Analysis of Prevalence and Clinical Features of Aortic Stenosis in Patients with and without Bicuspid Aortic Valve Using Machine Learning Methods.

Aortic stenosis (AS) is the most commonly diagnosed valvular heart disease, and its prevalence increases with the aging of the general population. However, AS is often diagnosed at a severe stage, necessitating surgical treatment, due to its long asymptomatic period. The objective of this study was to analyze the frequency of AS in a population of cardiovascular patients using echocardiography (ECHO) and to identify clinical factors and features associated with these patient groups. We utilized machine learning methods to analyze 84,851 echocardiograms performed between 2010 and 2018 at the National Medical Research Center named after V.A. Almazov. The primary indications for ECHO were coronary artery disease (CAD) and hypertension (HP), accounting for 33.5% and 14.2% of the cases, respectively. The frequency of AS was found to be 13.26% among the patients (n = 11,252). Within our study, 1544 patients had a bicuspid aortic valve (BAV), while 83,316 patients had a tricuspid aortic valve (TAV). BAV patients were observed to be younger compared to TAV patients. AS was more prevalent in the BAV group (59%) compared to the TAV group (12%), with a p-value of <0.0001. By employing a machine learning algorithm, we randomly identified significant features present in AS patients, including age, hypertension (HP), aortic regurgitation (AR), ascending aortic dilatation (AscAD), and BAV. These findings could serve as additional indications for earlier observation and more frequent ECHO in specific patient groups for the earlier detection of developing AS.

Machine Learning Methods for Pregnancy and Childbirth Risk Management.

Machine learning methods enable medical systems to automatically generate data-driven decision support models using real-world data inputs, eliminating the need for explicit rule design. In this research, we investigated the application of machine learning methods in healthcare, specifically focusing on pregnancy and childbirth risks. The timely identification of risk factors during early pregnancy, along with risk management, mitigation, prevention, and adherence management, can significantly reduce adverse perinatal outcomes and complications for both mother and child. Given the existing burden on medical professionals, clinical decision support systems (CDSSs) can play a role in risk management. However, these systems require high-quality decision support models based on validated medical data that are also clinically interpretable. To develop models for predicting childbirth risks and due dates, we conducted a retrospective analysis of electronic health records from the perinatal Center of the Almazov Specialized Medical Center in Saint-Petersburg, Russia. The dataset, which was exported from the medical information system, consisted of structured and semi-structured data, encompassing a total of 73,115 lines for 12,989 female patients. Our proposed approach, which includes a detailed analysis of predictive model performance and interpretability, offers numerous opportunities for decision support in perinatal care provision. The high predictive performance achieved by our models ensures precise support for both individual patient care and overall health organization management.

Application of Machine Learning Methods for Epilepsy Risk Ranking in Patients with Hematopoietic Malignancies Using.

Machine learning methods to predict the risk of epilepsy, including vascular epilepsy, in oncohematological patients are currently considered promising. These methods are used in research to predict pharmacoresistant epilepsy and surgical treatment outcomes in order to determine the epileptogenic zone and functional neural systems in patients with epilepsy, as well as to develop new approaches to classification and perform other tasks. This paper presents the results of applying machine learning to analyzing data and developing diagnostic models of epilepsy in oncohematological and cardiovascular patients. This study contributes to solving the problem of often unjustified diagnosis of primary epilepsy in patients with oncohematological or cardiovascular pathology, prescribing antiseizure drugs to patients with single seizure syndromes without finding a disease associated with these cases. We analyzed the hospital database of the V.A. Almazov Scientific Research Center of the Ministry of Health of Russia. The study included 66,723 treatment episodes of patients with vascular diseases (I10-I15, I61-I69, I20-I25) and 16,383 episodes with malignant neoplasms of lymphoid, hematopoietic, and related tissues (C81-C96 according to ICD-10) for the period from 2010 to 2020. Data analysis and model calculations indicate that the best result was shown by gradient boosting with mean accuracy cross-validation score = 0.96. f1-score = 98, weighted avg precision = 93, recall = 96, f1-score = 94. The highest correlation coefficient for G40 and different clinical conditions was achieved with fibrillation, hypertension, stenosis or occlusion of the precerebral arteries (0.16), cerebral sinus thrombosis (0.089), arterial hypertension (0.17), age (0.03), non-traumatic intracranial hemorrhage (0.07), atrial fibrillation (0.05), delta absolute neutrophil count (0.05), platelet count at discharge (0.04), transfusion volume for stem cell transplantation (0.023). From the clinical point of view, the identified differences in the importance of predictors in a broader patient model are consistent with a practical algorithm for organic brain damage. Atrial fibrillation is one of the leading factors in the development of both ischemic and hemorrhagic strokes. At the same time, brain infarction can be accompanied both by the development of epileptic seizures in the acute period and by unprovoked epileptic seizures and development of epilepsy in the early recovery and in a longer period. In addition, a microembolism of the left heart chambers can lead to multiple microfocal lesions of the brain, which is one of the pathogenetic aspects of epilepsy in elderly patients. The presence of precordial fibrillation requires anticoagulant therapy, the use of which increases the risk of both spontaneous and traumatic intracranial hemorrhage.

Application of Machine Learning Methods to Analyze Occurrence and Clinical Features of Ascending Aortic Dilatation in Patients with and without Bicuspid Aortic Valve.

Aortic aneurysm (AA) rapture is one of the leading causes of death worldwide. Unfortunately, the diagnosis of AA is often verified after the onset of complications, in most cases after aortic rupture. The aim of this study was to evaluate the frequency of ascending aortic aneurysm (AscAA) and aortic dilatation (AD) in patients with cardiovascular diseases undergoing echocardiography, and to identify the main risk factors depending on the morphology of the aortic valve. We processed 84,851 echocardiographic (ECHO) records of 13,050 patients with aortic dilatation (AD) in the Almazov National Medical Research Centre from 2010 to 2018, using machine learning methodologies. Despite a high prevalence of AD, the main reason for the performed ECHO was coronary artery disease (CAD) and hypertension (HP) in 33.5% and 14.2% of the patient groups, respectively. The prevalence of ascending AD (>40 mm) was 15.4% (13,050 patients; 78.3% (10,212 patients) in men and 21.7% (2838 patients) in women). Only 1.6% (n = 212) of the 13,050 patients with AD knew about AD before undergoing ECHO in our center. Among all the patients who underwent ECHO, we identified 1544 (1.8%) with bicuspid aortic valve (BAV) and 635 with BAV had AD (only 4.8% of all AD patients). According to the results of the random forest feature importance analysis, we identified the eight main factors of AD: age, male sex, vmax aortic valve (AV), aortic stenosis (AS), blood pressure, aortic regurgitation (AR), diabetes mellitus, and heart failure (HF). The known factors of AD-like HP, CAD, hyperlipidemia, BAV, and obesity, were also AD risk factors, but were not as important. Our study showed a high frequency of AscAA and dilation. Standard risk factors of AscAA such as HP, hyperlipidemia, or obesity are significantly more common in patients with AD, but the main factors in the formation of AD are age, male sex, vmax AV, blood pressure, AS, AR, HF, and diabetes mellitus. In males with BAV, AD incidence did not differ significantly, but the presence of congenital heart disease was one of the 12 main risk factors for the formation of AD and association with more significant aortic dilatation in AscAA groups.

Influence of Healthcare Organization Factors on Cardiovascular Diseases Mortality.

One serious pandemic can nullify years of efforts to extend life expectancy and reduce disability. The coronavirus pandemic has been a perturbing factor that has provided an opportunity to assess not only the effectiveness of health systems for cardio-vascular diseases (CVD), but also their sustainability. The goal of our research is to analyze the influence of public health factors on the mortality from circulatory diseases using machine learning methods. We analysed a very large dataset that consisted of the information collected from the national registers in Russia. We included data from 2015 to 2021. It included 340 factors that characterize organization of healthcare in Russia. The resulting area under receiver operating characteristic curve (AUC of ROC) of the Random Forest based regression model was 92% with a testing dataset. The models allow for automated retraining as time passes and epidemiological and other situations change. They also allow additional characteristics of regions and health care organizations to be added to existing training datasets depending on the target. The developed models allow the calculation of the probability of the target for 6-12 months with an error of 8%. Moreover, the models allow to calculate scenarios and the value of the target indicator when other indicators of the region change.

Dynamic Aortic Aneurism Risk Factors.

According to different systematic reviews incidence of thoracic aortic aneurysms (TAA) in the general population is increasing in frequency ranging from 5 to 10.4 per 100000 patients. However, only few studies have illustrated the role of different risk factors in the onset and progression of ascending aortic dilatation. Currently, noninvasive imaging techniques are used to assess the progression rate of aortic and aortic valve disease. Transthoracic (TT) Echocardiographic examination routinely includes evaluation of the aorta It is the most available screening method for diagnosis of proximal aortic dilatation. Since the predominant area of dilation is the proximal aorta, TT-echo is often sufficient for screening. We retrospectively analyzed the ECHO database with 78499 echocardiographic records in the Almazov National Medical Research Centre to identify patients with aneurysm. Detailed information including demographic characteristics, ECHO results and comorbidities were extracted from outpatient clinic and from hospital charts related to hospitalizations occurring within a year before index echocardiography was performed. Comorbid diseases were similarly extracted from outpatient clinic and/or hospital admissions. The classifier showed an AUC-ROC for predicting of aneurism detection after a repeated ECHO at 82%.

Lifestyle Cancer Survival Predictors: Influence of Vegetarian Diet on the Relapse of Endometrial Cancer.

Endometrial cancer (EC) is the most common gynecological tumor in high-income countries, and its incidence has increased over time. The most critical risk factor for EC is the long-term unopposed exposure to increased estrogens both exogenous and endogenous. Machine learning can be used as a promising tool to resolve longstanding challenges and support identification of the risk factors and their correlations before the clinical trials and make them more focused. In this paper we present the results of the research of the correlation analysis of Endometrial cancer risk factors. The study was performed with EC patients of the Almazov center in Saint-Petersburg, Russia. All women involved in the current study underwent radical surgical intervention due to EC. After initial cancer treatment, they were referred to the Almazov center outpatient specialists for follow-up visits. Many of them were readmitted of the inpatient clinic due to relapse. We extracted a variety of parameters related to lifestyle, dietary habits, socioeconomic, and reproductive features from the inpatient and outpatient databases of Almazov center. The medical records of the women with enough data were included in the study. Prediction of Progression-free survival (PFS) and overall survival (OS) were analyzed respectively. The AUC of ROC was calculated for PFS = 0.93 and for OS = 0.94.

Regulation Modelling and Analysis Using Machine Learning During the Covid-19 Pandemic in Russia.

Due to the specific circumstances related to the COVID-19 pandemic, many countries have enforced emergency measures such as self-isolation and restriction of movement and assembly, which are also directly affecting the functioning of their respective public health and judicial systems. The goal of this study is to identify the efficiency of the criminal sanctions in Russia that were introduced in the beginning of COVID-19 outbreak using machine learning methods. We have developed a regression model for the fine handed out, using random forest regression and XGBoost regression, and calculated the features importance parameters. We have developed classification models for the remission of the penalty and for setting a sentence using a gradient boosting classifier.

Gynecological Surgery and Machine Learning: Complications and Length of Stay Prediction.

In this study we are developing predictive models for a length of stay after a gynecological surgery, complications and the length of the surgery using machine learning methods. The study was performed with the data of patients with the diseases of the female reproductive system. The patients were admitted to the Almazov National Medical Research Centre (Saint-Petersburg, Russia) within the period 2010-2020. The study included 8170 electronic medical records of inpatient episodes including 3500 operation protocols. The data included anamnesis of life, anamnesis of disease, laboratory tests, severity, outcome of a surgery, main and comorbid diagnosis, complications, case outcome. The dataset was randomly split into 70% train and 30% test datasets. Validation with the test dataset provided the following prediction metrics for the length of stay after a surgery model. Training score: AUC of ROC: 0.9582230976834093; K-fold CV average score: -8.73; MSE: 5.65; RMSE: 2.83.

Prediction of a Due Date Based on the Pregnancy History Data Using Machine Learning.

Prediction of a labor due date is important especially for the pregnancies with high risk of complications where a special treatment is needed. This is especially valid in the countries with multilevel health care institutions like Russia. In Russia medical organizations are distributed into national, regional and municipal levels. Organizations of each level can provide treatment of different types and quality. For example, pregnancies with low risk of complications are routed to the municipal hospitals, moderate risk pregnancies are routed to the reginal and high risk of complications are routed to the hospitals of the national level. In the situation of resource deficiency especially on the national level it is necessary to plan admission date and a treatment team in advance to provide the best possible care. When pregnancy data is not standardized and semantically interoperable, data driven models. We have retrospectively analyzed electronic health records from the perinatal Center of the Almazov perinatal medical center in Saint-Petersburg, Russia. The dataset was exported from the medical information system. It consisted of structured and semi structured data with the total of 73115 lines for 12989 female patients. The proposed due date prediction data-driven model allows a high accuracy prediction to allow proper resource planning. The models are based on the real-world evidence and can be applied with limited amount of predictors.

Prediction of Childbirth Mortality Using Machine Learning.

Timely identification of risk factors in the early stages of pregnancy, risk management and mitigation, prevention, adherence management can reduce the number of adverse perinatal outcomes and complications for both mother and a child. We have retrospectively analyzed electronic health records from the perinatal Center of the Almazov specialized medical center in Saint-Petersburg, Russia. Correlation analysis was performed using Pearson correlation coefficient to select the most relevant predictors. We used APGAR score as a metrics for the childbirth outcomes. Score of 5 and less was considered as a negative outcome. To analyze the influence of the unstructured anamnesis data on the prediction accuracy we have run two prediction experiments for every classification task: 1. Without unstructured data and 2. With unstructured data. This study presents implementation of predictive models for adverse childbirth events that provides higher precision than state of the art models. This is due to the use of unstructured medical data in addition to the structured dataset that allowed to reach 0.92 precision. Identification of main risk factors using the results of the features importance analysis can support clinicians in early identification of possible complications and planning and execution preventive measures.

Identification of Diabetes Risk Factors in Chronic Cardiovascular Patients.

Specific predictive models for diabetes polyneuropathy based on screening methods, for example Nerve conduction studies (NCS, can reach up to AUC 65.8 - 84.7 % for the conditional diagnosis of DPN in primary care. Prediction methods that utilize data from personal health records deal with large non-specific datasets with different prediction methods. Li et al. utilized 30 independent variables, which allowed to implement a model with AUC = 0.8863 for a Multilayer perceptron (MLP). Linear regression (LR) based methods produced up to AUC = 0.8 %. This way, modern data mining and computational methods can be effectively adopted in clinical medicine to derive models that use patient-specific information to predict the development of diabetic polyneuropathy, however, there still is a space to improve the efficiency of the predictive models. The goal of this study is the implementation of machine learning methods for early risk identification of diabetes polyneuropathy based on structured electronic medical records. It was demonstrated that the machine learning methods allow to achieve up to 0.7982 precision, 0.8152 recall, 0.8064 f1-score, 0.8261 accuracy, and 0.8988 AUC using the neural network classifier.

Survival Analysis of COVID-19 Patients in Russia Using Machine Learning.

The current pandemic can likely have several waves and will require a major effort to save lives and provide optimal treatment. The efficient clinical resource planning and efficient treatment require identification of risk groups and specific clinical features of the patients. In this study we develop analyze mortality for COVID19 patients in Russia. We identify comorbidities and risk factors for different groups of patients including cardiovascular diseases and therapy. In the study we used a Russian national COVID registry, that provides sophisticated information about all the COVID-19 patients in Russia. To analyze Features importance for the mortality we have calculated Shapley values for the "mortality" class and ANN hidden layer coefficients for patient lifetime. We calculated the distribution of days spent in hospital before death to show how many days a patient occupies a bed depending on the age and the severity of the disease to allow optimal resource planning and enable age-based risk assessment. Predictors of the days spent in hospital were calculated using Pearson correlation coefficient. Decisions trees were developed to classify the patients into the groups and reveal the lethality factors.

Modelling of COVID-19 Morbidity in Russia.

The outbreak of COVID-19 has led to a crucial change in ordinary healthcare approaches. In comparison with emergencies re-allocation of resources for a long period of time is required and the peak utilization of the resources is also hard to predict. Furthermore, the epidemic models do not provide reliable information of the development of the pandemic's development, so it creates a high load on the healthcare systems with unforeseen duration. To predict morbidity of the novel COVID-19, we used records covering the time period from 01-03-2020 to 25-05-2020 and include sophisticated information of the morbidity in Russia. Total of 45238 patients were analyzed. The predictive model was developed as a combination of Holt and Holt-Winter models with Gradient boosting Regression. As we can see from the table 2, the models demonstrated a very good performance on the test data set. The forecast is quite reliable, however, due to the many uncertainties, only a real-world data can prove the correctness of the forecast.

Identification of risk factors for patients with diabetes: diabetic polyneuropathy case study.

BACKGROUND: Methods of data mining and analytics can be efficiently applied in medicine to develop models that use patient-specific data to predict the development of diabetic polyneuropathy. However, there is room for improvement in the accuracy of predictive models. Existing studies of diabetes polyneuropathy considered a limited number of predictors in one study to enable a comparison of efficiency of different machine learning methods with different predictors to find the most efficient one. The purpose of this study is the implementation of machine learning methods for identifying the risk of diabetes polyneuropathy based on structured electronic medical records collected in databases of medical information systems. METHODS: For the purposes of our study, we developed a structured procedure for predictive modelling, which includes data extraction and preprocessing, model adjustment and performance assessment, selection of the best models and interpretation of results. The dataset contained a total number of 238,590 laboratory records. Each record 27 laboratory tests, age, gender and presence of retinopathy or nephropathy). The records included information about 5846 patients with diabetes. Diagnosis served as a source of information about the target class values for classification. RESULTS: It was discovered that inclusion of two expressions, namely "nephropathy" and "retinopathy" allows to increase the performance, achieving up to 79.82% precision, 81.52% recall, 80.64% F1 score, 82.61% accuracy, and 89.88% AUC using the neural network classifier. Additionally, different models showed different results in terms of interpretation significance: random forest confirmed that the most important risk factor for polyneuropathy is the increased neutrophil level, meaning the presence of inflammation in the body. Linear models showed linear dependencies of the presence of polyneuropathy on blood glucose levels, which is confirmed by the clinical interpretation of the importance of blood glucose control. CONCLUSION: Depending on whether one needs to identify pathophysiological mechanisms for one's prospective study or identify early or late predictors, the choice of model will vary. In comparison with the previous studies, our research makes a comprehensive comparison of different decisions using a large and well-structured dataset applied to different decision support tasks.

Echocardiography Population Study in Russian Federation for 4P Medicine Using Machine Learning.

This article describes the study results of echocardiographic (ECHO) test data for 4P medicine applied to cardiovascular patients. Data from more than 145,000 echocardiographic tests were analyzed. One of the objectives of the study is the possibility to identify patterns and relationships in patient characteristics for more accurate appointment procedures based on the history of the disease and the individual characteristics of the patient. This is achieved by using classifications models based on machine learning methods. Early detection of disease risks and "accurate" appointment of diagnostic procedures makes a significant contribution to value-based medicine. Moreover, it was also possible to identify the classes and characteristics of patients for whom repeated diagnostic procedures are well founded. Calculation of personal risks from empirical retrospective data helps to detect the disease in early stages. Identifying patients with high risk of disease complications allow physicians to make right decisions about timely treatment, which can significantly improve the quality of treatment, and help to avoid diseases complications, optimize costs and improve the quality of medical care.

Graph-Based Predictive Modelling of Chronic Disease Development: Type 2 DM Case Study.

This study proposes a graph-based method for representing the dynamics of chronic diabetes as a complex process with different characteristics. The study was based on the case histories of 6864 patients with diabetes mellitus, 90% of whom suffer from type 2 diabetes. Our method allows to predict the sequence of events during the development of type 2 diabetes for each patient. Typical developmental trajectories of the disease were investigated, their clustering was carried out, the trajectory patterns were identified and studied. Based on the constructed directed graph reflecting transitions between different conditions of the patients, the clustering of diabetic statuses was carried out using the Modularity Class method; 8 clusters were selected, each of them was interpreted and studied. The method of the disease developmental trajectories creation by means of machine learning methods was described. Unlike static models of a disease course, this method considers complete past information on the patient and his or her previous events, using each event of the course of disease to predict the next event.

Dynamic Features Impact on the Quality of Chronic Heart Failure Predictive Modelling.

We study the way dynamics affects modelling in chronic heart failure (CHF) tasks. By dynamics we understand the patient history and the appearance of new events, states and variables changing in time. The goal is to understand what impact past data has on prediction quality. Three different experiments have been conducted: CHF episode results prediction (better, worse, no change), CHF stage classification and heart rate value prediction. For modelling we use clinical data of CHF patients. For each task the groups of static and dynamic features are selected and analyzed. For each task 3 machine learning algorithms were trained: XGBoost, Logistic Regression, and Random Forest for multi classification and Linear Regression, Decision Tree and XGBoost for the regression task. Different combinations of features were examined from both groups applying forward feature selection algorithm. The results confirm that the highest predictions quality is reached with combination of static and dynamic features.

Simulation of patient flow in multiple healthcare units using process and data mining techniques for model identification.

INTRODUCTION: An approach to building a hybrid simulation of patient flow is introduced with a combination of data-driven methods for automation of model identification. The approach is described with a conceptual framework and basic methods for combination of different techniques. The implementation of the proposed approach for simulation of the acute coronary syndrome (ACS) was developed and used in an experimental study. METHODS: A combination of data, text, process mining techniques, and machine learning approaches for the analysis of electronic health records (EHRs) with discrete-event simulation (DES) and queueing theory for the simulation of patient flow was proposed. The performed analysis of EHRs for ACS patients enabled identification of several classes of clinical pathways (CPs) which were used to implement a more realistic simulation of the patient flow. The developed solution was implemented using Python libraries (SimPy, SciPy, and others). RESULTS: The proposed approach enables more a realistic and detailed simulation of the patient flow within a group of related departments. An experimental study shows an improved simulation of patient length of stay for ACS patient flow obtained from EHRs in Almazov National Medical Research Centre in Saint Petersburg, Russia. CONCLUSION: The proposed approach, methods, and solutions provide a conceptual, methodological, and programming framework for the implementation of a simulation of complex and diverse scenarios within a flow of patients for different purposes: decision making, training, management optimization, and others.