032; OR 8.15, 95%CI 1.15-42.43, P = 0.014; OR 6.46, 95% CI 1.19-33.19 P = 0.031; respectively).

CONCLUSIONS

Early TTP (≤18 hours) appeared to be associated with worse outcomes for P. aeruginosa bacteremia children.

Development of the short version of the spielberger state-trait anxiety inventory

The Spielberger State-Trait Anxiety Inventory (STAI) has been widely used to measure the state and trait components of anxiety. We sought to develop a short, yet reliable and valid form of these scales for use in circumstances where the full-form is not feasible. We abbreviated the scales using item response theory analyses to retain the items that could discriminate the best among participants. One sample (N = 922) completed the state scale, a second sample (N = 2227) completed the trait scale, while a third sample (N = 250) completed the short forms. Our participants completed the Hungarian version of STAI alongside other measures to observe external validity. We calculated cut-off scores for the state (>9.5,) and trait (>13.5) scales. A total of 19.5% and 20.1% of the respondents reached the cut-off scores. The five-item short forms of STAI had sound psychometric properties that are comparable to those obtained on the full-form. The external validity of the scales is also demonstrated. We report detailed descriptive statistics that could be used in further studies as standards. The short scales are reliable measures that could be used in clinical screening and behavioural research; especially where practical considerations preclude the use of a longer questionnaire.

Estimating the area under a receiver operating characteristic curve using partially ordered sets

Ranked set sampling (RSS), known as a cost-effective sampling technique, requires that the ranker gives a complete ranking of the units in each set. Frey (2012) proposed a modification of RSS based on partially ordered sets, referred to as RSS-t in this paper, to allow the ranker to declare ties as much as he/she wishes. We consider the problem of estimating the area under a receiver operating characteristics (ROC) curve using RSS-t samples. The area under the ROC curve (AUC) is commonly used as a measure for the effectiveness of diagnostic markers. We develop six nonparametric estimators of the AUC with/without utilizing tie information based on different approaches. We then compare the estimators using a Monte Carlo simulation and an empirical study with real data from the National Health and Nutrition Examination Survey. The results show that utilizing tie information increases the efficiency of estimating the AUC. Suggestions about when to choose which estimator are also made available to practitioners.

Accurate likelihood inference for the volume under the ROC surface

BACKGROUND

With three ordered diagnostic categories, the volume under the receiver operating characteristic (ROC) surface, which is the extension of the area under the ROC curve for binary diagnostic outcomes, is the most commonly used measure for the overall diagnostic accuracy. For a continuous-scale diagnostic test, classical likelihood-based inference about the area under the ROC curve can be inaccurate, in particular when the sample size is small, and higher order inferential procedures have been proposed.

AIM

The goal of this paper is to illustrate higher order likelihood procedures for parametric inference in small samples, which provide accurate point estimates and confidence intervals for the volume under the ROC surface.

METHODS

Simulation studies are performed in order to illustrate the accuracy of the proposed methodology, and two applications to real data are discussed.

RESULTS

We show that likelihood modern inference provide refinements to classical inferential results. Furthermore, the freely available R package likelihoodAsy makes now their use almost automatic.

CONCLUSION

Modern likelihood inference based on higher-order asymptotic methods for the area under the ROC surface provide refinements to classical inferential results. A possible limitation of higher-order asymptotic methods for practical use is that their software implementation can be awkward. Nevertheless, use of the freely available R package likelihoodAsy makes such implementation straightforward.

Using Copula Functions to Estimate The AUC for Two Dependent Diagnostic Tests

When performing validation studies on diagnostic classification procedures, one or more biomarkers are typically measured in individuals. Some of these biomarkers may provide better information; moreover, more than one biomarker may be significant and may exhibit dependence between them. This proposal intends to estimate the Area Under the Receiver Operating  Characteristic Curve  (AUC)  for classifying individuals in a screening study. We analyze the dependence between the results of the tests by means of copula-type dependence (using FGM and Gumbel-Barnett copula functions), and studying the respective AUC under this type of dependence. Three different dependence-level values were evaluated for each copula function considered. In most of the reviewed literature, the authors assume a normal model to represent the performance of the biomarkers used for clinical diagnosis. There are situations in which assuming normality is not possible because that model is not suitable for one or both biomarkers. The proposed statistical model does not depend on some distributional assumption for the biomarkers used for diagnosis procedure, and additionally, it is not necessary to observe a strong or moderate linear dependence between them.

TAG-glucose (TyG) index in childhood: an estimate of cut-off points and the relation to cardiometabolic risk in 4- to 9-year-old children

OBJECTIVE

To propose cut-off points for the TAG-glucose (TyG) index in Brazilian children and evaluate the link to cardiometabolic risk.

DESIGN

A cross-sectional study with children from a municipality in Minas Gerais, Brazil. Anthropometric (weight, height, waist circumference and waist:height ratio), biochemical (lipid and glucose profile) and blood pressure (BP) tests were performed. Using the receiver operating characteristic curve, cut-off points for the TyG index were proposed according to sex using homoeostasis model of assessment - insulin resistance (IR) as the reference method.

SETTING

Viçosa, MG, Brazil.

PARTICIPANTS

Children aged 4-9 years (n 515).

RESULTS

The TyG index cut-off points to identify the risk of IR were 7·9 and 8·1 for boys and girls, respectively. We observed that 48·7 % of the children had an increased TyG index. The increased TyG index was associated with overweight, total body and central fat, increased BP and altered lipid profile. Children with an increased TyG index had a higher accumulation of cardiometabolic risk factors.

CONCLUSIONS

According to the cut-off points proposed by the current study, children at risk of IR estimated by the TyG index presented a higher cardiometabolic risk, including isolated risk factors, as to the higher accumulation of these.

Use of Typhoid Vi-Polysaccharide Vaccine as a Vaccine Probe to Delineate Clinical Criteria for Typhoid Fever

Blood cultures (BCs) detect an estimated 50% of typhoid fever cases. There is need for validated clinical criteria to define cases that are BC negative, both to help direct empiric antibiotic treatment and to better evaluate the magnitude of protection conferred by typhoid vaccines. To derive and validate a clinical rule for defining BC-negative typhoid fever, we assessed, in a cluster-randomized effectiveness trial of Vi-polysaccharide (ViPS) typhoid vaccine in Kolkata, India, 14,797 episodes of fever lasting at least 3 days during 4 years of comprehensive, BC-based surveillance of 70,865 persons. A recursive partitioning algorithm was used to develop a decision rule to predict BC-proven typhoid cases with a diagnostic specificity of 97–98%. To validate this rule as a definition for BC-negative typhoid fever, we assessed whether the rule defined culture-negative syndromes prevented by ViPS vaccine. In a training subset of individuals, we identified the following two rules: rule 1: patients aged < 15 years with prolonged fever accompanied by a measured body temperature ≥ 100°F, headache, and nausea; rule 2: patients aged ≥ 15 years with prolonged fever accompanied by nausea and palpable liver but without constipation. The adjusted protective efficacy of ViPS against clinical typhoid defined by these rules in persons aged ≥ 2 years in a separate validation subset was 33% (95% CI: 4–53%). We have defined and validated a clinical rule for predicting BC-negative typhoid fever using a novel vaccine probe approach. If validated in other settings, this rule may be useful to guide clinical care and to enhance typhoid vaccine evaluations.

Precision Diagnostics by Affinity-Mass Spectrometry: A Novel Approach for Fetal Growth Restriction Screening During Pregnancy

Fetal growth restriction (FGR) affects about 3% to 8% of pregnancies, leading to higher perinatal mortality and morbidity. Current strategies for detecting fetal growth impairment are based on ultrasound inspections. However, antenatal detection rates are insufficient and critical in countries with substandard care. To overcome difficulties with detection and to better discriminate between high risk FGR and low risk small for gestational age (SGA) fetuses, we investigated the suitability of risk assessment based on the analysis of a recently developed proteome profile derived from maternal serum in different study groups. Maternal serum, collected at around 31 weeks of gestation, was analyzed in 30 FGR, 15 SGA, and 30 control (CTRL) pregnant women who delivered between 31 and 40 weeks of gestation. From the 75 pregnant women of this study, 2 were excluded because of deficient raw data and 2 patients could not be grouped due to indeterminate results. Consistency between proteome profile and sonography results was obtained for 59 patients (26 true positive and 33 true negative). Of the proteome profiling 12 contrarious grouped individuals, 3 were false negative and 9 were false positive cases with respect to ultrasound data. Both true positive and false positive grouping transfer the respective patients to closer surveillance and thorough pregnancy management. Accuracy of the test is considered high with an area-under-curve value of 0.88 in receiver-operator-characteristics analysis. Proteome profiling by affinity-mass spectrometry during pregnancy provides a reliable method for risk assessment of impaired development in fetuses and consumes just minute volumes of maternal peripheral blood. In addition to clinical testing proteome profiling by affinity-mass spectrometry may improve risk assessment, referring pregnant women to specialists early, thereby improving perinatal outcomes.