Intelligent Clinical Decision Support Research Articles

The Evaluation of a Clinical Decision Support Tool Using Natural Language Processing to Screen Hospitalized Adults for Unhealthy Substance Use: Protocol for a Quasi-Experimental Design.

Automated and data-driven methods for screening using natural language processing (NLP) and machine learning may replace resource-intensive manual approaches in the usual care of patients hospitalized with conditions related to unhealthy substance use. The rigorous evaluation of tools that use artificial intelligence (AI) is necessary to demonstrate effectiveness before system-wide implementation. An NLP tool to use routinely collected data in the electronic health record was previously validated for diagnostic accuracy in a retrospective study for screening unhealthy substance use. Our next step is a noninferiority design incorporated into a research protocol for clinical implementation with prospective evaluation of clinical effectiveness in a large health system. This study aims to provide a study protocol to evaluate health outcomes and the costs and benefits of an AI-driven automated screener compared to manual human screening for unhealthy substance use. A pre-post design is proposed to evaluate 12 months of manual screening followed by 12 months of automated screening across surgical and medical wards at a single medical center. The preintervention period consists of usual care with manual screening by nurses and social workers and referrals to a multidisciplinary Substance Use Intervention Team (SUIT). Facilitated by a NLP pipeline in the postintervention period, clinical notes from the first 24 hours of hospitalization will be processed and scored by a machine learning model, and the SUIT will be similarly alerted to patients who flagged positive for substance misuse. Flowsheets within the electronic health record have been updated to capture rates of interventions for the primary outcome (brief intervention/motivational interviewing, medication-assisted treatment, naloxone dispensing, and referral to outpatient care). Effectiveness in terms of patient outcomes will be determined by noninferior rates of interventions (primary outcome), as well as rates of readmission within 6 months, average time to consult, and discharge rates against medical advice (secondary outcomes) in the postintervention period by a SUIT compared to the preintervention period. A separate analysis will be performed to assess the costs and benefits to the health system by using automated screening. Changes from the pre- to postintervention period will be assessed in covariate-adjusted generalized linear mixed-effects models. The study will begin in September 2022. Monthly data monitoring and Data Safety Monitoring Board reporting are scheduled every 6 months throughout the study period. We anticipate reporting final results by June 2025. The use of augmented intelligence for clinical decision support is growing with an increasing number of AI tools. We provide a research protocol for prospective evaluation of an automated NLP system for screening unhealthy substance use using a noninferiority design to demonstrate comprehensive screening that may be as effective as manual screening but less costly via automated solutions. ClinicalTrials.gov NCT03833804; https://clinicaltrials.gov/ct2/show/NCT03833804. DERR1-10.2196/42971.

Development and application of survey-based artificial intelligence for clinical decision support in managing infectious diseases: A pilot study on a hospital in central Vietnam.

In this study, we developed a simplified artificial intelligence to support the clinical decision-making of medical personnel in a resource-limited setting. We selected seven infectious disease categories that impose a heavy disease burden in the central Vietnam region: mosquito-borne disease, acute gastroenteritis, respiratory tract infection, pulmonary tuberculosis, sepsis, primary nervous system infection, and viral hepatitis. We developed a set of questionnaires to collect information on the current symptoms and history of patients suspected to have infectious diseases. We used data collected from 1,129 patients to develop and test a diagnostic model. We used XGBoost, LightGBM, and CatBoost algorithms to create artificial intelligence for clinical decision support. We used a 4-fold cross-validation method to validate the artificial intelligence model. After 4-fold cross-validation, we tested artificial intelligence models on a separate test dataset and estimated diagnostic accuracy for each model. We recruited 1,129 patients for final analyses. Artificial intelligence developed by the CatBoost algorithm showed the best performance, with 87.61% accuracy and an F1-score of 87.71. The F1-score of the CatBoost model by disease entity ranged from 0.80 to 0.97. Diagnostic accuracy was the lowest for sepsis and the highest for central nervous system infection. Simplified artificial intelligence could be helpful in clinical decision support in settings with limited resources.

Ontology for Symptomatic Treatment of Multiple Sclerosis.

Symptomatic treatment is an essential component in the overall treatment of multiple sclerosis (MS). However, knowledge in this regard is confusing and scattered. Physicians also have challenges in choosing symptomatic treatment based on the patient's condition. To share, update, and reuse this knowledge, the aim of this study was to provide an ontology for MS symptomatic treatment. The Symptomatic Treatment of Multiple Sclerosis Ontology (STMSO) was developed according to Ontology Development 101 and a guideline for developing good ontologies in the biomedical domain. We obtained knowledge and rules through a systematic review and entered this knowledge in the form of classes and subclasses in the ontology. We then mapped the ontology using the Basic Formal Ontology (BFO) and Ontology for General Medical Sciences (OGMS) as reference ontologies. The ontology was built using Protégé Editor in the Web Ontology Language format. Finally, an evaluation was done by experts using criterion-based approaches in terms of accuracy, clarity, consistency, and completeness. The knowledge extraction phase identified 110 articles related to the ontology in the form of 626 classes, 40 object properties, and 139 rules. Five general classes included "patient," "symptoms," "pharmacological treatment," "treatment plan," and "measurement index." The evaluation in terms of standards for biomedical ontology showed that STMSO was accurate, clear, consistent, and complete. STMSO is the first comprehensive semantic representation of the symptomatic treatment of MS and provides a major step toward the development of intelligent clinical decision support systems for symptomatic MS treatment.

Applied picture fuzzy sets with knowledge reasoning and linguistics in clinical decision support system

Motivation: Healthcare systems globally face significant resource and financial challenges. Moreover, these challenges have resulted in an existential paradigm shift driven by: (i) the growth in the demand for healthcare services is exacerbated by a global population characterised by an ageing demographic with increasingly complex healthcare needs, and (ii) rapid developments in healthcare technologies and drug therapies which can be seen in the new and emerging treatment options. A potential solution to address [or at least mitigate] these challenges is ‘telemedicine’ with nurse-led ‘triage’ systems; however, a limiting factor for ‘telemedicine’ is the management of imprecision and uncertainty in the diagnostic process. Contribution: In this paper we introduce a novel rule-based approach predicated on picture fuzzy sets to enable intelligent clinical decision support system which builds on previous research to create an approach predicated on picture fuzzy sets. Our principal contribution lies in the use of expert clinician preferences in a rule-based system which implements knowledge reasoning along with linguistic information to improve the diagnostic performance. Results: In ‘real-world’ case studies (using ethically approved anonymised patient data) we have investigated heart conditions, kidney stones, and kidney infections. Reported results for the proposed approach demonstrate a high level of accuracy in clinical diagnostic accuracy terms with reported accuracy in the range [92% to 95%] and a high confidence level when compared to alternative diagnostic matching methods.

The Impact of Cultural Embeddedness on the Implementation of an Artificial Intelligence Program at Triage: A Qualitative Study.

Triage requires rapid determination of acuity and resources. Current modalities allow for individual judgment, with varied application of algorithmic rules. Although artificial intelligence can improve triage accuracy, gaps remain in understanding implementation facilitators and barriers, especially those related to the cultural understandings by nurses of emergency department presentations. The purpose of this study was to explore the cultural and technological elements of the implementation of an artificial intelligence clinical decision support aid (i.e., KATE) in an emergency nursing triage process in an urban community hospital on the West Coast of the United States. An exploratory qualitative study using semi-structured small group and individual interviews and constant comparison analysis strategies. The sample comprised 13 emergency department triage nurses at one site. Campinha-Bacote's theory of cultural competence framed the study. Responses yielded the overall theme of We know these people and we know these things. Supporting categories included the problem of aire; just another checkbox; gut trumps data; higher acuity with no resources; and technology as a safety net. Participants reported reliance on clinical experience and cultural knowledge to assign acuity. The implementation of an artificial intelligence program was initially received skeptically due to the acontextual nature of AI, but grew to be perceived as a safety net for triage decision making among emergency nurses.

The Ethical Considerations of Artificial Intelligence in Clinical Decision Support

With the explosion in technological innovation facilitating the advent of artificially intelligent systems, specifically clinical decision support, a unique subset of ethical and sustainability concerns arises. Although this technology possesses remarkable potential to revolutionise the healthcare industry, it becomes apparent that an innovative ethical framework must be posited to facilitate integration into the mainstream. Due to the sensitive nature of healthcare, ethical oversights pertaining to incorporation of such technologies would lead to the detriment of its public perception, potentially stigmatising related systems for years to come. By delving into the literature surrounding the idiosyncratic ethical considerations of artificially intelligent clinical decision support in this paper, best practices which seek to mitigate the impact of these concerns emerge. The objective of this work is to assimilate these best practices, which are used in the synthesis of a six principle code of ethics which are as follows: protect healthcare professional authority, ensure technological non-maleficence, cultivate clinical decision support transparency, establish procedures for accountability determination, promote sustainability of artificially intelligence based clinical decision support and encourage equity in the training and deployment of clinical decision support. These principles are then applied to the real world of Watson for Oncology by IBM, to assess the adherence of the product to ethical and sustainability best practices.

Development of a Cost-Effective Intelligent Clinical Decision Support System for Breast Cancer Early Diagnosis and Triage

Women population screening using mammography has dramatically reduced breast cancer rates worldwide. Nowadays, in many countries, the prevention of breast cancer policy is based on frequent and repeated mammographies, followed by breast ultrasound and if necessary, by histological examination in the biological material of the biopsy. However, evaluating mammography findings is considered as a difficult process which can properly performed only by a highly experienced and well-trained medical staff. Subsequently, the interpretation of those findings could be easily influenced by subjective factors and therefore can be prone to diagnostic errors as evidenced in the present study. Breast MRI, is a diagnostic practice indicated in cases of high breast density and at the same time, a high-cost examination for healthcare systems. Furthermore, genetic testing that is used to diagnose hereditary breast cancer, represents a small proportion of breast cancers and at the same time it is a high-cost specialized test. Even the most widely used biopsy, Fine Needle Aspirate (FNA), in some cases involves risks and provides false negative results. This study presents a novel intelligent Clinical Decision Support System (CDSS), which uses data from common practice, non-invasive and low-cost diagnostic tests, together with medical health record, in order to provide clinicians with a viable, cost-effective and accurate diagnostic solution. After implementing several algorithms, Random Forest classifier showed the highest values of sensitivity 96,2 %, specificity 94,6%, PPV 96,2% and NPV 94,6%, being thus an effective algorithm in the development of our innovative CDSS model aiming to constitute a very useful tool in clinical practice for breast cancer early diagnosis.

How Health Systems Decide to Use Artificial Intelligence for Clinical Decision Support

Summary Despite the widespread enthusiasm for artificial intelligence (AI) in health care, adoption of AI-enabled clinical-decision support software (CDS) that supports the detection, diagnosis, and treatment of health conditions in specific patients, as well as broader population health management tools, remains slow. Although previous research has examined the challenges associated with adopting AI and the factors that drive adoption, this article explores the perspectives of multiple health systems that have adopted AI-enabled CDS, with a particular focus on how they evaluate the business case for AI-enabled CDS products. This article outlines how health systems choose which software products to deploy, the major cost factors that health systems consider when operationalizing AI-enabled CDS products, and how the value proposition of the software products varies across health systems. The authors found that adoption of AI is driven by a variety of considerations, including clinical utility, ease of use, and patient safety, but also hospital priorities, interoperability/ease of software integration, physician champions, payment models, and market dynamics.

Intelligent Clinical Decision Support.

Early recognition of pathologic cardiorespiratory stress and forecasting cardiorespiratory decompensation in the critically ill is difficult even in highly monitored patients in the Intensive Care Unit (ICU). Instability can be intuitively defined as the overt manifestation of the failure of the host to adequately respond to cardiorespiratory stress. The enormous volume of patient data available in ICU environments, both of high-frequency numeric and waveform data accessible from bedside monitors, plus Electronic Health Record (EHR) data, presents a platform ripe for Artificial Intelligence (AI) approaches for the detection and forecasting of instability, and data-driven intelligent clinical decision support (CDS). Building unbiased, reliable, and usable AI-based systems across health care sites is rapidly becoming a high priority, specifically as these systems relate to diagnostics, forecasting, and bedside clinical decision support. The ICU environment is particularly well-positioned to demonstrate the value of AI in saving lives. The goal is to create AI models embedded in a real-time CDS for forecasting and mitigation of critical instability in ICU patients of sufficient readiness to be deployed at the bedside. Such a system must leverage multi-source patient data, machine learning, systems engineering, and human action expertise, the latter being key to successful CDS implementation in the clinical workflow and evaluation of bias. We present one approach to create an operationally relevant AI-based forecasting CDS system.

Design and Development of an Intelligent Clinical Decision Support System Applied to the Evaluation of Breast Cancer Risk.

Breast cancer is currently one of the main causes of death and tumoral diseases in women. Even if early diagnosis processes have evolved in the last years thanks to the popularization of mammogram tests, nowadays, it is still a challenge to have available reliable diagnosis systems that are exempt of variability in their interpretation. To this end, in this work, the design and development of an intelligent clinical decision support system to be used in the preventive diagnosis of breast cancer is presented, aiming both to improve the accuracy in the evaluation and to reduce its uncertainty. Through the integration of expert systems (based on Mamdani-type fuzzy-logic inference engines) deployed in cascade, exploratory factorial analysis, data augmentation approaches, and classification algorithms such as k-neighbors and bagged trees, the system is able to learn and to interpret the patient’s medical-healthcare data, generating an alert level associated to the danger she has of suffering from cancer. For the system’s initial performance tests, a software implementation of it has been built that was used in the diagnosis of a series of patients contained into a 130-cases database provided by the School of Medicine and Public Health of the University of Wisconsin-Madison, which has been also used to create the knowledge base. The obtained results, characterized as areas under the ROC curves of 0.95–0.97 and high success rates, highlight the huge diagnosis and preventive potential of the developed system, and they allow forecasting, even when a detailed and contrasted validation is still pending, its relevance and applicability within the clinical field.

Development and Application of Survey-Based Artificial Intelligence for Clinical Decision Support in Managing Infectious Diseases a Pilot Study on a Hospital in Central Vietnam

Development and Application of Survey-Based Artificial Intelligence for Clinical Decision Support in Managing Infectious Diseases a Pilot Study on a Hospital in Central Vietnam

Development of an intelligent clinical decision support system for the early prediction of diabetic nephropathy

BackgroundDiabetic nephropathy (DN) is the most common microvascular complication of diabetes mellitus (DM) and is identified as a leading cause of the end-stage renal disease (ESRD). Considering the importance of early prediction of individuals at risk of this complication, the use of intelligent models through machine learning (ML) algorithms can be helpful. Therefore, this study aimed to identify the influential variables in predicting DN and fed them as inputs to develop an ML-based decision support system (DSS) for DN diagnosis. MethodsThe data of 327 patients with diabetes (types 1 and 2) were retrospectively analyzed. After data preparation, the genetic algorithm (GA) feature selection method was used to identify the predictor variables affecting DN. Then, several ML algorithms, including the support vector machine (SVM), decision tree (DT), K-nearest neighbors (KNN), and artificial neural networks (ANN) were used to train predictive models based on the selected features. Afterward, the performance of the developed models was evaluated using sensitivity, specificity, and accuracy criteria in 10 independent runs. Finally, the DSS was developed based on the best fit model in the C# programming language. ResultsOur findings illustrated that age, hemoglobin A1c (HbA1c) test, diastolic arterial pressure (DAP), systolic arterial pressure (SAP), fasting glycemia rate (FGR), and DM involvement time were the most important factors in predicting DN. Moreover, to predict the DN, GA combined with the DT algorithm obtained the highest performance in terms of accuracy, sensitivity, specificity, and area under the curve (AUC), equal to 98.9, 98.6, 99.2, and 98.9%, respectively. ConclusionsThe results revealed that GA combined with the DT classifier predicted DN with significant accuracy. Thus, the DSS developed based on DT can be considered a reliable tool to help physicians make decisions. Future studies are warranted to further validate the applicability of our model in clinical settings.

A Specialized Shell for Intelligent Systems of Prescribing Medication

This paper analyzes the existing decision support systems for prescription of drug therapy. The main principles of development and architecture of an intelligent clinical decision support system that is implemented as a specialized shell are described. The unique features of the system, as well as information and software components that are part of it, are shown. The presented examples demonstrate all the proposed solutions.

High precision brain tumor classification model based on deep transfer learning and stacking concepts

In this article, we proposed an intelligent clinical decision support system for the detection and classification of brain tumor from risk of malignancy index (RMI) images. To overcome the lack of labeled training data needed to train convolutional neural networks, we have used a deep transfer learning and stacking concepts. For this, we choosed seven convolutional neural networks (CNN) architectures already pre-trained on an ImageNet dataset that we precisely fit on magnetic resonance imaging (MRI) of brain tumors collected from the brain tumor segmentation (BraTS) 19 database. To improve the accuracy of our global model, we only predict as output the prediction that obtained the maximum score among the predictions of the seven pre-trained CNNs. We used a 10-way cross-validation approach to assess the performance of our main 2-class model: low-grade glioma (LGG) and high-grade glioma (HGG) brain tumors. A comparison of the results of our proposed model with those published in the literature, shows that our proposed model is more efficient than those published with an average test precision of 98.67%, an average f1 score of 98.62%, a test precision average of 98.06% and an average test sensitivity of 98.33%.

Medical practitioner's adoption of intelligent clinical diagnostic decision support systems: A mixed-methods study

Artificial intelligence-based clinical diagnostic decision support systems promise transformational improvements in doctors’ efficiency and accuracy. Nevertheless, low adoption rates suggest that this innovation could fail without adequate uptake. This study uses a mixed-methods approach to develop and test a model based on theories of Unified Theory of Acceptance and Use of Technology, status quo bias, and technology trust. The results show that performance expectancy, effort expectancy, social influence, initial trust, and resistance to change predict intention to use. Further, inertia, perceived threat, and risks (medico-legal and performance) determine resistance to change. Measures for alleviating resistance and improving adoption are proposed.

Artificial Intelligence for Clinical Decision Support in Sepsis.

Sepsis is one of the main causes of death in critically ill patients. Despite the continuous development of medical technology in recent years, its morbidity and mortality are still high. This is mainly related to the delay in starting treatment and non-adherence of clinical guidelines. Artificial intelligence (AI) is an evolving field in medicine, which has been used to develop a variety of innovative Clinical Decision Support Systems. It has shown great potential in predicting the clinical condition of patients and assisting in clinical decision-making. AI-derived algorithms can be applied to multiple stages of sepsis, such as early prediction, prognosis assessment, mortality prediction, and optimal management. This review describes the latest literature on AI for clinical decision support in sepsis, and outlines the application of AI in the prediction, diagnosis, subphenotyping, prognosis assessment, and clinical management of sepsis. In addition, we discussed the challenges of implementing and accepting this non-traditional methodology for clinical purposes.

Determination of the Most Important Diagnostic Criteria for COVID-19: A Step forward to Design an Intelligent Clinical Decision Support System

Determination of the Most Important Diagnostic Criteria for COVID-19: A Step forward to Design an Intelligent Clinical Decision Support System

Multi-disease prediction using LSTM recurrent neural networks

Prediction of future clinical events (e.g., disease diagnoses) is an important machine learning task in healthcare informatics research. In this work, we propose a deep learning approach to perform multi-disease prediction for intelligent clinical decision support. The proposed approach utilizes a long short-term memory network and extends it with two mechanisms (i.e., time-aware and attention-based) to conduct multi-label classification based on patients’ clinical visit records. The former mechanism (time-aware) is used to handle the temporal irregularity across clinical visits whereas the latter mechanism (attention-based) assists in determining the importance of each visit for the prediction task. Using a large clinical record data set (over 5 million records) collected from a hospital in Southeast China, we show that our proposed approach outperforms a variety of traditional and deep learning methods in predicting future disease diagnoses. We further study the impacts of different time interval choices for the time-aware mechanism and compare the performances of existing attention-based mechanisms with the one proposed in our study. Our work has implications for supporting physician diagnoses via the use of intelligent systems and more broadly for improving the quality of healthcare service.

An interpretable deep-learning model for early prediction of sepsis in the emergency department

SummarySepsis is a life-threatening condition with high mortality rates and expensive treatment costs. Early prediction of sepsis improves survival in septic patients. In this paper, we report our top-performing method in the 2019 DII National Data Science Challenge to predict onset of sepsis 4 h before its diagnosis on electronic health records of over 100,000 unique patients in emergency departments. A long short-term memory (LSTM)-based model with event embedding and time encoding is leveraged to model clinical time series and boost prediction performance. Attention mechanism and global max pooling techniques are utilized to enable interpretation for the deep-learning model. Our model achieved an average area under the curve of 0.892 and was selected as one of the winners of the challenge for both prediction accuracy and clinical interpretability. This study paves the way for future intelligent clinical decision support, helping to deliver early, life-saving care to the bedside of septic patients.

Architecture for Intensive Care Data Processing and Visualization in Real-time

Abstract Clinical data is growing every day. Ergo, to treat, store and publish such data is an emergent task. Furthermore, analysing data in real-time using streaming and processing technologies and methods, in order to obtain quality data, prepared to support decision making is of extreme value. Big Data emerged with the introduction of real-time processing, thus revolutionizing traditional technologies and techniques through the ability to deal with the volume, speed and variety of data. Countless studies have been proposed in the healthcare domain in search of solutions that allow the flow of data in real-time. However, the work presented hereby is distinguished by allowing the collection, processing, storage and analysis of Intensive Care Units (ICU) data, both collected in real-time from bedside monitors but also stored in a historical repository. The architecture proposed makes use of current technologies, like Nextgen Connector as message supplier and integrator, Elasticsearch as a search index, Kibana for viewing stored data and Grafana for real-time streaming. This article is part of the ICDS4IM project - Intelligent Clinical Decision Support in Intensive Care Medicine to support the experimentation of data processing techniques and technologies, based in HL7 format and collected in real-time so that it can be made available through Health Information Systems across the healthcare institutions.