Bayes or not Bayes, is this the question?

Midbrain and pallidal iron changes identify patients with REM sleep behaviour disorder and Parkinson's disease

Idiopathic REM sleep behaviour disorder (iRBD) is considered a prodromal form of Parkinson's Disease (PD), potentially exhibiting similar patterns of neurodegeneration, such as brain iron changes. We investigated midbrain and pallidal iron using quantitative susceptibility mapping (QSM) in 16 iRBD patients, 30 PD patients, and 38 age-matched healthy controls (HCs) with 3T MRI. QSM revealed elevated substantia nigra pars compacta (SNc) mean susceptibility in both iRBD and PD patient groups compared to HCs, though iRBD and PD QSM measures did not differ. There were no SN pars reticulata group differences. Mean susceptibility was reduced for PD relative to iRBD and HCs in the globus pallidus externa (GPe). Furthermore, mean susceptibility was reduced for PD relative to iRBD in the GP interna (GPi). GPe/GPi mean susceptibility decreased with PD subgroup motor severity. Consistent with this, QSM in left GPi and MDS-UPDRS-III scores correlated negatively in PD patients, as well as in iRBD and PD patients combined. PD patients also evidenced higher mean susceptibility in the right ventral tegmental area (VTA) compared to iRBD and HCs, consistent with later VTA degeneration. RBD symptomatology did not correlate with QSM values. Combining SNc, GPe, GPi, and VTA QSM values, we distinguished iRBD-HCs, PD-HCs, and iRBD-PD patients at single-subject levels (0.84, 0.86, and 0.81 accuracies), using ROC curve analyses with repeated k-folds cross-validation. Using 3T MRI, QSM values in SNc, GPe, GPi, and VTA demonstrate promise as investigational measures and diagnostic/progression biomarkers of prodromal and early PD.

Comparison of Deep Learning and Traditional Machine Learning Models for Predicting Mild Cognitive Impairment Using Plasma Proteomic Biomarkers

Mild cognitive impairment (MCI) is a clinical condition characterized by a decline in cognitive ability and progression of cognitive impairment. It is often considered a transitional stage between normal aging and Alzheimer's disease (AD). This study aimed to compare deep learning (DL) and traditional machine learning (ML) methods in predicting MCI using plasma proteomic biomarkers. A total of 239 adults were selected from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort along with a pool of 146 plasma proteomic biomarkers. We evaluated seven traditional ML models (support vector machines (SVMs), logistic regression (LR), naïve Bayes (NB), random forest (RF), k-nearest neighbor (KNN), gradient boosting machine (GBM), and extreme gradient boosting (XGBoost)) and six variations of a deep neural network (DNN) model-the DL model in the H2O package. Least Absolute Shrinkage and Selection Operator (LASSO) selected 35 proteomic biomarkers from the pool. Based on grid search, the DNN model with an activation function of "Rectifier With Dropout" with 2 layers and 32 of 35 selected proteomic biomarkers revealed the best model with the highest accuracy of 0.995 and an F1 Score of 0.996, while among seven traditional ML methods, XGBoost was the best with an accuracy of 0.986 and an F1 Score of 0.985. Several biomarkers were correlated with the APOE-ε4 genotype, polygenic hazard score (PHS), and three clinical cerebrospinal fluid biomarkers (Aβ42, tTau, and pTau). Bioinformatics analysis using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) revealed several molecular functions and pathways associated with the selected biomarkers, including cytokine-cytokine receptor interaction, cholesterol metabolism, and regulation of lipid localization. The results showed that the DL model may represent a promising tool in the prediction of MCI. These plasma proteomic biomarkers may help with early diagnosis, prognostic risk stratification, and early treatment interventions for individuals at risk for MCI.

Enhancing therapeutic reasoning: key insights and recommendations for education in prescribing

BACKGROUND

Despite efforts to improve undergraduate clinical pharmacology & therapeutics (CPT) education, prescribing errors are still made regularly. To improve CPT education and daily prescribing, it is crucial to understand how therapeutic reasoning works. Therefore, the aim of this study was to gain insight into the therapeutic reasoning process.

METHODS

A narrative literature review has been performed for literature on cognitive psychology and diagnostic and therapeutic reasoning.

RESULTS

Based on these insights, The European Model of Therapeutic Reasoning has been developed, building upon earlier models and insights from cognitive psychology. In this model, it can be assumed that when a diagnosis is made, a primary, automatic response as to what to prescribe arises based on pattern recognition via therapy scripts (type 1 thinking). At some point, this response may be evaluated by the reflective mind (using metacognition). If it is found to be incorrect or incomplete, an alternative response must be formulated through a slower, more analytical and deliberative process, known as type 2 thinking. Metacognition monitors the reasoning process and helps a person to form new therapy scripts after they have chosen an effective therapy. Experienced physicians have more and richer therapy scripts, mostly based on experience and enabling conditions, instead of textbook knowledge, and therefore their type 1 response is more often correct.

CONCLUSION

Because of the important role of metacognition in therapeutic reasoning, more attention should be paid to metacognition in CPT education. Both trainees and teachers should be aware of the possibility to monitor and influence these cognitive processes. Further research is required to investigate the applicability of these insights and the adaptability of educational approaches to therapeutic reasoning.

Corticosteroids in critically ill patients with community-acquired pneumonia: A systematic review and Bayesian meta-analysis

Introduction

This systematic review and meta-analysis aimed to evaluate the effectiveness and safety of adjunct systemic corticosteroid therapy in patients admitted to the intensive care unit (ICU) with bacterial community-acquired pneumonia (CAP).

Method

We searched MEDLINE, Embase and the Cochrane Library to identify randomised controlled trials (RCTs) published from the databases' inception to February 2024. All RCTs evaluating the effect of systemic corticosteroids on mortality, compared to standard of care among adult bacterial CAP patients admitted to ICU were included. Bayesian meta-analysis was conducted in line with the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines. Independent authors reviewed each study for eligibility, extracted data and assessed risk of bias in duplicate, with discrepancies referred to senior reviewers.

Results

A total of 6 RCTs comprising 1585 patients were included for analysis. In ICU patients with severe CAP who were treated with corticosteroids, there was no significant reduction in hospital mortality (risk ratio [RR] 0.70, 95% confidence interval [CI] 0.39-1.14, certainty of evidence: ⊕⊕⊝⊝ low) or all-cause mortality (RR 0.68, 95% CI 0.34-1.22, ⊕⊕⊝⊝ low) compared with placebo. The use of corticosteroids showed a significant reduction in mechanical ventilation post-intervention (RR 0.58, 95% CI 0.37-0.86, ⊕⊕⊕⊕ high) compared with placebo. In a subgroup analysis of patients treated with hydrocortisone, hospital mortality was significantly reduced (RR 0.45, 95% CI 0.20-0.88, ⊕⊕⊝⊝ low) compared with placebo. There was no significant increase in gastrointestinal bleeding, secondary infections or hyperglycaemia in patients treated with corticosteroids.

Conclusion

Corticosteroids significantly reduced mechanical ventilation requirements, and hydrocor-tisone significantly reduced hospital mortality. Further work is required to determine whether other corticosteroids reduce mortality among ICU patients with CAP.

Prenatal air pollution and children's autism traits score: Examination of joint associations with maternal intake of vitamin D, methyl donors, and polyunsaturated fatty acids using mixture methods

Background

Maternal nutrient intake may moderate associations between environmental exposures and children's neurodevelopmental outcomes, but few studies have assessed joint effects. We aimed to evaluate whether prenatal nutrient intake influences the association between air pollutants and autism-related trait scores.

Methods

We included 126 participants from the EARLI (Early Autism Risk Longitudinal Investigation, 2009-2012) cohort, which followed US pregnant mothers who previously had a child with autism. Bayesian kernel machine regression and traditional regression models were used to examine joint associations of prenatal nutrient intake (vitamins D, B12, and B6; folate, choline, and betaine; and total omega 3 and 6 polyunsaturated fatty acids, reported via food frequency questionnaire), air pollutant exposure (particulate matter <2.5 μm [PM2.5], nitrogen dioxide [NO2], and ozone [O3], estimated at the address level), and children's autism-related traits (measured by the Social Responsiveness Scale [SRS] at 36 months).

Results

Most participants had nutrient intakes and air pollutant exposures that met US standards. Bayesian kernel machine regression mixture models and traditional regression models provided little evidence of individual or joint associations of nutrients and air pollutants with SRS scores or of an association between the overall mixture and SRS scores.

Conclusion

In this cohort with a high familial likelihood of autism, we did not observe evidence of joint associations between air pollution exposures and nutrient intake with autism-related traits. Future work should examine the use of these methods in larger, more diverse samples, as our results may have been influenced by familial liability and/or relatively high nutrient intakes and low air pollutant exposures.

Machine Learning-Based Prediction of Binge Drinking among Adults in the United State: Analysis of the 2022 Health Information National Trends Survey

Little is known about the association of social media and belief in alcohol and cancer with binge drinking. This study aimed to perform feature selection and develop machine learning (ML) tools to predict occurrence of binge drinking among adults in the United State. A total of 5,886 adults including 1,252 who ever experienced with binge drinking were selected from the 2022 Health Information National Trends Survey (HINTS 6). Feature selection of 69 variables was conducted using Boruta and the Least Absolute Shrinkage and Selection Operator (LASSO). The Random Over Sampling Example (ROSE) method was utilized to deal with the imbalance data. Seven machine learning (ML) tools including the Support Vector Machines (SVMs) algorithms, Logistic Regression, Naïve Bayes, Random Forest, K-Nearest Neighbor, Gradient Boosting Machine, and XGBoost were applied to develop ML models to predict binge drinking. The overall prevalence of binge drinking among U.S. adults is 21.3%. Both Boruta and LASSO selected 28 identical variables. SVM with Radial Basis Function revealed the best model with the highest accuracy of 0.949 and sensitivity of 0.958. The top risk factors of binge drinking were tobacco use (e-cigarette use and smoking status), belief in alcohol (alcohol decreases the risk of future health), belief in cancer (prevention is not possible, worry about getting cancer), and social media (social media visits and sharing health information). These findings underscore the need for multiple health behavior interventions to enhance education related to alcohol use and cancer and how to effectively employ social media to improve health outcomes.

Rush to Charge, Dead to Drive: Application of Deadline Rush Model to Electric Vehicle User's Charging Behavior

OBJECTIVE

This work aims to estimate the portion of electric vehicle (EV) users who exhibit procrastination-like behavior, almost equivalent to an "empty" battery, before they decide to charge their vehicles.

BACKGROUND

There is a human tendency to procrastinate when a deadline approaches. Human behavior in the presence of deadlines has been studied in different fields to evaluate individuals' performance or organizational efficiency and effectiveness. However, this phenomenon has not been investigated among EV users.

METHOD

This study explores users' procrastination-like behavior among 69 Rhode Island public charging stations' data representing 70,611 charging events. The Deadline Rush Model is incorporated to model frequent users' charging profiles. To conduct a robust estimation, the Bayesian Mixture Model is implemented.

RESULTS

With the selection of an informative prior, the Bayesian Mixture Model estimated that almost one-third of frequent users procrastinate charging.

CONCLUSION

The majority of procrastination-like users have small battery sizes. Although procrastination-like users need to charge when they arrive at a location, that might not necessarily be true for a plug-in hybrid; thus, systematically, they can clog the system for other users whose needs are more pressing. Understanding unique and unexplored charging behaviors among EV users is beneficial to EV infrastructure stakeholders in reducing the adoption threshold by providing a reliable and ubiquitous charging network.

APPLICATION

The findings identify a different kind of demand on the EV infrastructure than previously modeled and can directly influence future decision-making criteria in terms of planning to optimize to accommodate EV drivers with different charging behaviors.

A Bayesian Network Meta-Analysis of the Effect of Targeted Therapies on the Total Length of Hospital Stay in Children with Drug-Induced Stevens-Johnson Syndrome/Toxic Epidermal Necrolysis Syndrome

Background: Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are rare potentially life-threatening hypersensitivity disorders characterized by widespread skin and mucosal involvement. However, there is no standardized evidence-based treatment to reduce the complications of SJS/TEN. This article aims to compare the efficacy of different treatments for pediatric SJS/TEN in terms of length of hospital stay (LOS) using a Bayesian network meta-analysis (NMA). A Bayesian NMA is used to compare and combine evidence from multiple studies and allows clinicians to estimate the relative effectiveness of different treatments/interventions while accounting for heterogeneity in the available evidence. Methods: We conducted a comprehensive electronic database search for studies compatible with our inclusion criteria. Six studies with 103 patients were included in the NMA; of them, 37 patients were treated with intravenous immunoglobulin (IVIG), 37 with systemic corticosteroids (CS), 23 with IVIG + CS, and 3 with Etanercept (ET) + CS. Patients with a median age of 10 years were included in the study. Results: CS had the highest probability of being the most optimal treatment for SJS/TEN in terms of shorter LOS based on the Surface Under the Cumulative Ranking curve levels, and CS + IVIG was associated with a statistically nonsignificant trend toward shorter LOS than IVIG alone. Remarkably, none of the treatments showed a significant benefit over the other interventions in terms of LOS. Conclusion: Current evidence suggests that coadministration of CS and IVIG may be associated with a shorter LOS than IVIG alone. Further research with larger randomized controlled trials is needed to reach a definitive conclusion about the efficacy of specific therapy on LOS in pediatric SJS/TEN and to establish more definitive treatment guidelines.

Brain dopamine receptor system is not altered in obesity: Bayesian and frequentist meta-analyses

Feeding induces dopamine release in the striatum, and a dysfunction of the dopaminergic reward system can lead to overeating, and obesity. Studies have reported inconsistent findings of dopamine receptor (DR) positron emission tomography scans in obesity. Here we investigated the association between DR availability and overweight/obesity using Bayesian and frequentist meta-analysis. We performed a systematic search of Embase, Medline, Scopus and Web of Science for studies that compared striatal DR availability between lean subjects and overweight/obese subjects. The standardized mean difference (Hedge's g) of DR availability was calculated after extraction of data from each study. Studies were divided into two groups according to the definition of overweight/obese subjects (body mass index [BMI] cutoff of 25 and 30 kg/m2 ). Both Bayesian and frequentist meta-analysis was done in R Statistical Software version 4.2.2 (The R Foundation for Statistical Computing). Nine studies were eligible for inclusion in this study. Three studies with C11-raclopride, one with C11-PNHO, two with F18-fallypride, one with I123-IBZM, one with C11-NMB and one with both C11-raclopride and C11-PNHO were included. In Bayesian meta-analysis, the standardized mean difference of DR availability between lean and overweight/obese subjects markedly overlapped with zero regardless of BMI cutoff for obesity. In frequentist meta-analysis, the pooled standardized mean difference of DR availability did not show the significant difference between lean and overweight/obese subjects. There was an effect of the radiopharmaceutical on the standardized mean difference of DR availability in meta-analysis of BMI cutoff of 25 kg/m2 . In conclusion, brain DR availability is not different between lean and overweight/obese subjects. However, the effect is dependent on the radiopharmaceutical and the degree of obesity. Further studies with multi-radiopharmaceutical in the same individuals are needed to understand the association between DR and obesity.

Interpretable prediction of 3-year all-cause mortality in patients with chronic heart failure based on machine learning

BACKGROUND

The goal of this study was to assess the effectiveness of machine learning models and create an interpretable machine learning model that adequately explained 3-year all-cause mortality in patients with chronic heart failure.

METHODS

The data in this paper were selected from patients with chronic heart failure who were hospitalized at the First Affiliated Hospital of Kunming Medical University, from 2017 to 2019 with cardiac function class III-IV. The dataset was explored using six different machine learning models, including logistic regression, naive Bayes, random forest classifier, extreme gradient boost, K-nearest neighbor, and decision tree. Finally, interpretable methods based on machine learning, such as SHAP value, permutation importance, and partial dependence plots, were used to estimate the 3-year all-cause mortality risk and produce individual interpretations of the model's conclusions.

RESULT

In this paper, random forest was identified as the optimal aools lgorithm for this dataset. We also incorporated relevant machine learning interpretable tand techniques to improve disease prognosis, including permutation importance, PDP plots and SHAP values for analysis. From this study, we can see that the number of hospitalizations, age, glomerular filtration rate, BNP, NYHA cardiac function classification, lymphocyte absolute value, serum albumin, hemoglobin, total cholesterol, pulmonary artery systolic pressure and so on were important for providing an optimal risk assessment and were important predictive factors of chronic heart failure.

CONCLUSION

The machine learning-based cardiovascular risk models could be used to accurately assess and stratify the 3-year risk of all-cause mortality among CHF patients. Machine learning in combination with permutation importance, PDP plots, and the SHAP value could offer a clear explanation of individual risk prediction and give doctors an intuitive knowledge of the functions of important model components.

Detailed statistical analysis plan for a secondary Bayesian analysis of the SafeBoosC-III trial: a multinational, randomised clinical trial assessing treatment guided by cerebral oximetry monitoring versus usual care in extremely preterm infants

BACKGROUND

Extremely preterm infants have a high mortality and morbidity. Here, we present a statistical analysis plan for secondary Bayesian analyses of the pragmatic, sufficiently powered multinational, trial-SafeBoosC III-evaluating the benefits and harms of cerebral oximetry monitoring plus a treatment guideline versus usual care for such infants.

METHODS

The SafeBoosC-III trial is an investigator-initiated, open-label, randomised, multinational, pragmatic, phase III clinical trial with a parallel-group design. The trial randomised 1601 infants, and the frequentist analyses were published in April 2023. The primary outcome is a dichotomous composite outcome of death or severe brain injury. The exploratory outcomes are major neonatal morbidities associated with neurodevelopmental impairment later in life: (1) bronchopulmonary dysplasia; (2) retinopathy of prematurity; (3) late-onset sepsis; (4) necrotising enterocolitis; and (5) number of major neonatal morbidities (count of bronchopulmonary dysplasia, retinopathy of prematurity, and severe brain injury). The primary Bayesian analyses will use non-informed priors including all plausible effects. The models will use a Hamiltonian Monte Carlo sampler with 1 chain, a sampling of 10,000, and at least 25,000 iterations for the burn-in period. In Bayesian statistics, such analyses are referred to as 'posteriors' and will be presented as point estimates with 95% credibility intervals (CrIs), encompassing the most probable results based on the data, model, and priors selected. The results will be presented as probability of any benefit or any harm, Bayes factor, and the probability of clinical important benefit or harm. Two statisticians will analyse the blinded data independently following this protocol.

DISCUSSION

This statistical analysis plan presents a secondary Bayesian analysis of the SafeBoosC-III trial. The analysis and the final manuscript will be carried out and written after we publicise the primary frequentist trial report. Thus, we can interpret the findings from both the frequentists and Bayesian perspective. This approach should provide a better foundation for interpreting of our findings.

TRIAL REGISTRATION

ClinicalTrials.org, NCT03770741. Registered on 10 December 2018.

Longitudinal exploration of biopsychosocial profiles in individuals with anorexia nervosa

BACKGROUND

Previous work in individuals with Anorexia Nervosa (AN) has demonstrated a range of psycho-social difficulties such as increased anxiety, depression, obsessive-compulsive symptoms, as well as difficulties in work and with interpersonal interactions. However, making inferences regarding the stability of these psycho-social difficulties from previous studies is challenging, due to lack of a control group and known frequentist statistical issues.

METHODS

134 participants, 40 healthy controls (HC) and 94 participants with AN, completed self-reported measures designed to explore eating disorder concerns, body mass index, mood symptoms, work and social functioning as well as traits associated with autism at two time points, two years apart. A principal component analysis and Bayesian mixed effects models were used to build and explore group differences in bio-psychosocial profiles at time points.

RESULTS

The Bayesian models demonstrated evidence for individuals with AN having higher scores for a component representing psycho-social difficulties and lower scores for a component representing biological difficulties compared to HC, at both time points. There was no evidence of a group difference for a component representing autism.

CONCLUSIONS

Our results demonstrate that persistent psycho-social difficulties are a feature in individuals with AN.

Designing and Implementing "Living and Breathing" Clinical Trials: An Overview and Lessons Learned from the COVID-19 Pandemic

The practice of medicine is characterized by uncertainty, and the findings of randomized clinical trials (RCTs) are meant to help curb that uncertainty. Traditional RCTs, however, have many limitations. To overcome some of these limitations, new trial paradigms rooted in the origins of evidence-based medicine are beginning to disrupt the traditional mold. These new designs recognize uncertainty permeates medical decision making and aim to capitalize on modern health system infrastructure to integrate investigation as a component of care delivery. This article provides an overview of "living, breathing" trials, including current state, anticipated developments, and areas of controversy.

Application of deep learning models for detection of subdural hematoma: a systematic review and meta-analysis

BACKGROUND

This study aimed to investigate the application of deep learning (DL) models for the detection of subdural hematoma (SDH).

METHODS

We conducted a comprehensive search using relevant keywords. Articles extracted were original studies in which sensitivity and/or specificity were reported. Two different approaches of frequentist and Bayesian inference were applied. For quality and risk of bias assessment we used Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2).

RESULTS

We analyzed 22 articles that included 1,997,749 patients. In the first step, the frequentist method showed a pooled sensitivity of 88.8% (95% confidence interval (CI): 83.9% to 92.4%) and a specificity of 97.2% (95% CI 94.6% to 98.6%). In the second step, using Bayesian methods including 11 studies that reported sensitivity and specificity, a sensitivity rate of 86.8% (95% CI: 77.6% to 92.9%) at a specificity level of 86.9% (95% CI: 60.9% to 97.2%) was achieved. The risk of bias assessment was not remarkable using QUADAS-2.

CONCLUSION

DL models might be an appropriate tool for detecting SDHs with a reasonably high sensitivity and specificity.

Add-on therapy for pulmonary fibrosis, a forthcoming era with implications for practice: the BI 101550 and RELIEF trials

The therapeutic landscape for idiopathic pulmonary fibrosis (IPF) and progressive fibrosing interstitial lung disease (PFILD) is increasingly complex, with add-on antifibrotic options now in clinical trials, or available for patients progressing on first-line therapy in both conditions. Here, we review two recent trials of potential add-on therapeutic options, the BI 101550 and RELIEF trials. BI 101550 was a phase 2 randomised control trial (RCT) of a novel phosphodiesterase-4 inhibitor in patients with IPF, with a primary end-point of change in forced vital capacity (ΔFVC) (in mL) at 12 weeks. The RELIEF trial was a phase 2 RCT in patients with PFILD, with a primary end-point of ΔFVC (absolute % predicted) over 48 weeks. Whilst the BI 101550 and RELIEF trials showed positive results in their primary end-points, the strengths and weaknesses of both trials are discussed with importance for their interpretation and clinical impact. We review current clinical practice in IPF and PFILD and place the BI101550 and RELIEF trial results in context, highlighting advances and problems with antifibrotic therapies.

Commentary on

Richeldi L, et al. Trial of a preferential phosphodiesterase 4B inhibitor for idiopathic pulmonary fibrosis. N Engl J Med 2022; 386: 2178-2187.Behr J, et al. Pirfenidone in patients with progressive fibrotic interstitial lung diseases other than idiopathic pulmonary fibrosis (RELIEF): a double-blind, randomised, placebo-controlled, phase 2b trial. Lancet Respir Med 2021; 9: 476-486.

Examining Prenatal Dietary Factors in Association with Child Autism-Related Traits Using a Bayesian Mixture Approach: Results from 2 United States Cohorts

Background

Prior work has suggested relationships between prenatal intake of certain nutrients and autism.

Objectives

We examined a broad set of prenatal nutrients and foods using a Bayesian modeling approach.

Methods

Participants were drawn from the Early Autism Risks Longitudinal Investigation (n = 127), a cohort following women with a child with autism through a subsequent pregnancy. Participants were also drawn from the Nurses' Health Study II (NHSII, n = 713), a cohort of United States female nurses, for comparison analyses. In both studies, information on prospectively reported prenatal diet was drawn from food frequency questionnaires, and child autism-related traits were measured by the Social Responsiveness Scale (SRS). Bayesian kernel machine regression was used to examine the combined effects of several nutrients with neurodevelopmental relevance, including polyunsaturated fatty acids (PUFAs), iron, zinc, vitamin D, folate, and other methyl donors, and separately, key food sources of these, in association with child SRS scores in crude and adjusted models.

Results

In adjusted analyses, the overall mixture effects of nutrients in Early Autism Risks Longitudinal Investigation and foods in both cohorts on SRS scores were not observed, though there was some suggestion of decreasing SRS scores with increasing overall nutrient mixture in NHSII. No associations were observed with folate within the context of this mixture, but holding other nutrients fixed, n-6 PUFAs were associated with lower SRS scores in NHSII. In both cohorts, lower SRS scores were observed with higher intake of some groupings of vegetables, though for differing types of vegetables across cohorts, and some vegetable groups were associated with higher SRS scores in NHSII.

Conclusions

Our work extends prior research and suggests the need to further consider prenatal dietary factors from a combined effects perspective. In addition, findings here point to potential differences in nutrient associations based on a family history of autism, which suggests the need to consider gene interactions in future work.

Evaluations of statistical methods for outlier detection when benchmarking in clinical registries: a systematic review

OBJECTIVES

Benchmarking is common in clinical registries to support the improvement of health outcomes by identifying underperforming clinician or health service providers. Despite the rise in clinical registries and interest in publicly reporting benchmarking results, appropriate methods for benchmarking and outlier detection within clinical registries are not well established, and the current application of methods is inconsistent. The aim of this review was to determine the current statistical methods of outlier detection that have been evaluated in the context of clinical registry benchmarking.

DESIGN

A systematic search for studies evaluating the performance of methods to detect outliers when benchmarking in clinical registries was conducted in five databases: EMBASE, ProQuest, Scopus, Web of Science and Google Scholar. A modified healthcare modelling evaluation tool was used to assess quality; data extracted from each study were summarised and presented in a narrative synthesis.

RESULTS

Nineteen studies evaluating a variety of statistical methods in 20 clinical registries were included. The majority of studies conducted application studies comparing outliers without statistical performance assessment (79%), while only few studies used simulations to conduct more rigorous evaluations (21%). A common comparison was between random effects and fixed effects regression, which provided mixed results. Registry population coverage, provider case volume minimum and missing data handling were all poorly reported.

CONCLUSIONS

The optimal methods for detecting outliers when benchmarking clinical registry data remains unclear, and the use of different models may provide vastly different results. Further research is needed to address the unresolved methodological considerations and evaluate methods across a range of registry conditions.

PROSPERO REGISTRATION NUMBER

CRD42022296520.

Characterizing the dimensional structure of early-life adversity in the Adolescent Brain Cognitive Development (ABCD) Study

Early-life adversity has profound consequences for youth neurodevelopment and adjustment; however, experiences of adversity are heterogeneous and interrelated in complex ways that can be difficult to operationalize and organize in developmental research. We sought to characterize the underlying dimensional structure of co-occurring adverse experiences among a subset of youth (ages 9-10) from the Adolescent Brain Cognitive Development (ABCD) Study (N = 7115), a community sample of youth in the United States. We identified 60 environmental and experiential variables that reflect adverse experiences. Exploratory factor analysis identified 10 robust dimensions of early-life adversity co-occurrence, corresponding to conceptual domains such as caregiver substance use and biological caregiver separation, caregiver psychopathology, caregiver lack of support, and socioeconomic disadvantage / neighborhood lack of safety. These dimensions demonstrated distinct associations with internalizing problems, externalizing problems, cognitive flexibility, and inhibitory control. Non-metric multidimensional scaling characterized qualitative similarity among the 10 identified dimensions. Results supported a nonlinear three-dimensional structure representing early-life adversity, including continuous gradients of "perspective", "environmental uncertainty", and "acts of omission/commission". Our findings suggest that there are distinct dimensions of early-life adversity co-occurrence in the ABCD sample at baseline, and the resulting dimensions may have unique implications for neurodevelopment and youth behavior.

Validity and Reliability of Short-Term Heart Rate Variability Parameters in Older People in Response to Physical Exercise

BACKGROUND

Currently, and to the best of our knowledge, there is no standardized protocol to measure the effect of low- to moderate-intensity physical exercise on autonomic modulation focused in older people.

AIM

Validate a test-retest short-term exercise protocol for measuring the autonomic response through HRV in older people.

METHODS

A test-retest study design was used. The participants were selected through intentional non-probabilistic sampling. A total of 105 older people (male: 21.9%; female: 78.1%) were recruited from a local community. The assessment protocol evaluated HRV before and immediately after the 2-min step test. It was performed twice on the same day, considering a time of three chronological hours between the two measurements.

RESULTS

The posterior distribution of estimated responses in the Bayesian framework suggests moderate to strong evidence favoring a null effect between measurements. In addition, there was moderate to robust agreement between heart rate variability (HRV) indices and assessments, except for low frequency and very low frequency, which showed weak agreement.

CONCLUSIONS

Our results provide moderate to strong evidence for using HRV to measure cardiac autonomic response to moderate exercise, suggesting that it is sufficiently reliable to show similar results to those shown in this test-retest protocol.

Mycobacterium tuberculosis complex detection in rural goat herds in South Africa using Bayesian latent class analysis

Animal tuberculosis affects a wide range of domestic and wild animal species, including goats (Capra hircus). In South Africa, Mycobacterium tuberculosis complex (MTBC) testing and surveillance in domestic goats is not widely applied, potentially leading to under recognition of goats as a potential source of M. bovis spread to cattle as well as humans and wildlife. The aim of this study was to estimate diagnostic test performance for four assays and determine whether M. bovis infection was present in goats sharing communal pastures with M. bovis positive cattle in the Umkhanyakude district of Northern Zululand, KwaZulu Natal. In 2019, 137 M. bovis-exposed goats were screened for MTBC infection with four diagnostic tests: the in vivo single intradermal comparative cervical tuberculin test (SICCT), in vitro QuantiFERON®-TB Gold (QFT) bovine interferon-gamma release assay (IGRA), QFT bovine interferon gamma induced protein 10 (IP-10) release assay (IPRA), and nasal swabs tested with the Cepheid GeneXpert® MTB/RIF Ultra (GXU) assay for detection of MTBC DNA. A Bayesian latent class analysis was used to estimate MTBC prevalence and diagnostic test sensitivity and specificity. Among the 137 M. bovis-exposed goats, positive test results were identified in 15/136 (11.0%) goats by the SICCT; 4/128 (3.1%) goats by the IPRA; 2/128 (1.6%) goats by the IGRA; and 26/134 (19.4%) nasal swabs by the GXU. True prevalence was estimated by our model to be 1.1%, suggesting that goats in these communal herds are infected with MTBC at a low level. Estimated posterior means across the four evaluated assays ranged from 62.7% to 80.9% for diagnostic sensitivity and from 82.9% to 97.9% for diagnostic specificity, albeit estimates of the former (diagnostic sensitivity) were dependent on model assumptions. The application of a Bayesian latent class analysis and multiple ante-mortem test results may improve detection of MTBC, especially when prevalence is low. Our results provide a foundation for further investigation to confirm infection in communal goat herds and identify previously unrecognized sources of intra- and inter-species transmission of MTBC.

Decoding Deep Learning applications for diagnosis and treatment planning

INTRODUCTION

Artificial Intelligence (AI), Machine Learning and Deep Learning are playing an increasingly significant role in the medical field in the 21st century. These recent technologies are based on the concept of creating machines that have the potential to function as a human brain. It necessitates the gathering of large quantity of data to be processed. Once processed with AI machines, these data have the potential to streamline and improve the capabilities of the medical field in diagnosis and treatment planning, as well as in the prediction and recognition of diseases. These concepts are new to Orthodontics and are currently limited to image processing and pattern recognition.

OBJECTIVE

This article exposes and describes the different methods by which orthodontics may benefit from a more widespread adoption of these technologies.

Bayesian spatio-temporal modelling of child anemia in Ethiopia using conditional autoregressive model

Anemia is a common health problem for women and under five children in low income countries. According to the WHO, anemia is considered a serious public health problem when the prevalence is greater than 40%. The prevalence of anemia among children under five in Ethiopia changes over time, and is spatially correlated because it is influenced by environmental, socio-economic and other related factors. However, to our knowledge, there is no small area level estimates of anemia among children under five in Ethiopia. Therefore, this study aimed to assess zonal level estimates of anemia using a Bayesian spatio-temporal conditional autoregressive modeling approach. The data for the study was extracted from the Ethiopian Demographic and Health Surveys (EDHS) from 2005 to 2016. A sample of 18,939 children aged 6-59 months were considered for this study. A Bayesian spatio-temporal conditional autoregressive model was implemented to identify the risk of child anemia. Smoothed relative risks along with the 95% credible interval were reported. The queen's adjacency matrix method was used in spatial smoothing and in estimating the relative risk. The prevalence of anemia among children aged 6-59 months in Ethiopia was 54% in 2005, 44% in 2011 and 57% in 2016. This study showed that low maternal education, low socio-economic status of women, and maternal anemia at zone level were strongly associated with child anemia in Ethiopia. Therefore, enhancing education for women, improving women's socioeconomic status, and mitigating maternal anemia are crucial to reduce the prevalence of childhood anemia in Ethiopia.

A 50-step walking test for analysis of recovery after decompressive surgery for thoracolumbar disc herniation in dogs

BACKGROUND

Despite its importance, there is no agreed definition of recovery of ambulation in dogs with spinal cord injury.

OBJECTIVES

To validate a new walking test in dogs recovering from thoracolumbar spinal cord injury.

ANIMALS

Two hundred twenty-four dogs weighing <20 kg: 120 normally ambulatory dogs, plus 104 dogs undergoing decompressive surgery for acute thoracolumbar intervertebral disc herniation.

METHODS

Prospective cohort studies. The distance each freely-ambulatory dog walked during 50 step cycles was regressed on ulna length. For each postsurgical dog, we recorded when the calculated 50-step distance was completed without falling, or their inability to complete this distance by 4 months or more after surgery. Bayesian analysis compared outcomes for presurgical neurologic categories; association of recovery with several preoperative variables was explored using logistic and time-to-event regression.

RESULTS

For control dogs, 50-step distance (m) = 1.384 × ulnar length (cm) + 2.773. In postsurgical dogs, the 50-step test provided decisive evidence that deep pain-negative dogs were less likely to recover ambulation than dogs with intact pain perception (12/29 recovered vs 71/75; Bayes factor [BF] = 5.9 × 106 ) and, if they did recover, it took much longer (median 91 days vs median 14 days; BF = 1.5 × 103 ). Exploratory analysis suggested that presurgical neurologic status (subhazard ratio [SHR] = 0.022; P < .001) and duration of presurgical anesthesia (SHR = 0.740; P = .04) were associated with rapidity of recovery.

CONCLUSIONS AND CLINICAL IMPORTANCE

This straightforward 50-step walking test provides robust data on ambulatory recovery well-suited to large scale pragmatic trials on treatment of thoracolumbar spinal cord injury in dogs.

Prophylaxis in healthcare workers during a pandemic: a model for a multi-centre international randomised controlled trial using Bayesian analyses

BACKGROUND

Coronavirus disease 2019 (COVID-19) has exposed the disproportionate effects of pandemics on frontline workers and the ethical imperative to provide effective prophylaxis. We present a model for a pragmatic randomised controlled trial (RCT) that utilises Bayesian methods to rapidly determine the efficacy or futility of a prophylactic agent.

METHODS

We initially planned to undertake a multicentre, phase III, parallel-group, open-label RCT, to determine if hydroxychloroquine (HCQ) taken once a week was effective in preventing severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection in healthcare workers (HCW) aged ≥ 18 years in New Zealand (NZ) and Ireland. Participants were to be randomised 2:1 to either HCQ (800 mg stat then 400 mg weekly) or no prophylaxis. The primary endpoint was time to Nucleic Acid Amplification Test-proven SARS-CoV-2 infection. Secondary outcome variables included mortality, hospitalisation, intensive care unit admissions and length of mechanical ventilation. The trial had no fixed sample size or duration of intervention. Bayesian adaptive analyses were planned to occur fortnightly, commencing with a weakly informative prior for the no prophylaxis group hazard rate and a moderately informative prior on the intervention log hazard ratio centred on 'no effect'. Stopping for expected success would be executed if the intervention had a greater than 0.975 posterior probability of reducing the risk of SARS-CoV-2 infection by more than 10%. Final success would be declared if, after completion of 8 weeks of follow-up (reflecting the long half-life of HCQ), the prophylaxis had at least a 0.95 posterior probability of reducing the risk of SARS-CoV-2 infection by more than 10%. Futility would be declared if HCQ was shown to have less than a 0.10 posterior probability of reducing acquisition of SARS-CoV-2 infection by more than 20%.

DISCUSSION

This study did not begin recruitment due to the marked reduction in COVID-19 cases in NZ and concerns regarding the efficacy and risks of HCQ treatment in COVID-19. Nonetheless, the model presented can be easily adapted for other potential prophylactic agents and pathogens, and pre-established collaborative models like this should be shared and incorporated into future pandemic preparedness planning.

TRIAL REGISTRATION

The decision not to proceed with the study was made before trial registration occurred.

Applied Bayesian Approaches for Research in Motor Neuron Disease

Statistical evaluation of empirical data is the basis of the modern scientific method. Available tools include various hypothesis tests for specific data structures, as well as methods that are used to quantify the uncertainty of an obtained result. Statistics are pivotal, but many misconceptions arise due to their complexity and difficult-to-acquire mathematical background. Even though most studies rely on a frequentist interpretation of statistical readouts, the application of Bayesian statistics has increased due to the availability of easy-to-use software suites and an increased outreach favouring this topic in the scientific community. Bayesian statistics take our prior knowledge together with the obtained data to express a degree of belief how likely a certain event is. Bayes factor hypothesis testing (BFHT) provides a straightforward method to evaluate multiple hypotheses at the same time and provides evidence that favors the null hypothesis or alternative hypothesis. In the present perspective, we show the merits of BFHT for three different use cases, including a clinical trial, basic research as well as a single case study. Here we show that Bayesian statistics is a viable addition of a scientist's statistical toolset, which can help to interpret data.

MRI Radiomics Features From Infarction and Cerebrospinal Fluid for Prediction of Cerebral Edema After Acute Ischemic Stroke

Neuroimaging biomarkers that predict the edema after acute stroke may help clinicians provide targeted therapies and minimize the risk of secondary injury. In this study, we applied pretherapy MRI radiomics features from infarction and cerebrospinal fluid (CSF) to predict edema after acute ischemic stroke. MRI data were obtained from a prospective, endovascular thrombectomy (EVT) cohort that included 389 patients with acute stroke from two centers (dataset 1, n = 292; dataset 2, n = 97), respectively. Patients were divided into edema group (brain swelling and midline shift) and non-edema group according to CT within 36 h after therapy. We extracted the imaging features of infarct area on diffusion weighted imaging (DWI) (abbreviated as DWI), CSF on fluid-attenuated inversion recovery (FLAIR) (CSF_FLAIR) and CSF on DWI (CSF_DWI), and selected the optimum features associated with edema for developing models in two forms of feature sets (DWI + CSF_FLAIR and DWI + CSF_DWI) respectively. We developed seven ML models based on dataset 1 and identified the most stable model. External validations (dataset 2) of the developed stable model were performed. Prediction model performance was assessed using the area under the receiver operating characteristic curve (AUC). The Bayes model based on DWI + CSF_FLAIR and the RF model based on DWI + CSF_DWI had the best performances (DWI + CSF_FLAIR: AUC, 0.86; accuracy, 0.85; recall, 0.88; DWI + CSF_DWI: AUC, 0.86; accuracy, 0.84; recall, 0.84) and the most stability (RSD% in DWI + CSF_FLAIR AUC: 0.07, RSD% in DWI + CSF_DWI AUC: 0.09), respectively. External validation showed that the AUC of the Bayes model based on DWI + CSF_FLAIR was 0.84 with accuracy of 0.77 and area under precision-recall curve (auPRC) of 0.75, and the AUC of the RF model based on DWI + CSF_DWI was 0.83 with accuracy of 0.81 and the auPRC of 0.76. The MRI radiomics features from infarction and CSF may offer an effective imaging biomarker for predicting edema.

Estimation of genetic parameters in dairy production in girolando cattle

Abstract Milk production is an important economic activity in Brazil. Dairy farmers would benefit from animal breeding programs that aid in identification and selection of animals with the best cost/benefit ratio to maximize productivity, and additionally provide advice on disposal of less productive animals. This study aims to estimate the heritability and repeatability of milk production corrected for 305 days (PL305) in a herd of Girolando cattle. We analyzed 528 lactations in 251 cows. For the analysis, uniform a priori distribution was defined for systematic effects. Gaussian and inverted Wishart distributions were defined as a priori distributions for random effects. The variance components were estimated based on Bayesian inference using the MCMCglmm function available in the MCMCglmm package of the R software. Convergence was verifed with the Geweke test available in the R software. The heritability and repeatability were estimated from the variance component results. Heritability was at 0.28, suggesting that selection for the milk production trait leads to efficient genetic progress in the herd. Phenotypic variance was mainly due to environmental variance; therefore, the phenotype of individuals should not be considered as indicator for additive genetic variance. Repeatability was at 0.93, indicating that the first performance of the animals based on milk production average is a good indicator of the second, and the data could be used for disposal decisions.

Estimação de parâmetros genéticos na produção leiteira em bovinos girolando

Resumo A produção de leite é uma das atividades econômicas mais importantes da agropecuária brasileira. Produtores podem usufruir de programas de melhoramento genético que permitem a identificação dos melhores animais e sua seleção para maximizar a produtividade com a melhor relação custo/benefício, além do aconselhamento do descarte de animais menos produtivos. Objetivou-se estimar a herdabilidade e repetibilidade da produção de leite corrigida para 305 dias (PL305) de um rebanho de bovinos da raça Girolando. Foram analisadas 528 lactações de 251 vacas. Para análise foi definida a distribuição uniforme a priori para efeitos sistemáticos. As distribuições de Wishart gaussiana e invertida foram definidas como distribuições a priori para efeitos aleatórios. Os componentes de variância foram estimados utilizando inferência bayesiana pela função MCMCglmm disponível no pacote MCMCglmm do software R. A convergência foi verificada pelo teste de Geweke disponível no software R. Após a obtenção dos componentes de variância foram estimados a herdabilidade e repetibilidade. A herdabilidade observada foi 0,28, o que sugere que a seleção para esta característica resultará em progresso genético eficiente no rebanho. A maior parte da variância fenotípica é devido a variância ambiental, com isso, o fenótipo dos indivíduos não é um bom indicador da variância genética aditiva. A repetibilidade foi de 0,93, indicando que o primeiro desempenho dos animais é considerado um bom indicador do segundo, podendo ser utilizadas em decisões de descarte.

Big Data to Knowledge: Application of Machine Learning to Predictive Modeling of Therapeutic Response in Cancer

Background

In recent years, the availability of high throughput technologies, establishment of large molecular patient data repositories, and advancement in computing power and storage have allowed elucidation of complex mechanisms implicated in therapeutic response in cancer patients. The breadth and depth of such data, alongside experimental noise and missing values, requires a sophisticated human-machine interaction that would allow effective learning from complex data and accurate forecasting of future outcomes, ideally embedded in the core of machine learning design.

Objective

In this review, we will discuss machine learning techniques utilized for modeling of treatment response in cancer, including Random Forests, support vector machines, neural networks, and linear and logistic regression. We will overview their mathematical foundations and discuss their limitations and alternative approaches in light of their application to therapeutic response modeling in cancer.

Conclusion

We hypothesize that the increase in the number of patient profiles and potential temporal monitoring of patient data will define even more complex techniques, such as deep learning and causal analysis, as central players in therapeutic response modeling.

Comparison of efficacy of various drugs used for medical expulsive therapy for distal ureter stones: A systematic review and network meta-analysis

OBJECTIVES

Medical expulsive therapy has been found to be effective for distal ureteric stones; however, which drug is most efficacious in terms of stone expulsion rate (SER) and stone expulsion time (SET) is not known. With this review we aimed to compare the efficacy of various drug treatments for distal ureter stones used as medical expulsive therapy in terms of SER and SET.

METHODS

Systematic literature search was conducted to include all the randomised study comparing various drug interventions for lower ureter stones. Standard preferred reporting items for systematic review and meta-analysis for network meta-analysis (PRISMA-NMA) were pursued.

RESULTS

In this review, 50 randomised studies with 12,382 patients were included. For stone expulsion rate (SER), compared with placebo all the treatment groups were more effective except nifedipine and sildenafil. According to the SUCRA values obtained, naftopidil plus steroid was the highest rank and nifedipine lowest. For stone expulsion time (SET), compared with placebo only tadalafil plus silodosin, nifedipine plus steroid, alfuzosin, silodosin, tadalafil and tamsulosin were more effective. SUCRA values were highest for tadalafil plus silodosin and least for naftopidil plus steroid. From subgroup analysis with individual drugs for SER, SUCRA values were highest for naftopidil followed by silodosin and SET was highest for silodosin and least for naftopidil.

CONCLUSION

For lower ureter stone, tadalafil plus silodosin is the best combination and silodosin best individual drug considering the SET and SER. Nifedipine as monotherapy is no more effective than control group.

Machine learning and statistical methods for predicting mortality in heart failure

Heart failure is a debilitating clinical syndrome associated with increased morbidity, mortality, and frequent hospitalization, leading to increased healthcare budget utilization. Despite the exponential growth in the introduction of pharmacological agents and medical devices that improve survival, many heart failure patients, particularly those with a left ventricular ejection fraction less than 40%, still experience persistent clinical symptoms that lead to an overall decreased quality of life. Clinical risk prediction is one of the strategies that has been implemented for the selection of high-risk patients and for guiding therapy. However, most risk predictive models have not been well-integrated into the clinical setting. This is partly due to inherent limitations, such as creating risk predicting models using static clinical data that does not consider the dynamic nature of heart failure. Another limiting factor preventing clinicians from utilizing risk prediction models is the lack of insight into how predictive models are built. This review article focuses on describing how predictive models for risk-stratification of patients with heart failure are built.