Healthcare Decision Support Research Articles

Improving Clinical Decision Support in Health Care Through AIs

Improving Clinical Decision Support in Health Care Through AIs

SympGAN: A systematic knowledge integration system for symptom–gene associations network

Phenotypes (i.e., symptoms and clinical signs) are essential for clinical diagnosis and research related to symptom science and precision health. As clinical observational manifestations of a disease, symptoms are clinically significant because they act as direct causes for patients to seek medical care and the primary indicators for clinicians to provide diagnosis/treatments. However, a comprehensive phenotypic knowledge base and high-quality symptom–gene associations are lacking. Therefore, a thorough understanding of the relationships between symptoms and other entities is urgently needed to support scientific research and clinical health care. In this paper, we constructed a systematic, large-scale, and high-quality symptom-gene associations network system named SympGAN (accessible at http://www.sympgan.org/). We provide access to the database with millions of associations between symptoms, genes, diseases, and drugs, as well as the system for users to search, analyze, knowledge inference, and present data visualization. We utilize state-of-the-art machine learning and deep learning algorithms as the backbone to form the final dataset. In addition, we utilize the RoBERTa-PubMed neural network for name entity recognition to assist in data screening. The knowledge graph is adopted to organize the relationships between different entities. We adopt ConvE, TuckER, and HypER methods for knowledge completion experiments to validate the quality of final knowledge graph triples. Based on the results, we provide online automatic knowledge inference interfaces. The system, SympGAN, has promising value for disease diagnosis, decision support in health care, precision health, and scientific research, as researchers and practitioners can easily access information about symptoms, diseases, targets, gene ontology, and drugs.

Doctor selection based on aspect-based sentiment analysis and neutrosophic TOPSIS method

Background:Physician-rating sites have become a convenient platform for patients to choose doctors. However, selecting an appropriate doctor through numerous online reviews is challengeable for patients. Although studies show that patients have different preferences for the aspects of healthcare services, existing doctor ranking methods rarely consider such preference information. Besides, they seldom handle the neutral sentiment information in patient reviews. Methods:To better assist patients in doctor selection, we propose a novel decision-making method that combines aspect-based sentiment analysis, single-valued neutrosophic sets and an extended Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method. The method utilizes Word2Vec to construct a feature dictionary of patient satisfaction. A rule-based approach is employed to extract the aspects and the related sentiments from patients’ text reviews. Moreover, it takes advantage of single-valued neutrosophic sets to address the positive, neutral and negative sentiment information. On this basis, we rank the doctors with an extended TOPSIS method considering the patient’s attitudinal character and the preference information. Results:A case study on a review dataset demonstrates the stability and flexibility of the method. The comparative analysis reveals that our method can efficiently reduce information loss and it is more practical than existing methods. Conclusion:Considering patients’ different preferences for the aspects of healthcare services, the method can better provide healthcare decision support for patients.

Bias Analysis in Healthcare Time Series (BAHT) Decision Support Systems from Meta Data.

One of the hindrances in the widespread acceptance of deep learning-based decision support systems in healthcare is bias. Bias in its many forms occurs in the datasets used to train and test deep learning models and is amplified when deployed in the real world, leading to challenges such as model drift. Recent advancements in the field of deep learning have led to the deployment of deployable automated healthcare diagnosis decision support systems at hospitals as well as tele-medicine through IoT devices. Research has been focused primarily on the development and improvement of these systems leaving a gap in the analysis of the fairness. The domain of FAccT ML (fairness, accountability, and transparency) accounts for the analysis of these deployable machine learning systems. In this work, we present a framework for bias analysis in healthcare time series (BAHT) signals such as electrocardiogram (ECG) and electroencephalogram (EEG). BAHT provides a graphical interpretive analysis of bias in the training, testing datasets in terms of protected variables, and analysis of bias amplification by the trained supervised learning model for time series healthcare decision support systems. We thoroughly investigate three prominent time series ECG and EEG healthcare datasets used for model training and research. We show the extensive presence of bias in the datasets leads to potentially biased or unfair machine-learning models. Our experiments also demonstrate the amplification of identified bias with an observed maximum of 66.66%. We investigate the effect of model drift due to unanalyzed bias in datasets and algorithms. Bias mitigation though prudent is a nascent area of research. We present experiments and analyze the most prevalently accepted bias mitigation strategies of under-sampling, oversampling, and the use of synthetic data for balancing the dataset through augmentation. It is important that healthcare models, datasets, and bias mitigation strategies should be properly analyzed for a fair unbiased delivery of service.

Spatial CT perfusion data helpful in automatically locating vessel occlusions for acute ischemic stroke patients.

Locating a vessel occlusion is important for clinical decision support in stroke healthcare. The advent of endovascular thrombectomy beyond proximal large vessel occlusions spurs alternative approaches to locate vessel occlusions. We explore whether CT perfusion (CTP) data can help to automatically locate vessel occlusions. We composed an atlas with the downstream regions of particular vessel segments. Occlusion of these segments should result in the hypoperfusion of the corresponding downstream region. We differentiated between seven-vessel occlusion locations (ICA, proximal M1, distal M1, M2, M3, ACA, and posterior circulation). We included 596 patients from the DUtch acute STroke (DUST) multicenter study. Each patient CTP data set was processed with perfusion software to determine the hypoperfused region. The downstream region with the highest overlap with the hypoperfused region was considered to indicate the vessel occlusion location. We assessed the indications from CTP against expert annotations from CTA. Our atlas-based model had a mean accuracy of 86% and could achieve substantial agreement with the annotations from CTA according to Cohen's kappa coefficient (up to 0.68). In particular, anterior large vessel occlusions and occlusions in the posterior circulation could be located with an accuracy of 80 and 92%, respectively. The spatial layout of the hypoperfused region can help to automatically indicate the vessel occlusion location for acute ischemic stroke patients. However, variations in vessel architecture between patients seemed to limit the capacity of CTP data to distinguish between vessel occlusion locations more accurately.

CAreFL: Enhancing smart healthcare with Contribution‐Aware Federated Learning

AbstractArtificial intelligence (AI) is a promising technology to transform the healthcare industry. Due to the highly sensitive nature of patient data, federated learning (FL) is often leveraged to build models for smart healthcare applications. Existing deployed FL frameworks cannot address the key issues of varying data quality and heterogeneous data distributions across multiple institutions in this sector. In this paper, we report our experience developing and deploying the Contribution‐Aware Federated Learning (CAreFL) framework for smart healthcare. It provides an efficient and accurate approach to fairly evaluate FL participants’ contributions to model performance without exposing their private data, and improves the FL model training protocol by allowing the best performing intermediate sub‐models to be distributed to participants for FL training. Since its deployment by Yidu Cloud Technology Inc. in March 2021, CAreFL has served eight well‐established medical institutions in China to build healthcare decision support models. It can perform contribution evaluations close to three times faster than the best existing approach and has improved the average accuracy of the resulting models by more than 2% compared to the previous system (which is significant in industrial settings). To the best of our knowledge, it is the first CAreFL successfully deployed in the healthcare industry.

Improving prescribing: a feasibility study of pharmacogenetic testing with clinical decision support in primary healthcare in Singapore.

The study of genetic variation as a factor influencing drug safety, efficacy, and effectiveness has brought about significant breakthroughs in understanding the clinical application of gene-drug interactions to better manage drug therapy. This study was designed to assess the feasibility of collecting buccal samples by general practitioners (GPs) at private practices in Singapore within a usual consultation, incorporating use of a pharmacogenetics-based medical decision support system to guide subsequent drug dosing. We used a prospective cohort study design, with GPs recruiting 189 patients between October 2020 and March 2021. The genotypes of 51 biallelic SNPs were determined using Illumina Infinium Global Screening Array. Seven GPs from 6 private practices recruited and obtained buccal samples from a total of 189 patients. All patients had at least one actionable variant. The prevalence of patients having 2, 3, or 4 variants was 37.0%, 32.8%, and 12.7%, respectively. Potential alterations to medications were identified using the Clinical Decision Support System. Patients were accepting and the GPs were enthusiastic about the potential of pharmacogenetics to personalize medicine for their patients. This is the first study in Singapore to demonstrate the feasibility of pharmacogenetic testing in primary care. The high prevalence of genetic variants underscores the potential use of pharmacogenetics in this setting.

Enabling Fairness in Healthcare Through Machine Learning

The use of machine learning systems for decision-support in healthcare may exacerbate health inequalities. However, recent work suggests that algorithms trained on sufficiently diverse datasets could in principle combat health inequalities. One concern about these algorithms is that their performance for patients in traditionally disadvantaged groups exceeds their performance for patients in traditionally advantaged groups. This renders the algorithmic decisions unfair relative to the standard fairness metrics in machine learning. In this paper, we defend the permissible use of affirmative algorithms; that is, algorithms trained on diverse datasets that perform better for traditionally disadvantaged groups. Whilst such algorithmic decisions may be unfair, the fairness of algorithmic decisions is not the appropriate locus of moral evaluation. What matters is the fairness of final decisions, such as diagnoses, resulting from collaboration between clinicians and algorithms. We argue that affirmative algorithms can permissibly be deployed provided the resultant final decisions are fair.

Contribution-Aware Federated Learning for Smart Healthcare

Artificial intelligence (AI) is a promising technology to transform the healthcare industry. Due to the highly sensitive nature of patient data, federated learning (FL) is often leveraged to build models for smart healthcare applications. Existing deployed FL frameworks cannot address the key issues of varying data quality and heterogeneous data distributions across multiple institutions in this sector. In this paper, we report our experience developing and deploying the Contribution-Aware Federated Learning (CAFL) framework for smart healthcare. It provides an efficient and accurate approach to fairly evaluate FL participants' contribution to model performance without exposing their private data, and improves the FL model training protocol to allow the best performing intermediate models to be distributed to participants for FL training. Since its deployment in Yidu Cloud Technology Inc. in March 2021, CAFL has served 8 well-established medical institutions in China to build healthcare decision support models. It can perform contribution evaluations 2.84 times faster than the best existing approach, and has improved the average accuracy of the resulting models by 2.62% compared to the previous system (which is significant in industrial settings). To our knowledge, it is the first contribution-aware federated learning successfully deployed in the healthcare industry.

Exposure to Ambient Air Pollutant PM10 in the Second Trimester of Pregnancy Is Associated with Preterm Birth: A Birth-Based Health Information Cohort Study.

Objectives We evaluated the effects of exposure to high concentrations of particulate matter (PM)10 on preterm birth (PTB) and identified a critical concentration of PM10 that could lead to PTB via a birth-based health information cohort study. Methods We conducted a birth-based cohort study consisting of nonanomalous singleton births at 22-42 weeks. PTB was defined as babies born alive before 37 weeks of pregnancy. Pregnancy period exposure averages were estimated for PM10 based on the China National Environmental Monitoring Centre (CNEMC). Pregnant women who lived within 50 km of the monitor station were recruited into this study. Logistic regression analyses were performed to determine the association between PTB and exposure to PM10 at different pregnancy periods with adjustment for confounding factors. Results The relative frequency of PTB was 8.7% in the study cohort of 5,291 singleton live births. A total of 1137 women had a high level of PM10 exposure (≥60 μg/m3) in the second trimester of pregnancy. The average concentrations of PM10 in the first, second, and third trimesters of pregnancy and throughout pregnancy were 53.8 μg/m3, 54.2 μg/m3, 55.6 μg/m3, and 54.3 μg/m3, respectively. The generalized additive model (GAM) analysis showed that there was a nonlinear correlation between PM10 and PTB in the second trimester of pregnancy (P < 0.001). The adjusted odds ratio between PTB and low concentration PM10 exposure (PM10 < 60 μg/m3) in the second trimester of pregnancy was 1.01 (95% CI 0.95-1.05). However, high PM10 exposure (PM10 ≥ 60 μg/m3) in the second trimester of pregnancy had an increased PTB risk even after adjustment for coexisting risk factors with an adjusted odds ratio of 1.78 (95% CI 1.69-1.87), and the incidence of PTB increased with an increase in PM10 exposure. Conclusions Our research discovered that exposure to high levels of PM10 increases the risk of PTB and the second trimester is the most vulnerable gestational period to ambient air pollution exposure. PM10 concentrations more than 60 μg/m3 are detrimental to pregnant women in their second trimester. This study has implications for health informatics-oriented healthcare decision support systems.

HDSS: a healthcare decision support system on combining domain knowledge and data analytics for predicting potential risk of mental health

HDSS: a healthcare decision support system on combining domain knowledge and data analytics for predicting potential risk of mental health

HDSS: A Healthcare Decision Support System on Combining Domain Knowledge and Data Analytics for Predicting Potential Risk of Mental Health

HDSS: A Healthcare Decision Support System on Combining Domain Knowledge and Data Analytics for Predicting Potential Risk of Mental Health

Population Preferences for Performance and Explainability of Artificial Intelligence in Health Care: Choice-Based Conjoint Survey.

BackgroundCertain types of artificial intelligence (AI), that is, deep learning models, can outperform health care professionals in particular domains. Such models hold considerable promise for improved diagnostics, treatment, and prevention, as well as more cost-efficient health care. They are, however, opaque in the sense that their exact reasoning cannot be fully explicated. Different stakeholders have emphasized the importance of the transparency/explainability of AI decision making. Transparency/explainability may come at the cost of performance. There is need for a public policy regulating the use of AI in health care that balances the societal interests in high performance as well as in transparency/explainability. A public policy should consider the wider public’s interests in such features of AI.ObjectiveThis study elicited the public’s preferences for the performance and explainability of AI decision making in health care and determined whether these preferences depend on respondent characteristics, including trust in health and technology and fears and hopes regarding AI.MethodsWe conducted a choice-based conjoint survey of public preferences for attributes of AI decision making in health care in a representative sample of the adult Danish population. Initial focus group interviews yielded 6 attributes playing a role in the respondents’ views on the use of AI decision support in health care: (1) type of AI decision, (2) level of explanation, (3) performance/accuracy, (4) responsibility for the final decision, (5) possibility of discrimination, and (6) severity of the disease to which the AI is applied. In total, 100 unique choice sets were developed using fractional factorial design. In a 12-task survey, respondents were asked about their preference for AI system use in hospitals in relation to 3 different scenarios.ResultsOf the 1678 potential respondents, 1027 (61.2%) participated. The respondents consider the physician having the final responsibility for treatment decisions the most important attribute, with 46.8% of the total weight of attributes, followed by explainability of the decision (27.3%) and whether the system has been tested for discrimination (14.8%). Other factors, such as gender, age, level of education, whether respondents live rurally or in towns, respondents’ trust in health and technology, and respondents’ fears and hopes regarding AI, do not play a significant role in the majority of cases.ConclusionsThe 3 factors that are most important to the public are, in descending order of importance, (1) that physicians are ultimately responsible for diagnostics and treatment planning, (2) that the AI decision support is explainable, and (3) that the AI system has been tested for discrimination. Public policy on AI system use in health care should give priority to such AI system use and ensure that patients are provided with information.

Development of Segmentation and Classification Algorithms for Computed Tomography Images of Human Kidney Stone

Computed tomography (CT) scan diagnostics procedures adopt the use of image information retrieval system with the help of radiographer’s expertise. However, this technique is prone to errors. Significant height of accuracy is required in healthcare decision support, as 20% of CT scans are associated with error. The application of artificial intelligence (AI) can improve performance level, mitigate human error, and enhance clinical decision support in the context of time and accuracy. The study introduced machine learning algorithm to analyze stream of anonymous CT scans of kidney. The research adopted deep learning approach for segmentation and classification of kidney stone (renal calculi) images in Python (with Keras and TensorFlow) environment. A control volume of data along with 336 kidney stone images were used to train the deep learning network with 10 testing images. The training images were divided into two sets (folders) as follows; one was labeled as STONE (containing 167 images) and the other as NO-STONE (containing 169 images); 10 iterations were performed for model training. The network layers were structured as input layer in the following with 2-D convolutional neural network machine learning (CNN-ML), ReLU activation, Maxpooling, and fully connected (dense) layer including the sigmoid activation layer. The training adopted a batch size of 8 with 10% validation. The output result, upon testing the model, has an accuracy of 90%, sensitivity value of 80% and effectiveness of 89%. The segmentation and classification algorithm model could be embedded in future CT diagnostic procedure to enhance medical decision support and accuracy.

Hierarchical Reinforcement Learning for Decision Support in Health Care

Optimal decision-making is critical within many organizations. A large number of these organizations are structured hierarchically but make sequential decisions (like health care, for example). Using data collected over time by these organizations, we have been able to successfully apply reinforcement learning (RL) to many sequential decision-making problems. In doing this, however, we are not taking advantage of the benefits of their hierarchical structures and the ways different layers affect each other, and thus are not able to learn optimal decision-making techniques.Hierarchical reinforcement learning (HRL) is a powerful tool for solving extended problems with spare rewards. HRL decomposes a RL problem into a hierarchy of subtasks to be solved individually using RL. Because of this, the fundamental concept of HRL applies nicely to our problem. Unfortunately, due to their inability to handle multiple agents, implement batch learning, and model concurrent activities, classic HRL frameworks are not quite suitable for our problem. In our work, we plan to formalize a new HRL framework that is capable of building sequential decision-making support models using datasets collected from stochastic behavioral policies. In this paper, we bring light to probable obstacles as well as potential solutions to existing problems.

Patients, primary care, and policy: Agent-based simulation modeling for health care decision support

Primary care systems are a cornerstone of universally accessible health care. The planning, analysis, and adaptation of primary care systems is a highly non-trivial problem due to the systems’ inherent complexity, unforeseen future events, and scarcity of data. To support the search for solutions, this paper introduces the hybrid agent-based simulation model SiM-Care. SiM-Care models and tracks the micro-interactions of patients and primary care physicians on an individual level. At the same time, it models the progression of time via the discrete-event paradigm. Thereby, it enables modelers to analyze multiple key indicators such as patient waiting times and physician utilization to assess and compare primary care systems. Moreover, SiM-Care can evaluate changes in the infrastructure, patient behavior, and service design. To showcase SiM-Care and its validation through expert input and empirical data, we present a case study for a primary care system in Germany. Specifically, we study the immanent implications of demographic change on rural primary care and investigate the effects of an aging population and a decrease in the number of physicians, as well as their combined effects.

CEFEs: A CNN Explainable Framework for ECG Signals

In the healthcare domain, trust, confidence, and functional understanding are critical for decision support systems, therefore, presenting challenges in the prevalent use of black-box deep learning (DL) models. With recent advances in deep learning methods for classification tasks, there is an increased use of deep learning in healthcare decision support systems, such as detection and classification of abnormal Electrocardiogram (ECG) signals. Domain experts seek to understand the functional mechanism of black-box models with an emphasis on understanding how these models arrive at specific classification of patient medical data. In this paper, we focus on ECG data as the healthcare data signal to be analyzed. Since ECG is a one-dimensional time-series data, we target 1D-CNN (Convolutional Neural Networks) as the candidate DL model. Majority of existing interpretation and explanations research has been on 2D-CNN models in non-medical domain leaving a gap in terms of explanation of CNN models used on medical time-series data. Hence, we propose a modular framework, CNN Explanations Framework for ECG Signals (CEFEs), for interpretable explanations. Each module of CEFEs provides users with the functional understanding of the underlying CNN models in terms of data descriptive statistics, feature visualization, feature detection, and feature mapping. The modules evaluate a model’s capacity while inherently accounting for correlation between learned features and raw signals which translates to correlation between model’s capacity to classify and it’s learned features. Explainable models such as CEFEs could be evaluated in different ways: training one deep learning architecture on different volumes/amounts of the same dataset, training different architectures on the same data set or a combination of different CNN architectures and datasets. In this paper, we choose to evaluate CEFEs extensively by training on different volumes of datasets with the same CNN architecture. The CEFEs’ interpretations, in terms of quantifiable metrics, feature visualization, provide explanation as to the quality of the deep learning model where traditional performance metrics (such as precision, recall, accuracy, etc.) do not suffice.

Mobile Healthcare Design Research: A Special Issue for Information Systems Researchers

Abstract Special Section Editorial - Mobile Decision Support and Analytics for Healthcare: Citizen, Organization, Governmental and Technological Perspectives

Evaluations of Decision Support Tools Are Preference-Sensitive and Interest-Conflicted: The Case of Deliberation Aids.

The questions 'What constitutes a good health care decision?', and, by extension, 'What constitutes good healthcare decision support?' continue to be asked. The most developed answers focus largely, often exclusively, on the quality of the 'deliberation' component as the determinant of the quality of the decision or decision aid. We argue that these answers and resulting aids reflect the preferences of healthcare professionals and aid developers and that these preferences are closely aligned with their interests. Some interests are material, but many professional, institutional, intellectual, methodological, and ethical. Successful promotion of a particular preference-sensitive, interest-conflicted decision aid does not change its ontological nature. Conflicts of interest are therefore universal and of concern only when this ontology is denied and if aids based on alternative interest-based preferences, such as technologies involving numerical analytic calculation, are subjected to discrimination.

Knowledge-Based Decision Support in Healthcare via Near Field Communication.

The benefits of automatic identification technologies in healthcare have been largely recognized. Nevertheless, unlocking their potential to support the most knowledge-intensive medical tasks requires to go beyond mere item identification. This paper presents an innovative Decision Support System (DSS), based on a semantic enhancement of Near Field Communication (NFC) standard. Annotated descriptions of medications and patient’s case history are stored in NFC transponders and used to help caregivers providing the right therapy. The proposed framework includes a lightweight reasoning engine to infer possible incompatibilities in treatment, suggesting substitute therapies. A working prototype is presented in a rheumatology case study and preliminary performance tests are reported. The approach is independent from back-end infrastructures. The proposed DSS framework is validated in a limited but realistic case study, and performance evaluation of the prototype supports its practical feasibility. Automated reasoning on knowledge fragments extracted via NFC enables effective decision support not only in hospital centers, but also in pervasive IoT-based healthcare contexts such as first aid, ambulance transport, rehabilitation facilities and home care.