Medical Decision Support System Research Articles

Development and testing of a medical decision support system in endoscopy

.

Early Detection of Cardiac Diseases from Electrocardiogram Using Artificial Intelligence Techniques

Heart disease is one of the leading causes of death. Early diagnosis improves the treatment quality and reduces the mortality rates. Using Electrocardiogram (ECG) signals in health diagnosis has attracted researchers to develop a qualified intelligent medical decision support system. In this research, a smart early detection system for three common heart diseases, namely Apnea, Atrial Fibrillation (AF), and Heart Failure (HF) is developed, by analyzing the ECG signal for the patients using artificial intelligence (AI) techniques. The designed system is implemented using three different artificial intelligence techniques, namely Artificial Neural Networks (ANN), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN). Electrocardiogram (ECG) signals from PhysioNet and data collected by the researchers are analyzed, and four features are extracted; then, they are used as an input for the classification algorithms. The results show the many advantages of using the proposed artificial intelligence techniques, which can result in saving the lives of patients with heart diseases. The accuracy of classification using K-nearest Neighbor is 92.4% at k=1. Using the Artificial Neural Networks provides an accuracy of 95.7% at 33 epochs, while the highest classification accuracy is obtained using Support Vector Machine with a classification accuracy of 97.8%.

An intelligent medical decision support system for diagnosis of heart abnormalities in ECG signals

This work describes a novel method to detect a Bundle branch block and myocardial infarction from the multi-lead ECG signal. The clinical characteristics of BBB and MI extracted by using a derivative filter and continuous wavelet transform (CWT). The signal with the frequency below 50 Hz obtained and derivative-based filter applied to extract features. The continuous wavelet transforms also applied to the signals of BBB and MI. The CWT coefficients extracted, and the signals reconstructed from the wavelet to obtain the features. The feature vectors generated from each lead of both the methods computed using parameters such as spectral entropy, mean of peaks, total energy from power spectrum density, form factor, and root mean squared value. The results of both the derivative-based filter and CWT analyzed by applying these features to the classifiers. The accuracy of classification of diseases computed using SVM, KNN, Levenberg-Marquardt Neural Network (LMNN), and scaled conjugate gradient backpropagation network (SCG NN). The best accuracy obtained from the derivative filter and wavelet transform method is 96.4% using LMNN and SCGNN classifier and 96.4% using KNN and LMNN classifier respectively.

An interpretable knowledge-based decision support system and its applications in pregnancy diagnosis

This paper aims to propose an interpretable knowledge-based decision support system (IKBDSS) that will assist physicians to predict the risk level of a disease. Our system enables to integrate both historical cases extracted from database and opinions provided by different experts in order to set up a medical knowledge base and provide relevant advises by inferring from the knowledge base. To present various experts’ opinions, the Multi-granularity Linguistic Term Sets (MLTS) model is used to address the ambiguity and intangibility of knowledge. Our work mainly focuses on knowledge acquisition, similarity degree calculation and consistency checking process. It is worth mentioning that a criterion weights calculation method is introduced to objectively obtain the weights based on knowledge from experts, rather than subjectively predefined.The developed system leads to a better performance in specificity, sensitivity and F1 score compared to other methods in the literature. To conclude, our work contributes to: (1) The development of a medical decision support system to combine clinical records and domain knowledge to predict diagnosis. (2) The decision-making process ensures interpretability, which increases the reliability of our system in terms of being a decision supporter. (3) The criterion weights are calculated based on the professional knowledge presented in MLTS form, and this process improves the capacity of providing diagnostic recommendations.

A medical decision support system for predicting the severity level of COVID-19

The main assay tool of COVID-19, as a pandemic, still has significant faults. To ameliorate the current situation, all facilities and tools in this realm should be implemented to encounter this epidemic. The current study has endeavored to propose a self-assessment decision support system (DSS) for distinguishing the severity of the COVID-19 between confirmed cases to optimize the patient care process. For this purpose, a DSS has been developed by the combination of the data-driven Bayesian network (BN) and the Fuzzy Cognitive Map (FCM). First, all of the data are utilized to extract the evidence-based paired (EBP) relationships between symptoms and symptoms’ impact probability. Then, the results are evaluated in both independent and combined scenarios. After categorizing data in the triple severity levels by self-organizing map, the EBP relationships between symptoms are extracted by BN, and their significance is achieved and ranked by FCM. The results show that the most common symptoms necessarily do not have the key role in distinguishing the severity of the COVID-19, and extracting the EBP relationships could have better insight into the severity of the disease.

FORMALIZING THE STAGES OF MAMMOGRAPHIC EXAMINATIONS IN DESIGNINGA MEDICAL DECISION SUPPORT SYSTEM

The paper considers the formalization of the stages and modeling of the mammographic examination procedure in the design of medical computer decision support systems. The mammographic examination process is presented in a generalizedmodel,which consists offunctional, structural, and mathematical models. The functional model (context diagram) is made using the functional modeling methodology.When analyzingthe context diagram, four main functional blocks were identified: register a patient; perform registration and analysis of mammograms; carry out diagnostics; form a survey protocol. If there are standards for maintaining medi-cal records and drawing up examination protocols, the first and last blocksare easily automated.The article focuseson the second and third blocks.At the mammogram analysis stage, the sub-stages “Perform preliminary processing”and “Perform morphological analysis”are essential. Preprocessing of mammograms (adaptive filtering, changing brightness or increasing contrast, etc.) is carried out using digital image processing methods to improve visualization quality. The result of morphological analysis is selecting struc-tural elements and forminga set of diagnostic signs in the form of parameters of the found structural elements.Becausesome ele-ments of mammograms (microcalcifications) have an irregular structure, specialized morphological analysis methods are used, based on taking into account the features of the images under consideration and their transformation methods in the form of the useful sig-nal, in particular, fractal dimension models.The developed formalized models made it possible to reasonably design the decision support system’s structureduring mammographic examinations, information, mathematical,software, and hardwareto increase med-ical services’efficiencyand minimize the risks of medical errors.

Сравнительный анализ мобильного приложения для измерения параметров сагиттального баланса «СпиноМетр» с системой Surgimap: апробация межэкспертной надежности

The purpose of the study was to assess the reliability, accuracy and effectiveness of a mobile application SpinoMeter for measuring the sagittal balance of the spino-pelvic complex in comparison with the desktop preoperative planning system Surgimap. Methods SpinoMeter mobile application was developed in 2019 at the laboratory for Medical Decision Support Systems of Saratov State University. SpinoMeter is intended for measuring the actual and for calculating the optimal parameters of the sagittal balance of the spino-pelvic complex of a particular patient. Pelvic incidence (PI), lumbar lordosis (LL), sacral slope (SS) and pelvic tilt (PT) were measured with the help of Surgimap and SpinoMeter. Inter-observer reliability was assessed on X-rays in a standing position of 52 patients having consequences of injuries and pathology of the lumbar spine with imbalance syndrome. The research site was the Scientific Research Institute of Traumatology, Orthopedics and Neurosurgery of Saratov State Medical University. Measurements of the sagittal balance parameters were presented in tables and diagrams. For all data, the hypothesis of a normal distribution of the measurement results was tested using the Kolmogorov-Smirnov test. To test the hypothesis of homogeneity of the measurement results for normally distributed samples, the paired Student t-test was used. In the absence of a normal distribution, the Wilcoxon test was used for related samples. Results For PI, PT, SS and L1-S1 lordosis at a significance level of 5 %, the hypothesis of uniformity of measurements was confirmed. The average value of L4-S1 lordosis, measured by SpinoMeter, was shifted down relative to the measurements of this parameter by the Surgimap program. The shift did not exceed 0.7 %. Conclusion SpinoMeter mobile application is equivalent to the Surgimap system in terms of measuring the sagittal balance parameters based on radiographs in a standing position.

WITHDRAWN: A comparative study on machine learning based heart disease prediction

Over the last few decades, the population worldwide is suffering from heart disease, which is considered one of the most significant fatalities. About 17.7 million people die on average every year because of heart disease, the World Health Organization (WHO). There will be many difficulties in the prognosis of heart disease due to various risk factors like diabetes, high blood pressure, high cholesterol, abnormal pulse rate, and many other factors. The main goal is to save humans' lives by detecting abnormalities in heart conditions, which would be achieved by identifying and processing raw data collected based on heart information. The healthcare industry has found that Machine Learning (ML) is a useful and accurate decision-making technique in the data collection produced in large quantities. The medical decision support systems developed were effective based on the software and the different algorithms proposed by many researchers. Here a study is done based on the various techniques using the different algorithms and their performance analysis. The predicting model was introduced with several combined features, and among the multiple methods and were other classification techniques. Many existing ways discussed, among which the accuracy level was found as 88.7% using the Hybrid Random Forest with a Linear Model (HRFLM) technique.

Pharmacotherapy of allergic rhinitis: position of a clinical pharmacologist

Pharmacotherapy for allergic rhinitis is based on different categories of drugs used either in monotherapy or in combination regimens. The current clinical guidelines suggest a stepwise approach to pharmacotherapy for allergic rhinitis. The use of intranasal corticosteroids is considered as the preferred second-stage pharmacotherapy. Inadequate control of AR symptoms in first-line therapy is a common problem. Integrated care pathways (ICP), developed taking into account the data obtained about patients using a mobile application, suggest the use of intranasal corticosteroids as the first line of therapy, including in patients with intermittent rhinitis who have not previously received treatment when assessing the condition according to the VAS for more than 5 points, in patients who received earlier treatment when assessing the condition according to the VAS less than 5 points. According to the data in the medical decision support system and continuing medical education UpToDate, inhaled corticosteroids are considered as the first-line drugs for the pharmacotherapy of allergic rhinitis. In terms of pharmacodynamic efficacy, intranasal corticosteroids are comparable to each other. The selection criteria can be considered: the value of systemic absorption; lipophilicity; the start time of the action; frequency of introduction, organoleptic properties; the possibility of influencing non-nasal symptoms. The use of sprays containing both a glucocorticoid and an antihistamine (mometasone furoate/azelastine hydrochloride) opens up additional pharmacotherapeutic possibilities in the treatment of allergic rhinitis.

Data-Driven Online Recommender Systems with Costly Information Acquisition

In numerous recommender systems, collecting useful information from users is costly, implying that the recommender system has to make active choices by simultaneously learning the observations of the features' states to make useful recommendations to users among available products and services. This paper integrates information acquisition decisions into recommender system. To solve the aforementioned dual learning problem, we propose two different algorithms, namely Sim-OOS and Seq-OOS, where observations are made simultaneously and sequentially, respectively. We prove that both algorithms guarantee a sub linear regret. The developed recommender system can be applied to a variety of real-world applications, including medical informatics, smart transportation, finance, and cyber-security where collecting information before making decisions results in an excessive cost. We validate and evaluate our proposed policies in a medical decision-support system that recommends tests and treatments for breast cancer patients.

Virtual Care Technologies, Wearable Health Monitoring Sensors, and Internet of Medical Things-based Smart Disease Surveillance Systems in the Diagnosis and Treatment of COVID-19 Patients

Digital epidemiological surveillance in monitoring, detection, and prevention of COVID-19 is optimized by use of medical artificial intelligence, clinical and diagnostic decision support systems, machine learning-based real-time data sensing and processing, and smart healthcare devices and applications. Study Design, Survey Methods, and Materials The interviews were conducted online and data were weighted by five variables (age, race/ethnicity, gender, education, and geographic region) using the Census Bureau's American Community Survey to reflect reliably and accurately the demographic composition of the United States. (Pustokhina et al., 2020) Body sensor networks integrate interconnected bio-sensors and wearable healthcare devices (Kovacova and Lazaroiu, 2021;Lyons and Lazaroiu, 2020) that assess abnormal alterations in vital physiological signs and share medical imaging data for patient diagnosis and monitoring, being instrumental in chronic diseases by use of deep learning-based applications. Conclusions, Implications, Limitations, and Further Research Directions Artificial intelligence-enabled wearable medical devices, virtualized care systems, and wireless biomedical sensing devices are pivotal in COVID-19 screening, testing, and treatment.

Artificial Intelligence-enabled Wearable Medical Devices, Clinical and Diagnostic Decision Support Systems, and Internet of Things-based Healthcare Applications in COVID-19 Prevention, Screening, and Treatment

Building our argument by drawing on data collected from Accenture, GlobalWebIndex, GoMo Health, KPMG, McKinsey, Oracle, Sermo, STAT, Statista, and Workplace Intelligence, we performed analyses and made estimates regarding how predictive big data analytics, body sensor networks, medical wearable devices, decision support systems, and wireless sensing applications can be harnessed in real-time continuous remote monitoring of patients vital signs configuring clinical data in pervasive mobile patient-centric healthcare. Introduction The extensive data of COVID-19 patients can be assimilated and inspected by cutting-edge machine learning algorithms to grasp the pattern of viral transmission, optimize diagnostic swiftness and precision, advance adequate therapeutic methods, and identify the most vulnerable individuals according to personalized genetic and physiological features. Methodology and Empirical Analysis Building our argument by drawing on data collected from Accenture, Global-WebIndex, GoMo Health, KPMG, McKinsey, Oracle, Sermo, STAT, Statista, and Workplace Intelligence, we performed analyses and made estimates regarding how predictive big data analytics, body sensor networks, medical wearable devices, decision support systems, and wireless sensing applications can be harnessed in real-time continuous remote monitoring of patients' vital signs configuring clinical data in pervasive mobile patient-centric healthcare. Study Design, Survey Methods, and Materials The interviews were conducted online and data were weighted by five variables (age, race/ethnicity, gender, education, and geographic region) using the Census Bureau's American Community Survey to reflect reliably and accurately the demographic composition of the United States.

FPGA implementation of fuzzy medical decision support system for disc hernia diagnosis

The aim of this study was to create a decision support system for disc hernia diagnostics based on real measurements of foot force values from sensors and fuzzy logic, as well as to implement the system on Field Programmable Gate Array (FPGA). The results show that the created fuzzy logic system had the 92.8% accuracy for pre-operational diagnosis and very high match between the Matlab and FPGA output (94.2% match for pre-operational condition, and 100% match for the post-operational and after physical therapy conditions). Interestingly enough, our system is also able to detect improvements in patient condition after the surgery and physical therapy. The main benefit of using FPGAs in this study is to create an inexpensive, portable expert system for real time acquisition, processing and providing the objective recommendation for disc hernia diagnosis and tracking the condition improvement.

Explainable Artificial Intelligence Based Framework for Non-Communicable Diseases Prediction

The rapid rise of non-communicable diseases (NCDs) becomes one of the serious health issues and the leading cause of death worldwide. In recent years, artificial intelligence-based systems have been developed to assist clinicians in decision-making to reduce morbidity and mortality. However, a common drawback of these modern studies is related to explanations of their output. In other words, understanding the inner logic behind the predictions is hidden to the end-user. Thus, clinicians struggle to interpret these models because of their black-box nature, and hence they are not acceptable in the medical practice. To address this problem, we have proposed a Deep Shapley Additive Explanations (DeepSHAP) based deep neural network framework equipped with a feature selection technique for NCDs prediction and explanation among the population in the United States. Our proposed framework comprises three components: First, representative features are done based on the elastic net-based embedded feature selection technique; second a deep neural network classifier is tuned with the hyper-parameters and used to train the model with the selected feature subset; third, two kinds of model explanation are provided by the DeepSHAP approach. Herein, (I) explaining the risk factors that affected the model’s prediction from the population-based perspective; (II) aiming to explain a single instance from the human-centered perspective. The experimental results indicated that the proposed model outperforms various state-of-the-art models. In addition, the proposed model can improve the medical understanding of NCDs diagnosis by providing general insights into the changes in disease risk at the global and local levels. Consequently, DeepSHAP based explainable deep learning framework contributes not only to the medical decision support systems but also can provide to real-world needs in other domains.

Design and implementation of the system to support making medical decisions aid to the population

The publication discusses the issues and results of designing an automated system for supporting medical decision-making on issues of antirabic prevention and aid to the population. The need to provide medical care in the framework of preventing deaths from rabies infection and carrying out preventive measures to prevent such infections is relevant not only in the Russian Federation, but also in many countries of the world. Algorithms for providing medical care and carrying out rabies vaccination are complex, requiring certain professional skills from medical professionals, as well as care and accuracy from patients. Therefore, in manual mode, they are time-consuming both in terms of registering information about the treatment performed, and in terms of the speed of forming medical decisions when providing medical care or when correcting violations of previously defined algorithms. For computer support of these processes, an automated system for supporting medical decision-making has been developed using the tools of the Clarion 10 relational database management system. For various categories of patients and citizens undergoing rabies prevention, experts have developed and entered into the database templates of treatment courses, which are assigned by the system after analyzing the clinical data of patients and the initial or repeated fact of their request for help. These templates and algorithms for analyzing the possibilities of their application are similar to products that allow you to link assumptions and conclusions when making decisions. The automated medical decision support system allows you to register patients and persons undergoing rabies prevention, assign them treatment courses and vaccination schemes, adjust treatment and prevention methods in case of violations of previously prescribed ones, and generate statistical reports. The developed software product received the author’s certificate no. 2018663452 dated 26.10.2018 from the Federal service for intellectual property (Rospatent). The software product has been tested in first aid rooms in Moscow, in the practice of the polyclinic department of the First City hospital named after N. I. Pirogov.

Defining Optimal Exercises for Efficient Detection of Parkinson’s Disease Using Machine Learning and Wearable Sensors

Our society exhibits a worldwide trait of a quickly growing cohort of patients with neurodegenerative diseases, such as Parkinson's disease (PD). According to the analysts, there is a plausible “PD pandemic” to occur within the next two decades. Nowadays, the research in the area focuses on how to detect, predict, or classify PD and similar diseases without addressing the point of what activities or exercises a subject should do to improve the performance of these tasks. In this article, we propose a method based on machine learning (ML) and wearable sensors to identify the optimal exercises for the efficient detection of PD in patients. We first define 15 common tasks that are typically used to diagnose PD in modern clinical practice. However, these exercises still carry a high risk of misdiagnosis and, moreover, not all of them work well in the scope of existing ML solutions to support the diagnosis. Herein, we collect the data in a real clinical setting using a compact wearable wireless sensor node entailing a board gyroscope, accelerometer, and magnetometer. Application of ML methods to the collected data reveals three “most efficient” exercises to assist diagnosticians with the highest discriminating power (0.9 ROC AUC in each task). The proposed solution can be implemented as a medical decision support system for real-time PD diagnostics.

Evolutionary computing for clinical dataset classification using a novel feature selection algorithm

The medical diagnostic decision support system uses machine learning and data mining algorithms to detect and diagnose diseases. Several deaths can be avoided if the diseases are detected and cured in the early stages of infection. Feature selection is a major pre-processing method used to obtain the most significant features, thereby enhancing the data mining model's classification accuracy. This work proposes a new feature selection algorithm to perform feature selection as a multi-objective optimization problem. The minimization of classification error rate and minimization of the feature subset's cardinality are the two contradictory objectives that need to be optimized simultaneously. The proposed work is applied for five clinical datasets such as lung cancer, breast cancer, diabetes, fertility, and immunotherapy and the results are compared with existing techniques based on 6 other datasets. This work converts the real-valued Jaya Optimization Algorithm into binary space. It also handles premature convergence and sensitivity–specificity trade-off. The proposed algorithm's efficiency is assessed and analyzed based on average classification accuracy, sensitivity, specificity, number of features selected, percentage feature selection, and CPU computation time. The proposed algorithm improves the effectiveness of data mining based medical diagnostic decision support system.

An Intelligent IoT Based Healthcare System Using Fuzzy Neural Networks

Healthcare facilities in modern age are key challenge especially in developing countries where remote areas face lack of high-quality hospitals and medical experts. As artificial intelligence has revolutionized various fields of life, health has also benefited from it. The existing architecture of store-and-forward method of conventional telemedicine is facing some problems, some of which are the need for a local health center with dedicated staff, need for medical equipment to prepare patient reports, time constraint of 24–48 hours in receiving diagnosis and medication details from a medical expert in a main hospital, cost of local health centers, and need for Wi-Fi connection. In this paper, we introduce a novel and intelligent healthcare system that is based on modern technologies like Internet of things (IoT) and machine learning. This system is intelligent enough to sense and process a patient’s data through a medical decision support system. This system is low-cost solution for the people of remote areas; they can use it to find out whether they are suffering from a serious health issue and cure it accordingly by contacting near hospitals. The results of the experiments also show that the proposed system is efficient and intelligent enough to provide health facilities. The results presented in this paper are the proof of the concept.

XAI-based Medical Decision Support System Model

XAI-based Medical Decision Support System Model

Decision support system for predicting metabolic risks in patients with comorbid non-alcoholic fatty liver disease and gallstone disease after cholecystectomy

183 patients with non-alcoholic fatty liver disease (NAFLD) were included into the case control study. Objective. Development a medical decision support system to improve the prediction of metabolic risks in persons with comorbid NAFLD and gallstone disease (GD) after cholecystectomy. Materials and methods. The main group was represented by patients with NAFLD and GD (n = 88), of which 53 patients underwent cholecystectomy. The comparison group was represented by patients with NAFLD without GD (n = 95). A standard laboratory and instrumental examinations were performed, including elastometry to assess of the stage of liver fibrosis and assessment of the level of hormones leptin, its soluble receptor, adiponectin and insulin. Results. Patients suffering from GD and NAFLD had a symptom of dyspepsia and general weakness. High prevalence of type 2 diabetes (rs = 0.164; р ≤ 0.01) and сoronary heart disease (25.00 % versus 9.47 % in the comparison group; р ≤ 0.01), high level of LDL and GGT (rs = 0.228, р ≤ 0.01 and rs = 0.298, р ≤ 0.01) were found in patients with GD after cholecystectomy. The phenomenon of insulin and leptin resistance, high levels of adiponectin were detected in patients suffering from NAFLD and GD. Hyperleptinemia was observed in NAFLD patients with GD after cholecystectomy (H = 5.812, p ≤ 0.05, rs = 0.313, p ≤ 0.05). Cholecystectomy in patients suffering from GD and NAFLD was associated with the formation of progressive stages of liver fibrosis (rs = 0.366; р ≤ 0.01). Conclusion. We have developed a decision support system to assess of the possibility of cholecystectomy in patients with NAFLD and GD according to the level metabolic risk of cardiovascular diseases and NAFLD progression.