Bayesian Regression Research Articles

Associations between urinary phytoestrogen mixed metabolites and osteoarthritis risk

Background This study aims to explore the relationship between urinary phytoestrogen mixed metabolites and the risk of osteoarthritis (OA). Methods Using data from the National Health and Nutrition Examination Survey (NHANES), a Weighted Quantile Sum (WQS) regression analysis was conducted to determine the dominant metabolites. Additionally, a Bayesian kernel machine regression (BKMR) model was utilized to explore the combined effects of phytoestrogen mixed metabolites on OA. Results Compared to the lowest quartile group, the highest quartile group of Enterodiol showed a 46% increased risk of OA (OR = 1.46, 95% CI: 1.09–1.96), while the highest quartile group of Enterlactone showed a 30% decreased risk of OA (OR = 0.70, 95% CI: 0.52–0.96). The WQS regression model analysis revealed a positive relationship between urinary phytoestrogen mixed metabolites and OA risk, with Enterodiol found to have the highest weight in this association. The BKMR model indicated that the association between urinary phytoestrogens and OA increased with concentration but did not reach statistical significance. The univariate exposure-response function demonstrated a positive association between Enterodiol and OA. Conclusions There is a positive relationship between urinary phytoestrogen mixed metabolites and OA, with Enterodiol being an important factor influencing OA risk.

Market Predictability Before the Closing Bell Rings

This study examines the predictability of the last 30 min of intraday stock price movements within the US financial market. The analysis encompasses several potential explanatory variables, including returns from each 30 min intraday trading session, overnight returns, the federal reserve fund rate decision days and the subsequent three days, the US dollar index, month effects, weekday effects, and market volatilities. Market-adaptive trading strategies are developed and backtested on the basis of the study’s insights. Unlike the commonly employed multiple linear regression methods with Gaussian errors, this research utilizes a Bayesian linear regression model with Student-t error terms to more accurately capture the heavy tails characteristic of financial returns. A comparative analysis of these two approaches is conducted and the limitations inherent in the traditionally used method are discussed. Our main findings are based on data from 2007 to 2018. We observed that well-studied factors such as overnight effects and intraday momentum have diminished over time. Some other new factors were significant, such as lunchtime returns during boring days and the tug-of-war effect over the days after a federal fund rate change decision. Ultimately, we incorporate findings derived from data spanning 2022 to 2024 to provide a contemporary perspective on the examined components, followed by a discussion of the study’s limitations.

The Relationship Between Training Load and Injury in Competitive Swimming: A Two-Year Longitudinal Study

Training load monitoring is employed to quantify training demands, to determine individual physiological adaptions and to examine the dose–response relationship, ultimately reducing the likelihood of injury and making a meaningful impact on performance. The purpose of this study is to explore the relationship between training load and injury in competitive swimmers, using the session rate of perceived exertion (sRPE) method. Data were collected using a prospective, longitudinal study design across 104 weeks. Data were collected from 34 athletes centralised in two of Swim Ireland’s National Centres. Bayesian mixed effects logistic regression models were used to analyse the relationship between sRPE-TL and medical attention injuries. The average weekly swim volume was 33.5 ± 12.9 km. The weekly total training load (AU) averaged 3838 ± 1616.1. A total of 58 medical attention injury events were recorded. The probability of an association between training load and injury ranged from 70% to 98%; however, evidence for these relationships was deemed weak or highly uncertain. The findings suggest that using a single training load metric in isolation cannot decisively inform when an injury will occur. Instead, coaches should utilise monitoring tools to ensure that the athletes are exposed to an appropriate training load to optimise physiological adaptation. Future research should strive to investigate the relationship between additional risk factors (e.g., wellbeing, lifestyle factors or previous injury history), in combination with training load and injury, in competitive swimmers.

The spatial overlap between risk of intimate partner violence against women and suicide-related emergency calls

Abstract Intimate partner violence against women (IPVAW) is a global problem, and suicidal behaviour is one of its most serious consequences. However, no study has investigated the spatial overlap between the two. The aim of this paper is to analyse whether there is spatial overlap between IPVAW and suicide-related calls in the same neighbourhoods, and to identify common risk factors. Geocoded data on IPVAW (N = 2060) and suicide-related calls (N = 4916) from the city of Valencia, Spain, between 2019 and 2021 were collected from the 112 emergency service and the Spanish National Police Corps. We used two analytical approaches: (i) a Bayesian spatial Poisson regression modelling analysis to identify the risk factors associated with IPVAW and suicide-related calls; (ii) Bayesian joint spatial modelling to examine the common spatial distribution of these outcomes. Relative risk maps are also displayed and analysed. The study revealed a higher incidence of IPVAW and suicide-related calls in neighbourhoods with lower income and population density, and greater residential instability and immigrant concentration. There was a 70% similarity between high- and low-risk areas for these incidents, with a moderate positive correlation of 0.43. Notably, 98% of IPVAW risk variance was explained by shared factors, while 30% applied to suicide-related calls; high IPVAW risk was also found in peripheral areas, while suicide-related calls were concentrated in the city centre. The findings highlight the need to identify areas with shared risk and neighbourhood characteristics for the development of more effective intervention strategies.

Specification Choices in Quantile Regression for Empirical Macroeconomics

ABSTRACTQuantile regression has become widely used in empirical macroeconomics, in particular for estimating and forecasting tail risks. This paper examines various choices in the specification of quantile regressions for macro applications, including how and to what extent to include shrinkage and whether to apply shrinkage in a classical or Bayesian framework. We focus on forecasting accuracy, measured with quantile scores and quantile‐weighted continuous ranked probability scores at a range of quantiles from the left to right tail. Across applications, we find that shrinkage is generally helpful to quantile forecast accuracy, with Bayesian quantile regression dominating frequentist quantile regression.JEL Classification: C53, E17, E37, F47

Using three statistical methods to analyze the associations between a mixture of multi-nutrients and risk of mild cognitive impairment in an elderly population in Northern China

Few studies have considered nutrients as a mixture and their impact on Mild Cognitive Impairment (MCI). The generalized linear regression (GLM), weighted quantile sum (WQS) regression, and Bayesian kernel machine regression (BKMR) models are fitted to estimate the association between intake of a mixture of nutrients and MCI. Comparing the results from these three models, vitamin E and vitamin B6 were identified as the most important factors associated with the risk of MCI. Considering the characteristics of BKMR, it may be more advantageous to use BKMR to estimate the combined the joint effects of nutrients mixture. In the future, studies need to move from a “one nutrient at a time” approach to simultaneous analyses of multiple nutrients intakes in order to understand and quantify the joint effect of nutrients mixture on health.

A Decade-Long Comparison of Heavy Metal(loid)s in the River and Children’s Health Risk Assessment of an E-Waste Recycling Area

Informal e-waste recycling can cause serious heavy metal(loid) pollution to nearby bodies of water, thereby increasing the risk of heavy metal exposure to local residents. This study investigates metal(loid) pollution levels in the Lianjiang River, which flows through Guiyu, an e-waste recycling town in Shantou, Guangdong. In 2009 and 2020, water samples from the Lianjiang River were taken, and the levels of 12 metal(loid)s (V, Cr, Mn, Sr, Ni, Cu, Zn, As, Se, Cd, Ba, and Pb) were measured by ICP-MS. In total, 380 valid questionnaires from Guiyu kindergarten children were selected to assess child health. Exposure health risks for children were calculated using two forms of exposure (dermal and inhalation) and statistical methods (multiple linear regression and Bayesian kernel machine regression—BKMR). The concentrations of Mn, Ni, Pb, and Cd in 2009 were significantly higher than those in 2020. The concentration of heavy metal(loid)s had been decreasing in 2020. The non-carcinogenic risk levels of the 12 metal(loid)s in both exposure routes were at an acceptable risk level. The average carcinogenic risk levels for As, Cr, Ni, Pb, and Cd exceeded the ILCRi <10–6 acceptable range. According to MLR, it was found that daily oral intake doses of Pb [β(95% CI): −0.949 (−1.596, −0.863), p < 0.001 and Se [β(95% CI): −0.911 (−1.888, −0.092), p = 0.031] were negatively associated with body mass index. A BKMR model was developed, through which the synergistic effects of co-exposure to 12 heavy metal(loid)s on growth and development indicators in children were analyzed. Concentrations of heavy metal(loid)s in rivers near e-waste recycling sites have been decreasing for 10 years. It was found that the growth and development of children are affected by the intake of heavy metal(loid)s in water. The reduction in heavy metal(loid) contamination in e-waste recycling areas needs to be continued, and concern about its impact on children’s health must remain.

Development of a recommendation engine for university study programme selection: A regression-based approach

This paper is the third in a series focused on bridging the gap between secondary and higher education. Our primary objective is to develop a robust theoretical framework for an innovative e-business model called the Undergraduate Study Programme Search System (USPSS). This system considers multiple criteria to reduce the likelihood of exam failure or the need for multiple retakes, while maximizing the chances of successful program completion. Testing of the proposed algorithm demonstrated that the Stochastic Gradient Boosted Regression Trees method outperforms the current method used in Lithuania for admitting applicants to 47 educational programs. Specifically, it is more accurate than the Probabilistic Neural Network for 25 programs, the Ensemble of Regression Trees for 24 programs, the Single Regression Tree for 18 programs, the Random Forest Regression for 16 programs, the Bayesian Additive Regression Trees for 13 programs, and the Regression by Discretization for 10 programs.

Association of joint exposure to organophosphorus flame retardants and phthalate acid esters with gestational diabetes mellitus: a nested case-control study

BackgroundOrganic phosphate flame retardants (OPFRs) and phthalate acid esters (PAEs) are common endocrine-disrupting chemicals that cause metabolic disorders. This study aimed to assess the association between joint exposure to OPFRs and PAEs during early pregnancy in women with gestational diabetes mellitus (GDM).MethodsSeven OPFRs and five PAEs were detected in the urine of 65 GDM patients and 100 controls using gas chromatography-tandem triple quadrupole mass spectrometry (GC-MS). The association of OPFRs and PAEs with GDM was assessed using logistic regression, weighted quantile sum (WQS) regression, and Bayesian kernel machine regression (BKMR) models.ResultsLevels of dibutyl phthalate (DBP), di-2-ethylhexyl phthalate (DEHP), diethyl phthalate (DEP), dimethyl phthalate (DMP), tris (2-butoxyethyl) phosphate (TBEP), tributyl phosphate (TBP), tris (2-chloroethyl) phosphate (TCEP), tris (1,3-dichloro-2-propyl) phosphate (TDCPP), tri-ortho-cresyl phosphate (TOCP), and triphenyl phosphate (TPHP) increased in the GDM group, and the OPFRs and PAEs, except for BBP and TMCP, were associated with GDM in the logistic regression analysis. In the WQS model, the mixture of OPFRs and PAEs was significantly positively associated with GDM (OR = 3.29, 95%CI = 1.27–8.51, P = 0.014), with TDCPP having the highest WQS index weight. BKMR analysis reinforced these results, showing that the overall association of joint exposure to the OPFRs and PAEs with GDM increased at exposure levels of the 55th to 75th percentiles. Independent exposure to TDCPP (OR = 1.42, 95%CI = 1.09–1.86, P = 0.011) and TBEP (OR = 1.29, 95%CI = 1.04–1.60, P = 0.023) were associated with an increased risk of GDM.ConclusionsEnvironmental exposure to OPFRs and PAEs is significantly associated with GDM. These findings provide evidence for the adverse effects of exposure to OPFRs and PAEs on the health of pregnant women.

The association between circulating short-chain fatty acids and blood pressure in Chinese elderly population

The gut microbiome primarily generates short-chain fatty acids (SCFAs) by fermenting dietary fibers. Though previous studies have linked SCFAs to blood pressure, there remains a lack of research on the relationship between SCFAs levels in the serum of elderly individuals and blood pressure. Based on this, we investigated the associations of serum SCFAs with blood pressure in Chinese older adults in a cross-sectional study. In this report, we recruited 1013 older adults over 60 years of age from June to September 2016 in Lu 'an City, China. Using Ultra High Performance Liquid Chromatography-Quadrupole-Exactive-Orbitrap-Mass Spectrometry (UHPLC-QE-Orbitrap MS), we measured the level of various SCFAs, including acetic acid (AA), propanoic acid (PA), butyric acid (BA), isobutyric acid (iso-BA), valeric acid (VA), isovaleric acid (iso-VA), and caproic acid (CA), in serum samples collected from Chinese elderly adults. The study recruited 1013 older adults in total. Multiple logistic regression analysis shows that AA (OR = 0.696, 95%CI: 0.501–0.966) and VA (OR = 0.713, 95%CI: 0.516–0.985) are negatively associated with hypertension. Linear regression analysis shows a negative correlation between AA (β = -3.89, 95% CI: -7.12 – -0.66) and the systolic blood pressure (SBP) levels, and a significant negative association between iso-VA (β = -2.11, 95% CI: -3.94 – -0.29) and diastolic blood pressure (DBP) levels. Whether in unadjusted or adjusted linear regression models, we all observe significant positive associations between CA and blood pressure levels. In the Bayesian kernel-machine regression (BKMR) models, the trends between the mixture of SCFAs and hypertension, SBP are inverse, but not significant; we also observe a significant negative correlation between AA and SBP, and a significant negative association between iso-VA and DBP levels, while CA is significantly positively correlated with SBP and DBP. Collectively, our results advocate for considering SCFA as a potential intervention to lower blood pressure, and especially AA may be a possible target for research. This may provide new perspectives for understanding the role of SCFAs in hypertension.

Study on association between human plasma polybrominated diphenyl ethers (PBDEs) and thyroid homeostasis and carcinoma.

Polybrominated diphenyl ethers (PBDEs), a category of persistent organic pollutants (POPs), are ubiquitous in the environment and have been reported to have endocrine-disrupting and tumor-promoting activities. However, the evidence for the correlation between plasma PBDEs levels, thyroid homeostasis and thyroid carcinoma in humans remains limited. Herein, we analyzed eight PBDE congeners in 53 patients undergoing thyroid surgery. BDE-209 was identified as the most abundant PBDE congener in plasma (median, 11.36ng/g lipid). BDE-100 concentration was positively associated with free triiodothyronine (FT3), indicating a potential interference with thyroid function. Point-biserial correlation analysis revealed positive associations between certain plasma PBDE congeners (BDE-28, BDE-47, BDE-99, BDE-100, and BDE-154) and aggressive pathological parameters. There was no significant correlation between PBDEs and BRAF or RAS mutations in PTC patients, indicating that PBDEs may not directly promote the initiation and progression of thyroid cancer through these genetic mutations. It implies the complexity of the relationship between PBDEs exposure and thyroid cancer development. Although not statistically significant, Bayesian kernel-machine regression (BKMR) analysis of single-exposure effects model found that BDE-47 and BDE-99 were positively associated with the risk of malignant neoplasms. The present study not only contributes to the growing evidence regarding the impact of PBDEs on thyroid function but also provides new insights into the association between exposure to certain PBDE congeners and the aggressive pathological parameters of thyroid cancer. Large-scale prospective studies are still needed to support our findings.

Oxidative stress mediates the association of organophosphate flame retardants with metabolic obesity in U.S. adults: A combined epidemiologic and bioinformatic study

Obesity is a global public health issue, with limited epidemiologic studies on the relationship and mechanisms between organophosphate flame retardants (OPFRs) and metabolic obesity phenotypes (MOPs). We aimed to explore the link between OPFRs metabolite (m-OPFRs) and MOPs using a combined epidemiologic and bioinformatic approach. We used cross-sectional survey data from the U.S. National Health and Nutrition Examination Survey (2011–2018) to analyze the relationship between m-OPFRs and metabolic health obesity (MHO), as well as metabolic unhealthy obesity (MUO). The dataset encompasses eligible adults to assess the impact of individual, mixed, and mediated effects on the outcome variables through multivariate logistic regression, Bayesian kernel machine regression (BKMR), and mediation analysis. Multiple logistic regression models, stratified by tertiles of exposure showed that bis(1,3-dichloro-2-propyl) phosphate (BDCIPP) levels in the body significantly increased the risk of MHO, with OR and 95%CI of 1.454 (1.082, 1.953) for the second tertile (T2) and 1.598 (1.126, 2.268) for the third tertile (T3), compared to the first tertile (T1). Increased levels of BDCIPP in T3 (1.452(1.013, 2.081)) are associated with MUO, compared to T1. Mixed m-OPFRs and MHO risk in BMKR were positively correlated, with BDCIPP being the primary contributor. We found that the serum uric acid (SUA) and white blood cell count (WBC) indicators significantly mediated the association between BDCIPP and MHO (P < 0.05). Our study suggests that OPFRs, either individual or mixed, are associated with two distinct MOPs, with oxidative stress playing an important role. In addition, in silico analysis was used to screen for shared genes, and eight shared genes and eleven biological pathways identified during the screening process were used to construct the adverse outcome pathway, which suggests that exposure to OPFRs may activate the peroxisome proliferator-activated receptor (PPAR) pathway, thereby increasing the risk of obesity. Further studies are needed to validate our findings.

Effects of fine particulate matter bound heavy metals on intentional self-harm deaths in Guangzhou, China, insight from core chemical constituents.

Fine particulate matter (PM2.5) is one of the major threats to human health, and may partly responsible for intentional self-harm deaths, while the limited results seemed contradictory. Further analysis on PM2.5 constituents may provide more reliable evidence. Heavy metals are crucial toxic components of PM2.5 that may induce suicide behavior. What role do PM2.5-bound heavy metals play in a threat to intentional self-harm death is still unclear. Two-year data of daily PM2.5-bound heavy metals (including metalloids) and daily intentional self-harm deaths were collected in Guangzhou. Bayesian kernel machine regression, weighted quantile sum, and quantile-based g-computation models were employed to depict the relationships between heavy metals and intentional self-harm deaths. The number of intentional self-harm deaths was 217 and 283 for 2015 and 2016, respectively. A positive correlation was found between the combined effect of the 13 heavy metals and intentional self-harm deaths. Nickel, cadmium, and iron were the primary contributors to this positive correlation. Heavy metal components play significant roles in PM2.5-related intentional self-harm deaths, and targeted source control measures are warranted to protect residents from suicide.

Quadrivalent Conjugate Vaccine and Invasive Meningococcal Disease in US Adolescents and Young Adults

Beginning in 2005, the US implemented routine immunization of adolescents with a quadrivalent conjugate vaccine (MenACWY) for the prevention of invasive meningococcal disease (IMD). To assess whether MenACWY immunization was associated with a reduced IMD burden among the US adolescent population and how the downward trajectory of IMD that began in the mid-1990s might have evolved in the absence of vaccination efforts. In this decision analytical study, a bayesian hierarchical Poisson regression model was developed to investigate the potential trajectory of IMD among US adolescents and young adults without vaccination and evaluate the direct association of vaccination with IMD burden. The model included the entire age-stratified US population and was fitted to national incidence data for serogroups C, W, and Y from January 1, 2001, to December 31, 2021, with stratification by vaccination status for IMD cases. Simulated counterfactual scenario of absent vaccination from 2005 to 2021, while retaining the incidence rate of IMD for unvaccinated individuals estimated during model fitting. The main outcomes were the estimated numbers of IMD cases and deaths averted by MenACWY vaccination among US adolescents and young adults aged 11 to 23 years. Among the entire US population from 2005 to 2021, MenACWY vaccination prevented an estimated 172 (95% credible interval [CrI], 85-345) cases of IMD among US adolescents 11 to 15 years of age and 328 (95% CrI, 164-646) cases of IMD among those aged 16 to 23 years. Absent vaccination, the cumulative incidence of IMD in these age groups would have been at least 59% higher than reported over the same period with vaccination. Using case fatality rates of unvaccinated individuals derived from national data, vaccination averted an estimated 16 (95% CrI, 8-31) deaths among adolescents aged 11 to 15 years and 38 (95% CrI, 19-75) deaths among those aged 16 to 23 years. This decision analytical model suggests that the MenACWY vaccination program in the US was associated with a reduced burden of meningococcal disease. Without vaccination, the incidence rates per 100 000 adolescents and young adults would have been substantially higher than those observed during the vaccine era.

Change in striatal functional connectivity networks across 2 years due to stimulant exposure in childhood ADHD: results from the ABCD sample.

Widely prescribed for Attention-Deficit/Hyperactivity Disorder (ADHD), stimulants (e.g., methylphenidate) have been studied for their chronic effects on the brain in prospective designs controlling dosage and adherence. While controlled approaches are essential, they do not approximate real-world stimulant exposure contexts where medication interruptions, dosage non-compliance, and polypharmacy are common. Brain changes in real-world conditions are largely unexplored. To fill this gap, we capitalized on the observational design of the Adolescent Brain Cognitive Development (ABCD) study to examine effects of stimulants on large-scale bilateral cortical networks' resting-state functional connectivity (rs-FC) with 6 striatal regions (left and right caudate, putamen, and nucleus accumbens) across two years in children with ADHD. Bayesian hierarchical regressions revealed associations between stimulant exposure and change in rs-FC of multiple striatal-cortical networks, affiliated with executive and visuo-motor control, which were not driven by general psychotropic medication. Of these connections, three were selective to stimulants versus stimulant naive: reduced rs-FC between caudate and frontoparietal network, and between putamen and frontoparietal and visual networks. Comparison with typically developing children in the ABCD sample revealed stronger rs-FC reduction in stimulant-exposed children for putamen and frontoparietal and visual networks, suggesting a normalizing effect of stimulants. 14% of stimulant-exposed children demonstrated reliable reduction in ADHD symptoms, and were distinguished by stronger rs-FC reduction between right putamen and visual network. Thus, stimulant exposure for a two-year period under real-world conditions modulated striatal-cortical functional networks broadly, had a normalizing effect on a subset of networks, and was associated with potential therapeutic effects involving visual attentional control.

Intelligent Helicopter Turbine Engine Fault Diagnosis Using Multi-Head Attention

A turbine engine provides power to the helicopter, enabling the helicopter to travel and hover in the air. Since the rotorcraft operates at high altitudes, ensuring safety and maintaining a healthy operational status are crucial at all times. Therefore, a prognostics and health management (PHM) system for the turbine engine must be implemented to predict any anomalies or faults to prevent catastrophic accidents. This research proposes a novel fault diagnosis method for helicopter turbine engines based on operational data acquired from actual aircraft. First, the proposed method predicts engine torque using other operational data while accounting for uncertainty. A Bayesian regression approach is employed to predict the engine torque. The torque margin, defined as the difference between the actual torque and the estimated torque, is then used to diagnose engine faults. Specifically, a multi-head attention mechanism is incorporated to capture interactions between various engine parameters. Additionally, domain adaptation techniques are applied to enhance the model's generalization performance, ensuring robustness across diverse operating conditions. The proposed method is validated using seven different datasets, each acquired from a helicopter engine. Four datasets were used for training, while the remaining three were allocated for testing and validation. The results indicated that the proposed method accurately predicted torque. Furthermore, the fault diagnosis showed promising results, leading to a 3rd-place finish in the 2024 PHM Society Data Challenge in terms of validation score.

Population-based cross-sectional survey of cervical cancer screening prevalence and socio-demographic correlates in Bangladeshi women.

Cervical cancer, albeit preventable, is the second-most deadly gynecological cancer in developing nations. Little is known about cervical cancer among Bangladeshi women. This study aims to estimate the prevalence of cervical cancer screening and demographic correlates to identify potential variabilities in screening rates among different demographic groups and regions. This study used secondary data from the WHO STEPS 2018 Survey. We used Bayesian regression to perform the bivariate analyses between the outcome and each explanatory factor, as it generates more acceptable results and improves parameter estimates. The top-ranked socio-demographic factors were identified using a two-step cluster analysis. This method determines the relevance of predictor variables and automatically establishes the number of clusters. The prevalence of Bangladeshi women who had ever been screened for cervical cancer was 6.2%. In the adjusted model, women with the following socio-demographic factors had a higher likelihood of developing cervical cancer: being 18-29 years old (AOR = 3.3, 95% CI: 0.24, 15.27) or 45-59 years old (AOR = 2.8, 95% CI: 1.22, 6.0), currently married (AOR = 2.3, 95% CI: 1.36, 3.70), and employed (AOR = 2.4, 95% CI: 1.40, 4.06). Women in the Barisal division were found to have higher odds of being screened for cervical cancer (AOR = 21, 95% CI: 0.66, 121.97). Cluster analysis found residence status predisposes women to cervical cancer screening. There is a significant potential for substantial reductions in the burden of cervical cancer in Bangladesh by strengthening the application of cervical cancer screening. Future studies should examine how socioeconomic status, culture, and healthcare access affect cervical cancer screening trends for different divisions in Bangladesh. An independent national cancer registry is urgently needed to evaluate screening trends and outcomes.

Impact of active case finding for tuberculosis with mass chest X-ray screening in Glasgow, Scotland, 1950-1963: An epidemiological analysis of historical data.

Community active case finding (ACF) for tuberculosis was widely implemented in Europe and North America between 1940 and 1970, when incidence was comparable to many present-day high-burden countries. Using an interrupted time series analysis, we analysed the effect of the 1957 Glasgow mass chest X-ray campaign to inform contemporary approaches to screening. Case notifications for 1950 to 1963 were extracted from public health records and linked to demographic data. We fitted Bayesian multilevel regression models to estimate annual relative case notification rates (CNRs) during and after a mass screening intervention implemented over 5 weeks in 1957 compared to the counterfactual scenario where the intervention had not occurred. We additionally estimated case detection ratios and incidence. From 11 March 1957 to 12 April 1957, 714,915 people (622,349 of 819,301 [76.0%] resident adults ≥15 years) were screened with miniature chest X-ray; 2,369 (0.4%) were diagnosed with tuberculosis. Pre-intervention (1950 to 1956), pulmonary CNRs were declining at 2.3% per year from a CNR of 222/100,000 in 1950. With the intervention in 1957, there was a doubling in the pulmonary CNR (RR: 1.95, 95% uncertainty interval [UI] [1.81, 2.11]) and 35% decline in the year after (RR: 0.65, 95% UI [0.59, 0.71]). Post-intervention (1958 to 1963) annual rates of decline (5.4% per year) were greater (RR: 0.77, 95% UI [0.69, 0.85]), and there were an estimated 4,599 (95% UI [3,641, 5,683]) pulmonary case notifications averted due to the intervention. Effects were consistent across all city wards and notifications declined in young children (0 to 5 years) with the intervention. Limitations include the lack of data in historical reports on microbiological testing for tuberculosis, and uncertainty in contributory effects of other contemporaneous interventions including slum clearances, introduction of BCG vaccination programmes, and the ending of postwar food rationing. A single, rapid round of mass screening with chest X-ray (probably the largest ever conducted) likely resulted in a major and sustained reduction in tuberculosis case notifications. Synthesis of evidence from other historical tuberculosis screening programmes is needed to confirm findings from Glasgow and to provide insights into ongoing efforts to successfully implement ACF interventions in today's high tuberculosis burden countries and with new screening tools and technologies.

Evaluation of Bayesian Linear Regression models for gene set prioritization in complex diseases.

Genome-wide association studies (GWAS) provide valuable insights into the genetic architecture of complex traits, yet interpreting their results remains challenging due to the polygenic nature of most traits. Gene set analysis offers a solution by aggregating genetic variants into biologically relevant pathways, enhancing the detection of coordinated effects across multiple genes. In this study, we present and evaluate a gene set prioritization approach utilizing Bayesian Linear Regression (BLR) models to uncover shared genetic components among different phenotypes and facilitate biological interpretation. Through extensive simulations and analyses of real traits, we demonstrate the efficacy of the BLR model in prioritizing pathways for complex traits. Simulation studies reveal insights into the model's performance under various scenarios, highlighting the impact of factors such as the number of causal genes, proportions of causal variants, heritability, and disease prevalence. Comparative analyses with MAGMA (Multi-marker Analysis of GenoMic Annotation) demonstrate BLR's superior performance, especially in highly overlapped gene sets. Application of both single-trait and multi-trait BLR models to real data, specifically GWAS summary data for type 2 diabetes (T2D) and related phenotypes, identifies significant associations with T2D-related pathways. Furthermore, comparison between single- and multi-trait BLR analyses highlights the superior performance of the multi-trait approach in identifying associated pathways, showcasing increased statistical power when analyzing multiple traits jointly. Additionally, enrichment analysis with integrated data from various public resources supports our results, confirming significant enrichment of diabetes-related genes within the top T2D pathways resulting from the multi-trait analysis. The BLR model's ability to handle diverse genomic features, perform regularization, conduct variable selection, and integrate information from multiple traits, genders, and ancestries demonstrates its utility in understanding the genetic architecture of complex traits. Our study provides insights into the potential of the BLR model to prioritize gene sets, offering a flexible framework applicable to various datasets. This model presents opportunities for advancing personalized medicine by exploring the genetic underpinnings of multifactorial traits.

Bayesian additive regression trees for predicting childhood asthma in the CHILD cohort study

BackgroundAsthma is a heterogeneous disease that affects millions of children and adults. There is a lack of objective gold standard diagnosis that spans the ages; instead, diagnoses are made by clinician assessment based on a cluster of signs, symptoms and objective tests dependent on age. Yet, there is a clear morbidity associated with chronic asthma symptoms. Machine learning has become a popular tool to improve asthma diagnosis and classification. There is a paucity of literature on the use of Bayesian machine learning algorithms to predict asthma diagnosis in children. This paper develops a prediction model using the Bayesian additive regression trees (BART) and compares its performance to various machine learning algorithms in predicting the diagnosis of childhood asthma.MethodsClinically relevant variables collected at or before 3 years of age from 2794 participants in the CHILD Cohort Study were used to predict physician-diagnosed asthma at age 5. BART and six other commonly used machine learning algorithms, namely adaptive boosting, logistic regression, decision tree, neural network, random forest, and support vector machine were trained. Measures of performance including sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve were calculated. The confidence intervals were calculated using Bootstrapping samples. Important predictors and interaction effects associated with asthma were also identified using BART.ResultsBART, logistic regression and random forest showed the highest area under the ROC curve compared to other machine learning algorithms. Based on BART, recurrent wheeze, respiratory infection and food sensitization at 3 years of age were the most important predictors. The three most important interaction effects were found to be interaction terms of respiratory infection at 3 years and recurrent wheezing at 3 years, maternal asthma and paternal asthma, and maternal wheezing and inhalant sensitization of child at 3 years.ConclusionsBART demonstrated promising prediction performance when compared to other machine learning algorithms. Future research could validate the BART in an external cohort to evaluate its reliability and generalizability.