A Bayesian network model for predicting cardiovascular risk

Relationship between insulin resistance surrogate markers with diabetes and dyslipidemia: A Bayesian network analysis of Korean adults

Insulin resistance (IR) can be optimally assessed using the euglycemic clamp, but practical clinical limitations necessitate surrogate markers. This study leveraged the Bayesian network analysis to evaluate three established IR markers: the Homeostatic Model Assessment of IR (HOMA-IR) using insulin level and fasting blood glucose (FBG), TG-Glucose (TyG) index using triglycerides (TG) and FBG, and TG-to-HDL ratio (TG/HDL ratio) using TG and high-density lipoprotein (HDL), based on the Korean National Health and Nutrition Examination Survey data (2019-2021). Our analysis revealed a sequential association pattern (TG/HDL ratio → TyG index → HOMA-IR), positioning the TyG index as a central connecting marker. The HOMA-IR exhibited strong predictive power for diabetes, while the TG/HDL ratio was most effective for assessing dyslipidemia. However, both had limited crossover utility. In contrast, the TyG index bridged this gap, demonstrating robust predictive capability for both conditions. The Markov blanket analysis illuminated the distinctive metabolic signatures of each marker: The TyG index displayed balanced glucose-lipid metabolic contributions, the HOMA-IR predominantly reflected glucose metabolism and obesity characteristics, and the TG/HDL ratio emphasized lipid metabolism. Notably, the TyG index's predictive performance showed significant enhancement when integrated with obesity information, contrasting with the HOMA-IR's minimal response owing to its inherent incorporation of obesity characteristics. These findings position the TyG index as a superior clinical marker, offering both comprehensive predictive capability and enhanced performance through synergistic integration with obesity measures. While each marker demonstrated reliability, the TyG index's unique combination of versatility and scalability establishes it as an effective tool for comprehensive metabolic risk assessment.

A Bayesian Belief Network model for the estimation of risk of cardiovascular events in subjects with type 1 diabetes

OBJECTIVES

Cardiovascular diseases (CVDs) represent a major risk for people with type 1 diabetes (T1D). Our aim here is to develop a new methodology that overcomes some of the problems and limitations of existing risk calculators. First, they are rarely tailored to people with T1D and, in general, they do not deal with missing values for any risk factor. Moreover, they do not take into account information on risk factors dependencies, which is often available from medical experts.

METHOD

This study introduces a Bayesian Belief Network (BBN) model to quantify CVD risk in individuals with T1D. The developed methodology is applied to a large T1D dataset and its performances are assessed. A simulation study is also carried out to quantify the parameter estimation properties.

RESULTS

The performances of individual risk estimation, as measured by the area under the ROC curve and by the C-index, are about 0.75 for both real and simulated data with comparable sample sizes.

CONCLUSIONS

We observe a good predictive ability of the proposed methodology with accurate parameter estimation. The BBN approach takes into account causal relationships between variables, providing a comprehensive description of the system. This makes it possible to derive useful tools for optimising intervention.

Causal associations between scapular morphology and shoulder condition estimated with Bayesian statistics

BACKGROUND AND OBJECTIVE

While there is a reported correlation between shoulder condition and scapular morphology, the precise impact of typical anatomical variables remains a subject of ongoing debate. This study aimed to evaluate this causal association, by emphasizing the importance of scientific modeling before statistical analysis.

METHODS

We examined the effect of scapular anatomy on shoulder condition, and conditioning on sex, age, height, and weight. We considered the two most common pathologies: primary osteoarthritis (OA) and cuff tear arthropathy (CTA). We combined the other pathologies into a single category (OTH) and included a control category (CTRL) of adult subjects without pathology. We represented acromion and glenoid morphology by acromion angle (AA), acromion posterior angle (APA), acromion tilt angle (ATA), glenoid inclination angle (GIA), and glenoid version angle (GVA). GVA was negative for posterior orientation. These variables were automatically calculated from CT scans of 396 subjects in the 4 shoulder condition groups by a deep learning model. We applied do-calculus to assess the identifiability of the causal associations and used a multinomial logistic regression Bayesian model to estimate them. To isolate the effect of each anatomical variable on each shoulder condition, we increased it from -2 to 2 z-score while constraining all other variables to their average value, and reported the effect on shoulder condition probability as percentage points (pp) for females and males.

RESULTS

Increasing AA reduced the probability of OA by 44 pp for females and 17 pp for males while increasing the probability of CTA by 36 pp for females and 33 pp for males. Increasing APA raised the probability of OA by 15 pp for females and 4 pp for males and increased the probability of CTA by 12 pp for females and 4 pp for males. Increasing ATA increased the probability of OA by 15 pp for females but decreased it by 25 pp for males, while also raising the probability of CTA by 11 pp for females and 21 pp for males. Increasing GIA decreased the probability of OA by 55 pp for females and 23 pp for males while increasing the probability of CTA by 45 pp for females and 31 pp for males. GVA (more anterior), decreased the probability of OA by 33 pp for females and 63 pp for males. The effects of APA and ATA were less important compared to the other variables. Overall, morphological effects were more pronounced for females than for males, except for GVA's impact on OA.

CONCLUSIONS

We developed a Bayesian causal model to answer interventional questions about the scapular anatomy's effect on shoulder condition. Our results, consistent with clinical knowledge, hold promise for aiding in early pathology detection and optimizing surgical planning within clinical settings.

The secrets to high-level green technology innovation of China's waste power battery recycling enterprises

Green technology innovation (GTI) in China's waste power battery recycling (WPBR) sector is a key driver for sustainable resource management, environmental protection, and economic prosperity. Using the PSR-BN-GPT-4 model and multi-source data, this study explores China's WPBRenterprises' high-level GTI mechanism. The research concludes that (1) Compared to traditional expert knowledge, the Bayesian network model based on GPT-4 exhibits superior causal reasoning capability. (2) The current level of GTI in China's WPBR industry is relatively low, with the probability of high-level GTI being only 19%. (3) Key factors identified include incentives like R&D investment, bottlenecks such as green finance policy tools, and hindrances like government procurement policy tools. (4) "Supporting Infrastructure Policy Tools - Recycling Outlets Number - Market Potential -Green Technology Innovation" and "Green Finance Policy Tools - R&D Investment - Green Technology Innovation" are two critical paths for enhancing the high-level development of GTI in WPBR enterprises. The study offers valuable insights for governmental, industrial, and corporate decision-making regarding GTI in battery recycling.

Predicting risk factors for Epstein-Barr virus reactivation using Bayesian network analysis: a population-based study of high-risk areas for nasopharyngeal cancer

Background and objective

Nasopharyngeal carcinoma (NPC) is a rare disease in most parts of the world, but it is highly prevalent in South China. Epstein-Barr virus (EBV) is one of the major risk factors for NPC. Hence, understanding the factors associated with the reactivation of EBV from the latent stage is crucial for preventing NPC. This study aimed to investigate the risk factors for EBV reactivation associated with NPC in high-prevalence areas in China using a Bayesian network (BN) model combined with structural equation modeling tools.

Methods

The baseline information for this study was derived from NPC screening data from a population-based prospective cohort in Sihui City, Guangdong Province, China. We divided the data into a training dataset and a test dataset. We then constructed an interaction networktionba BN prediction model to explore the risk factors for EBV reactivation, which was compared with a conventional logistic regression model.

Results

A total of 12,579 participants were included in the analyses, with 1596 participant pairs finally included after the use of a nested case-control study. The results of multivariable logistic regression showed that only being older than 60 years (OR = 1.718, 95% CI = 1.273,2.322) and being a current smoker (OR = 1.477, 95% CI = 1.167 - 1.872) were the risk factors for EBV reactivation. The results of the model constructed using BN showed that age and smoking were directly associated with EBV reactivation. In contrast, sex, education level, tea drinking, cooking, and family history of cancer were indirectly associated with EBV reactivation. Further, we predicted the risk of EBV reactivation using Bayesian inference and visualized the BN inference. Model prediction performance was evaluated using the test dataset. The results showed that the BN model slightly outperformed the traditional logistic regression model in all metrics.

Conclusions

BN not only reflects the complex interaction between factors but also visualizes the prediction results. It has a promising application potential in the risk prediction of EBV reactivation associated with NPC.

Colorectal cancer risk mapping through Bayesian networks

BACKGROUND AND OBJECTIVE

Only about 14% of eligible EU citizens finally participate in colorectal cancer (CRC) screening programs despite it being the third most common type of cancer worldwide. The development of CRC risk models can enable predictions to be embedded in decision-support tools facilitating CRC screening and treatment recommendations. This paper develops a predictive model that aids in characterizing CRC risk groups and assessing the influence of a variety of risk factors on the population.

METHODS

A CRC Bayesian Network is learnt by aggregating extensive expert knowledge and data from an observational study and making use of structure learning algorithms to model the relations between variables. The network is then parametrised to characterize these relations in terms of local probability distributions at each of the nodes. It is finally used to predict the risks of developing CRC together with the uncertainty around such predictions.

RESULTS

A graphical CRC risk mapping tool is developed from the model and used to segment the population into risk subgroups according to variables of interest. Furthermore, the network provides insights on the predictive influence of modifiable risk factors such as alcohol consumption and smoking, and medical conditions such as diabetes or hypertension linked to lifestyles that potentially have an impact on an increased risk of developing CRC.

CONCLUSION

CRC is most commonly developed in older individuals. However, some modifiable behavioral factors seem to have a strong predictive influence on its potential risk of development. Modeling these effects facilitates identifying risk groups and targeting influential variables which are subsequently helpful in the design of screening and treatment programs.

Bayesian network model of ethno-racial disparities in cardiometabolic-based chronic disease using NHANES 1999-2018

Background

Ethno-racial disparities in cardiometabolic diseases are driven by socioeconomic, behavioral, and environmental factors. Bayesian networks offer an approach to analyze the complex interaction of the multi-tiered modifiable factors and non-modifiable demographics that influence the incidence and progression of cardiometabolic disease.

Methods

In this study, we learn the structure and parameters of a Bayesian network based on 20 years of data from the US National Health and Nutrition Examination Survey to explore the pathways mediating associations between ethno-racial group and cardiometabolic outcomes. The impact of different factors on cardiometabolic outcomes by ethno-racial group is analyzed using conditional probability queries.

Results

Multiple pathways mediate the indirect association from ethno-racial group to cardiometabolic outcomes: (1) ethno-racial group to education and to behavioral factors (diet); (2) education to behavioral factors (smoking, physical activity, and-via income-to alcohol); (3) and behavioral factors to adiposity-based chronic disease (ABCD) and then other cardiometabolic drivers. Improved diet and physical activity are associated with a larger decrease in probability of ABCD stage 4 among non-Hispanic White (NHW) individuals compared to non-Hispanic Black (NHB) and Hispanic (HI) individuals.

Conclusion

Education, income, and behavioral factors mediate ethno-racial disparities in cardiometabolic outcomes, but traditional behavioral factors (diet and physical activity) are less influential among NHB or HI individuals compared to NHW individuals. This suggests the greater contribution of unmeasured individual- and/or neighborhood-level structural determinants of health that impact cardiometabolic drivers among NHB and HI individuals. Further study is needed to discover the nature of these unmeasured determinants to guide cardiometabolic care in diverse populations.

An innovative artificial intelligence-based method to compress complex models into explainable, model-agnostic and reduced decision support systems with application to healthcare (NEAR)

BACKGROUND AND OBJECTIVE

In everyday clinical practice, medical decision is currently based on clinical guidelines which are often static and rigid, and do not account for population variability, while individualized, patient-oriented decision and/or treatment are the paradigm change necessary to enter into the era of precision medicine. Most of the limitations of a guideline-based system could be overcome through the adoption of Clinical Decision Support Systems (CDSSs) based on Artificial Intelligence (AI) algorithms. However, the black-box nature of AI algorithms has hampered a large adoption of AI-based CDSSs in clinical practice. In this study, an innovative AI-based method to compress AI-based prediction models into explainable, model-agnostic, and reduced decision support systems (NEAR) with application to healthcare is presented and validated.

METHODS

NEAR is based on the Shapley Additive Explanations framework and can be applied to complex input models to obtain the contributions of each input feature to the output. Technically, the simplified NEAR models approximate contributions from input features using a custom library and merge them to determine the final output. Finally, NEAR estimates the confidence error associated with the single input feature contributing to the final score, making the result more interpretable. Here, NEAR is evaluated on a clinical real-world use case, the mortality prediction in patients who experienced Acute Coronary Syndrome (ACS), applying three different Machine Learning/Deep Learning models as implementation examples.

RESULTS

NEAR, when applied to the ACS use case, exhibits performances like the ones of the AI-based model from which it is derived, as in the case of the Adaptive Boosting classifier, whose Area Under the Curve is not statistically different from the NEAR one, even the model's simplification. Moreover, NEAR comes with intrinsic explainability and modularity, as it can be tested on the developed web application platform (https://neardashboard.pythonanywhere.com/).

CONCLUSIONS

An explainable and reliable CDSS tailored to single-patient analysis has been developed. The proposed AI-based system has the potential to be used alongside the clinical guidelines currently employed in the medical setting making them more personalized and dynamic and assisting doctors in taking their everyday clinical decisions.

Development and application of an optimised Bayesian shrinkage prior for spectroscopic biomedical diagnostics

BACKGROUND AND OBJECTIVE

Classification of vibrational spectra is often challenging for biological substances containing similar molecular bonds, interfering with spectral outputs. To address this, various approaches are widely studied. However, whilst providing powerful estimations, these techniques are computationally extensive and frequently overfit the data. Shrinkage priors, which favour models with relatively few predictor variables, are often applied in Bayesian penalisation techniques to avoid overfitting.

METHODS

Using the logit-normal continuous analogue of the spike-and-slab (LN-CASS) as the shrinkage prior and modelling, we have established classification for accurate analysis, with the established system found to be faster than conventional least absolute shrinkage and selection operator, horseshoe or spike-and-slab. These were examined versus coefficient data based on a linear regression model and vibrational spectra produced via density functional theory calculations. Then applied to Raman spectra from saliva to classify the sample sex.

RESULTS

Subsequently applied to the acquired spectra from saliva, the evaluated models exhibited high accuracy (AUC>90 %) even when number of parameters was higher than the number of observations. Analyses of spectra for all Bayesian models yielded high-classification accuracy upon cross-validation. Further, for saliva sensing, LN-CASS was found to be the only classifier with 100 %-accuracy in predicting the output based on a leave-one-out cross validation.

CONCLUSIONS

With potential applications in aiding diagnosis from small spectroscopic datasets and are compatible with a range of spectroscopic data formats. As seen with the classification of IR and Raman spectra. These results are highly promising for emerging developments of spectroscopic platforms for biomedical diagnostic sensing systems.

Unraveling atherosclerotic cardiovascular disease risk factors through conditional probability analysis with Bayesian networks: insights from the AZAR cohort study

This study aimed at modelling the underlying predictor of ASCVD through the Bayesian network (BN). Data for the AZAR Cohort Study, which evaluated 500 healthcare providers in Iran, was collected through examinations, and blood samples. Two BNs were used to explore a suitable causal model for analysing the underlying predictor of ASCVD; Bayesian search through an algorithmic approach and knowledge-based BNs. Results showed significant differences in ASCVD risk factors across background variables' levels. The diagnostic indices showed better performance for the knowledge-based BN (Area under ROC curve (AUC) = 0.78, Accuracy = 76.6, Sensitivity = 62.5, Negative predictive value (NPV) = 96.0, Negative Likelihood Ratio (LR-) = 0.48) compared to Bayesian search (AUC = 0.76, Accuracy = 72.4, Sensitivity = 17.5, NPV = 93.2, LR- = 0.83). In addition, we decided on knowledge-based BN because of the interpretability of the relationships. Based on this BN, being male (conditional probability = 63.7), age over 45 (36.3), overweight (51.5), Mets (23.8), diabetes (8.3), smoking (10.6), hypertension (12.1), high T-C (28.5), high LDL-C (23.9), FBS (12.1), and TG (25.9) levels were associated with higher ASCVD risk. Low and normal HDL-C levels also had higher ASCVD risk (35.3 and 37.4), while high HDL-C levels had lower risk (27.3). In conclusion, BN demonstrated that ASCVD was significantly associated with certain risk factors including being older and overweight male, having a history of Mets, diabetes, hypertension, having high levels of T-C, LDL-C, FBS, and TG, but Low and normal HDL-C and being a smoker. The study may provide valuable insights for developing effective prevention strategies for ASCVD in Iran.