Decision support system for triage management: A hybrid approach using rule-based reasoning and fuzzy logic

The effects of applying artificial intelligence to triage in the emergency department: A systematic review of prospective studies

INTRODUCTION

Accurate and rapid triage can reduce undertriage and overtriage, which may improve emergency department flow. This study aimed to identify the effects of a prospective study applying artificial intelligence-based triage in the clinical field.

DESIGN

Systematic review of prospective studies.

METHODS

CINAHL, Cochrane, Embase, PubMed, ProQuest, KISS, and RISS were searched from March 9 to April 18, 2023. All the data were screened independently by three researchers. The review included prospective studies that measured outcomes related to AI-based triage. Three researchers extracted data and independently assessed the study's quality using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) protocol.

RESULTS

Of 1633 studies, seven met the inclusion criteria for this review. Most studies applied machine learning to triage, and only one was based on fuzzy logic. All studies, except one, utilized a five-level triage classification system. Regarding model performance, the feed-forward neural network achieved a precision of 33% in the level 1 classification, whereas the fuzzy clip model achieved a specificity and sensitivity of 99%. The accuracy of the model's triage prediction ranged from 80.5% to 99.1%. Other outcomes included time reduction, overtriage and undertriage checks, mistriage factors, and patient care and prognosis outcomes.

CONCLUSION

Triage nurses in the emergency department can use artificial intelligence as a supportive means for triage. Ultimately, we hope to be a resource that can reduce undertriage and positively affect patient health.

PROTOCOL REGISTRATION

We have registered our review in PROSPERO (registration number: CRD 42023415232).

Nurses' experiences of using a computer-based triage decision support system in the emergency department

BACKGROUND

Emergency department triage is a vital and complex decision-making process. There is limited knowledge about nurses' experiences with triage decision support systems in emergency departments.

AIM

This study was conducted to examine nurses' experiences with a computer-based triage decision support system in the emergency department.

STUDY DESIGN

This is a qualitative and phenomenological study. Data were collected through interviews from 14 triage nurses who used a triage decision support system in the emergency department of a university hospital. The nurses were recruited for semi-structured interviews. All interviews were recorded on a voice recorder and then transcribed. Data were analysed with the inductive content analysis method. The interviewer asked comprehensive questions about the nurses' experiences with the triage decision support system.

RESULTS

Three main and 11 sub-themes were elicited as a result of the analysis of the in-depth interviews: (a) the facilitating the triage decision theme, which included help in case of a dilemma, team collaboration, monitoring/supervision, and error reduction sub-themes; (b) the contribution to professionalism theme, which included ease of learning and teaching triage, professional autonomy, creating a database, and evidence-based practice sub-themes; (c) the areas that need improvement theme, which included reducing screen clicks, the effect of the hospital automation system performance, and clinical descriptors not included in the algorithm sub-themes.

CONCLUSION

Triage nurses stated that the decision support system was beneficial and facilitated decision-making. The decision support system enabled triage nurses to make their own decisions using their clinical knowledge and experience, without a restriction on their professional autonomy, and this was perceived positively. It was emphasized that this system could be a support tool in educating nurses new to triage. However, nurses stated that the pace of the decision support system was affected by the performance of the hospital automation system and that the large number of steps used for electing items caused a waste of time. They also added that these were technical areas that needed improvement.

RELEVANCE TO CLINICAL PRACTICE

The study provides important data that will help healthcare organizations and professionals better understand the emergency department nurse triage decision support system and gain a versatile, comprehensive, and general understanding.

Precision emergency medicine

BACKGROUND

Precision health is a burgeoning scientific discipline that aims to incorporate individual variability in biological, behavioral, and social factors to develop personalized health solutions. To date, emergency medicine has not deeply engaged in the precision health movement. However, rapid advances in health technology, data science, and medical informatics offer new opportunities for emergency medicine to realize the promises of precision health.

METHODS

In this article, we conceptualize precision emergency medicine as an emerging paradigm and identify key drivers of its implementation into current and future clinical practice. We acknowledge important obstacles to the specialty-wide adoption of precision emergency medicine and offer solutions that conceive a successful path forward.

RESULTS

Precision emergency medicine is defined as the use of information and technology to deliver acute care effectively, efficiently, and authentically to individual patients and their communities. Key drivers and opportunities include leveraging human data, capitalizing on technology and digital tools, providing deliberate access to care, advancing population health, and reimagining provider education and roles. Overcoming challenges in equity, privacy, and cost is essential for success. We close with a call to action to proactively incorporate precision health into the clinical practice of emergency medicine, the training of future emergency physicians, and the research agenda of the specialty.

CONCLUSIONS

Precision emergency medicine leverages new technology and data-driven artificial intelligence to advance diagnostic testing, individualize patient care plans and therapeutics, and strategically refine the convergence of the health system and the community.

The Value of Emergency Care Data Set (ECDS) Presentation Codes for Predicting Mortality and Inpatient Admission

BACKGROUND

 Early identification of patients at higher risk of death and hospital admission is an important problem in Emergency Departments (ED). Most triage scales were developed before current electronic healthcare records were developed. The implementation of a national Emergency Care Data Set (ECDS) allows for the standardised recording of presenting complaints and the use of Electronic Patient Records (EPR) offers the potential for automated triage. The mortality risk and need for hospital admission associated with the different presenting complaints in a standardised national data set has not been previously reported. This study aimed to quantify the risks of death and hospitalisation from presenting complaints. This would be valuable in developing automated triage tools and decision support software.

METHODS

We conducted an observational retrospective cohort study on patients who visited a single ED in 2021. The presenting complaints related to subsequent attendances were excluded. This patient list was then manually matched with a routinely collected list of deaths. All deaths that occurred within 30 days of attendance were included.

RESULTS

Data was collected from 84,999 patients, of which 1,159 people died within 30 days of attendance. The mortality rate was the highest in cardiac arrest [32 (78.1%)], cardiac arrest due to trauma [2(50%)] and respiratory arrest [3(50%)]. Drowsy [17(12%)], hypothermia [3(13%)] and cyanosis [1(10%)] were also high-risk categories. Chest pain [34(0.6%)] was not a high-risk presenting complaint.

CONCLUSION

The initial presenting complaint in ECDS may be useful to identify people at higher and lower risk of death. This information is useful for building automated triage models.

Optimization of ultrasound-aided extraction of bioactive ingredients from Vitis vinifera seeds using RSM and ANFIS modeling with machine learning algorithm

Plant materials are a rich source of polyphenolic compounds with interesting health-beneficial effects. The present study aimed to determine the optimized condition for maximum extraction of polyphenols from grape seeds through RSM (response surface methodology), ANFIS (adaptive neuro-fuzzy inference system), and machine learning (ML) algorithm models. Effect of five independent variables and their ranges, particle size (X1: 0.5-1 mm), methanol concentration (X2: 60-70% in distilled water), ultrasound exposure time (X3: 18-28 min), temperature (X4: 35-45 °C), and ultrasound intensity (X5: 65-75 W cm-2) at five levels (- 2, - 1, 0, + 1, and + 2) concerning dependent variables, total phenolic content (y1; TPC), total flavonoid content (y2; TFC), 2, 2-diphenyl-1-picrylhydrazyl free radicals scavenging (y3; %DPPH*sc), 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) free radicals scavenging (y4; %ABTS*sc) and Ferric ion reducing antioxidant potential (y5; FRAP) were selected. The optimized condition was observed at X1 = 0.155 mm, X2 = 65% methanol in water, X3 = 23 min ultrasound exposure time, X4 = 40 °C, and X5 = 70 W cm-2 ultrasound intensity. Under this situation, the optimal yields of TPC, TFC, and antioxidant scavenging potential were achieved to be 670.32 mg GAE/g, 451.45 mg RE/g, 81.23% DPPH*sc, 77.39% ABTS*sc and 71.55 μg mol (Fe(II))/g FRAP. This optimal condition yielded equal experimental and expected values. A well-fitted quadratic model was recommended. Furthermore, the validated extraction parameters were optimized and compared using the ANFIS and random forest regressor-ML algorithm. Gas chromatography-mass spectroscopy (GC-MS) and liquid chromatography-mass spectroscopy (LC-MS) analyses were performed to find the existence of the bioactive compounds in the optimized extract.

The effect of computer-based decision support system on emergency department triage: Non-randomised controlled trial

BACKGROUND

Deciding on triage in emergency departments is difficult and requires comprehensive knowledge and experience.

PURPOSE

This study was conducted to evaluate the effect of a "computer-based emergency department triage decision support system (DSS)," which was designed and integrated into the hospital information management system, on triage decision accuracy and triage duration by using real patient data.

METHODS

Single-group, pretest-posttest non-randomised clinical trial. The study was conducted with the real data of patients who had been triaged in the adult emergency department of a university hospital. The pretest was applied between July 16 and September 16, 2019, and the post-test on September 1 and October 31, 2020. In the pre-test and post-test phases of the study, triage decision accuracy rates, and triage duration were evaluated. In the post-test phase, Emergency Triage Decision Support System (ETDSS) was prepared with a rule-based decision trees method using the Emergency Severity Index Version 4 and The Australasian Triage Scale and was integrated into the hospital information management system. The effect of the developed ETDSS was evaluated. The mean, standard deviation, frequency and percentage values were calculated for the descriptive characteristics. Independent samples t-test, analysis of variance, Sidak paired comparison, and Bonferroni tests were applied.

RESULTS

The effect of the computer-based emergency triage DSS on triage management was tested based on the data of 16,409 patients in the pretest phase and 7,765 patients in the posttest phase. While the accuracy rate of nurses' triage decisions was 57.8% in the pretest, it was found to increase to 64.9% in the posttest. The mean duration of triage was 1.47 ± 0.72 in the pretest and 1.79 ± 0.85 min in the posttest.

CONCLUSIONS

The DSS increased triage decision accuracy independently of professional and triage experience and brought the triage duration closer to the time recommended in the literature. Clinically, this is associated with patient safety, quality improvement processes, and professional accountability.

Taxonomy of hybrid architectures involving rule-based reasoning and machine learning in clinical decision systems: A scoping review

BACKGROUND

As the application of Artificial Intelligence (AI) technologies increases in the healthcare sector, the industry faces a need to combine medical knowledge, often expressed as clinical rules, with advances in machine learning (ML), which offer high prediction accuracy at the expense of transparency of decision making.

PURPOSE

This paper seeks to review the present literature, identify hybrid architecture patterns that incorporate rules and machine learning, and evaluate the rationale behind their selection to inform future development and research on the design of transparent and precise clinical decision systems.

METHODS

PubMed, IEEE Explore, and Google Scholar were queried in search for papers from 1992 to 2022, with the keywords: "clinical decision system", "hybrid clinical architecture", "machine learning and clinical rules". Excluded articles did not use both ML and rules or did not provide any explanation of employed architecture. A proposed taxonomy was used to organize the results, analyze them, and depict them in graphical and tabular form. Two researchers, one with expertise in rule-based systems and another in ML, reviewed identified papers and discussed the work to minimize bias, and the third one re-reviewed the work to ensure consistency of reporting.

RESULTS

The authors screened 957 papers and reviewed 71 that met their criteria. Five distinct architecture archetypes were determined: Rules are Embedded in ML architecture (REML) (most used), ML pre-processes input data for Rule-Based inference (MLRB), Rule-Based method pre-processes input data for ML prediction (RBML), Rules influence ML training (RMLT), Parallel Ensemble of Rules and ML (PERML), which was rarely observed in clinical contexts.

CONCLUSIONS

Most architectures in the reviewed literature prioritize prediction accuracy over explainability and trustworthiness, which has led to more complex embedded approaches. Alternatively, parallel (PERML) architectures may be employed, allowing for a more transparent system that is easier to explain to patients and clinicians. The potential of this approach warrants further research.

OTHER

A limitation of the study may be that it reviews scientific literature, while algorithms implemented in clinical practice may present different distributions of motivations and implementations of hybrid architectures.

The Use and Structure of Emergency Nurses' Triage Narrative Data: Scoping Review

BACKGROUND

Emergency departments use triage to ensure that patients with the highest level of acuity receive care quickly and safely. Triage is typically a nursing process that is documented as structured and unstructured (free text) data. Free-text triage narratives have been studied for specific conditions but never reviewed in a comprehensive manner.

OBJECTIVE

The objective of this paper was to identify and map the academic literature that examines triage narratives. The paper described the types of research conducted, identified gaps in the research, and determined where additional review may be warranted.

METHODS

We conducted a scoping review of unstructured triage narratives. We mapped the literature, described the use of triage narrative data, examined the information available on the form and structure of narratives, highlighted similarities among publications, and identified opportunities for future research.

RESULTS

We screened 18,074 studies published between 1990 and 2022 in CINAHL, MEDLINE, Embase, Cochrane, and ProQuest Central. We identified 0.53% (96/18,074) of studies that directly examined the use of triage nurses' narratives. More than 12 million visits were made to 2438 emergency departments included in the review. In total, 82% (79/96) of these studies were conducted in the United States (43/96, 45%), Australia (31/96, 32%), or Canada (5/96, 5%). Triage narratives were used for research and case identification, as input variables for predictive modeling, and for quality improvement. Overall, 31% (30/96) of the studies offered a description of the triage narrative, including a list of the keywords used (27/96, 28%) or more fulsome descriptions (such as word counts, character counts, abbreviation, etc; 7/96, 7%). We found limited use of reporting guidelines (8/96, 8%).

CONCLUSIONS

The breadth of the identified studies suggests that there is widespread routine collection and research use of triage narrative data. Despite the use of triage narratives as a source of data in studies, the narratives and nurses who generate them are poorly described in the literature, and data reporting is inconsistent. Additional research is needed to describe the structure of triage narratives, determine the best use of triage narratives, and improve the consistent use of triage-specific data reporting guidelines.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

RR2-10.1136/bmjopen-2021-055132.

A fuzzy approach to support decision-making in the triage process for suspected COVID-19 patients in Brazil

Triage is a fundamental process in hospitals and emergency care units, as it allows for the classification and prioritization of patient care based on the severity of their clinical conditions. In Brazil, the triage of suspected COVID-19 cases is performed using a specific protocol, which involves manual steps, requiring the completion of four different forms, by four health care professionals. Aiming to investigate the possibility of improving the triage processes in Brazil, this article proposes the use of computational techniques for decision-making based on fuzzy inference systems. We argue that fuzzy set theory is appropriate to the problem because it allows the use of natural language to express the patient's symptoms, making it easier for health care professionals. After modelling the problem in a fuzzy system we applied a pilot test. The model includes symptoms that health professionals currently use to analyse COVID-19 cases. The results suggest that the model presents convergence with the sample data, highlighting its potential application in supporting triage for the classification of the severity of COVID-19 cases. Among the benefits of the proposed model, we emphasize contributions as the reduction of the time and number of professionals required for triage as well as the reduction of exposure of health care professionals and other patients suspected of carrying the virus. In this context, this research provides an opportunity to obtain social contributions regarding the services in public hospitals improvement.

A classification framework for identifying bronchitis and pneumonia in children based on a small-scale cough sounds dataset

Bronchitis and pneumonia are the common respiratory diseases, of which pneumonia is the leading cause of mortality in pediatric patients worldwide and impose intense pressure on health care systems. This study aims to classify bronchitis and pneumonia in children by analyzing cough sounds. We propose a Classification Framework based on Cough Sounds (CFCS) to identify bronchitis and pneumonia in children. Our dataset includes cough sounds from 173 outpatients at the West China Second University Hospital, Sichuan University, Chengdu, China. We adopt aggregation operation to obtain patients’ disease features because some cough chunks carry the disease information while others do not. In the stage of classification in our framework, we adopt Support Vector Machine (SVM) to classify the diseases due to the small scale of our dataset. Furthermore, we apply data augmentation to our dataset to enlarge the number of samples and then adopt Long Short-Term Memory Network (LSTM) to classify. After 45 random tests on RAW dataset, SVM achieves the best classification accuracy of 86.04% and standard deviation of 4.7%. The precision of bronchitis and pneumonia is 93.75% and 87.5%, and their recall is 88.24% and 93.33%. The AUC of SVM and LSTM classification models on the dataset with pitch-shifting data augmentation reach 0.92 and 0.93, respectively. Extensive experimental results show that CFCS can effectively classify children into bronchitis and pneumonia.

Machine learning methods applied to triage in emergency services: A systematic review

BACKGROUND

In emergency services is important to accurately assess and classify symptoms, which may be improved with the help of technology. One mechanism that could help and improve predictions from health records or patient flow is machine learning (ML).

AIM

To analyse the effectiveness of ML systems in triage for making predictions at the emergency department in comparison with other triage scales/scores.

METHODS

Following the PRISMA recommendations, a systematic review was conducted using CINAHL, Cochrane, Cuiden, Medline and Scopus databases with the search equation "Machine learning AND triage AND emergency".

RESULTS

Eleven studies were identified. The studies show that the use of ML methods consistently predict important outcomes like mortality, critical care outcomes and admission, and the need for hospitalization in comparison with scales like Emergency Severity Index or others. Among the ML models considered, XGBoost and Deep Neural Networks obtained the highest levels of prediction accuracy, while Logistic Regression performed obtained the worst values.

CONCLUSIONS

Machine learning methods can be a good instrument for helping triage process with the prediction of important emergency variables like mortality or the need for critical care or hospitalization.

Automation in nursing decision support systems: A systematic review of effects on decision making, care delivery, and patient outcomes

OBJECTIVE

The study sought to summarize research literature on nursing decision support systems (DSSs ); understand which steps of the nursing care process (NCP) are supported by DSSs, and analyze effects of automated information processing on decision making, care delivery, and patient outcomes.

MATERIALS AND METHODS

We conducted a systematic review in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. PubMed, CINAHL, Cochrane, Embase, Scopus, and Web of Science were searched from January 2014 to April 2020 for studies focusing on DSSs used exclusively by nurses and their effects. Information about the stages of automation (information acquisition, information analysis, decision and action selection, and action implementation), NCP, and effects was assessed.

RESULTS

Of 1019 articles retrieved, 28 met the inclusion criteria, each studying a unique DSS. Most DSSs were concerned with two NCP steps: assessment (82%) and intervention (86%). In terms of automation, all included DSSs automated information analysis and decision selection. Five DSSs automated information acquisition and only one automated action implementation. Effects on decision making, care delivery, and patient outcome were mixed. DSSs improved compliance with recommendations and reduced decision time, but impacts were not always sustainable. There were improvements in evidence-based practice, but impact on patient outcomes was mixed.

CONCLUSIONS

Current nursing DSSs do not adequately support the NCP and have limited automation. There remain many opportunities to enhance automation, especially at the stage of information acquisition. Further research is needed to understand how automation within the NCP can improve nurses' decision making, care delivery, and patient outcomes.

Emergency nurse triage in the hospital information management system: A quality improvement study

BACKGROUND

Assessment of nurse triage decision accuracy and triage times is currently carried out through paper-based methods. This quality improvement study aims to develop a method that can assess the accuracy and duration of nurse triage decisions based on a computerized system and to share an example of the application of this method.

METHODS

This is a descriptive quality improvement study. The study was carried out in two stages between March and May 2019. The functionality of the developed method was examined using 3835 patients' triage data, which were obtained between June 1 and 14, 2019.

RESULTS

With this study, the determination of the accuracy and duration of nurse triage decisions was accomplished with a computerized process based on real patient outputs, and the accuracy and duration of these decisions were continuously measured, monitored, and assessed, which is different from paper-based methods. The functionality of the method was evaluated with data from 3835 real patients. The triage decision accuracy rate was 64.4%, and the average duration of triage was 81.3s. Positive feedback on the method was received from all triage nurses.

CONCLUSION

The study result outputs can be integrated into quality processes and can be used internationally as performance assessment criteria and quality indicators for triage nursing.

A Context-Aware Middleware for Context Modeling and Reasoning: A Case-Study in Smart Cultural Spaces

The proliferation of smart things and the subsequent emergence of the Internet of Things has motivated the deployment of intelligent spaces that provide automated services to users. Context-awareness refers to the ability of the system to be aware of the virtual and physical environment, allowing more efficient personalization. Context modeling and reasoning are two important aspects of context-aware computing, since they enable the representation of contextual data and inference of high-level, meaningful information. Context-awareness middleware systems integrate context modeling and reasoning, providing abstraction and supporting heterogeneous context streams. In this work, such a context-awareness middleware system is presented, which integrates a proposed context model based on the adaptation and combination of the most prominent context categorization schemata. A hybrid reasoning procedure, which combines multiple techniques, is also proposed and integrated. The proposed system was evaluated in a real-case-scenario cultural space, which supports preventive conservation. The evaluation showed that the proposed system efficiently addressed both conceptual aspects, through means of representation and reasoning, and implementation aspects, through means of performance.

The potential of artificial intelligence to improve patient safety: a scoping review

Artificial intelligence (AI) represents a valuable tool that could be used to improve the safety of care. Major adverse events in healthcare include: healthcare-associated infections, adverse drug events, venous thromboembolism, surgical complications, pressure ulcers, falls, decompensation, and diagnostic errors. The objective of this scoping review was to summarize the relevant literature and evaluate the potential of AI to improve patient safety in these eight harm domains. A structured search was used to query MEDLINE for relevant articles. The scoping review identified studies that described the application of AI for prediction, prevention, or early detection of adverse events in each of the harm domains. The AI literature was narratively synthesized for each domain, and findings were considered in the context of incidence, cost, and preventability to make projections about the likelihood of AI improving safety. Three-hundred and ninety-two studies were included in the scoping review. The literature provided numerous examples of how AI has been applied within each of the eight harm domains using various techniques. The most common novel data were collected using different types of sensing technologies: vital sign monitoring, wearables, pressure sensors, and computer vision. There are significant opportunities to leverage AI and novel data sources to reduce the frequency of harm across all domains. We expect AI to have the greatest impact in areas where current strategies are not effective, and integration and complex analysis of novel, unstructured data are necessary to make accurate predictions; this applies specifically to adverse drug events, decompensation, and diagnostic errors.

A Conceptual Framework to Study the Implementation of Clinical Decision Support Systems (BEAR): Literature Review and Concept Mapping

Background

The implementation of clinical decision support systems (CDSSs) as an intervention to foster clinical practice change is affected by many factors. Key factors include those associated with behavioral change and those associated with technology acceptance. However, the literature regarding these subjects is fragmented and originates from two traditionally separate disciplines: implementation science and technology acceptance.

Objective

Our objective is to propose an integrated framework that bridges the gap between the behavioral change and technology acceptance aspects of the implementation of CDSSs.

Methods

We employed an iterative process to map constructs from four contributing frameworks—the Theoretical Domains Framework (TDF); the Consolidated Framework for Implementation Research (CFIR); the Human, Organization, and Technology-fit framework (HOT-fit); and the Unified Theory of Acceptance and Use of Technology (UTAUT)—and the findings of 10 literature reviews, identified through a systematic review of reviews approach.

Results

The resulting framework comprises 22 domains: agreement with the decision algorithm; attitudes; behavioral regulation; beliefs about capabilities; beliefs about consequences; contingencies; demographic characteristics; effort expectancy; emotions; environmental context and resources; goals; intentions; intervention characteristics; knowledge; memory, attention, and decision processes; patient–health professional relationship; patient’s preferences; performance expectancy; role and identity; skills, ability, and competence; social influences; and system quality. We demonstrate the use of the framework providing examples from two research projects.

Conclusions

We proposed BEAR (BEhavior and Acceptance fRamework), an integrated framework that bridges the gap between behavioral change and technology acceptance, thereby widening the view established by current models.

Risk of mortality and cardiopulmonary arrest in critical patients presenting to the emergency department using machine learning and natural language processing

Emergency department triage is the first point in time when a patient's acuity level is determined. The time to assign a priority at triage is short and it is vital to accurately stratify patients at this stage, since under-triage can lead to increased morbidity, mortality and costs. Our aim was to present a model that can assist healthcare professionals in triage decision making, namely in the stratification of patients through the risk prediction of a composite critical outcome-mortality and cardiopulmonary arrest. Our study cohort consisted of 235826 adult patients triaged at a Portuguese Emergency Department from 2012 to 2016. Patients were assigned to emergent, very urgent or urgent priorities of the Manchester Triage System (MTS). Demographics, clinical variables routinely collected at triage and the patients' chief complaint were used. Logistic regression, random forests and extreme gradient boosting were developed using all available variables. The term frequency-inverse document frequency (TF-IDF) natural language processing weighting factor was applied to vectorize the chief complaint. Stratified random sampling was used to split the data into train (70%) and test (30%) data sets. Ten-fold cross validation was performed in train to optimize model hyper-parameters. The performance obtained with the best model was compared against the reference model-a regularized logistic regression trained using only triage priorities. Extreme gradient boosting exhibited good calibration properties and yielded areas under the receiver operating characteristic and precision-recall curves of 0.96 (95% CI 0.95-0.97) and 0.31 (95% CI 0.26-0.36), respectively. The predictors ranked with higher importance by this model were the Glasgow coma score, the patients' age, pulse oximetry and arrival mode. Compared to the reference, the extreme gradient boosting model using clinical variables and the chief complaint presented higher recall for patients assigned MTS-3 and can identify those who are at risk of the composite outcome.

Machine Learning-Based Prediction of Korean Triage and Acuity Scale Level in Emergency Department Patients

Objectives

Triage is a process to accurately assess and classify symptoms to identify and provide rapid treatment to patients. The Korean Triage and Acuity Scale (KTAS) is used as a triage instrument in all emergency centers. The aim of this study was to train and compare machine learning models to predict KTAS levels.

Methods

This was a cross-sectional study using data from a single emergency department of a tertiary university hospital. Information collected during triage was used in the analysis. Logistic regression, random forest, and XGBoost were used to predict the KTAS level.

Results

The models with the highest area under the receiver operating characteristic curve (AUROC) were the random forest and XGBoost models trained on the entire dataset (AUROC = 0.922, 95% confidence interval 0.917-0.925 and AUROC = 0.922, 95% confidence interval 0.918-0.925, respectively). The AUROC of the models trained on the clinical data was higher than that of models trained on text data only, but the models trained on all variables had the highest AUROC among similar machine learning models.

Conclusions

Machine learning can robustly predict the KTAS level at triage, which may have many possibilities for use, and the addition of text data improves the predictive performance compared to that achieved by using structured data alone.

An effective density-based clustering and dynamic maintenance framework for evolving medical data streams

BACKGROUND

Medical data stream clustering has become an integral part of medical decision systems since it extracts highly-sensitive information from a tremendous flow of medical data. However, clustering and maintaining of medical data streams is still a challenging task. That is because the evolving of medical data streams imposes various challenges for clustering such as the ability to discover the arbitrary shape of a cluster, the ability to group data streams without a predefined number of clusters, and the ability to maintain the data clusters dynamically.

OBJECTIVE

To support the online medical decisions, there is a need to address the clustering challenges. Therefore, in this paper, we propose an effective density-based clustering and dynamic maintenance framework for grouping the patients with similar symptoms into meaningful clusters and monitoring the patients' status frequently.

METHODS

For clustering, we generate a set of initial medical data clusters based on the combination of Piece-wise Aggregate Approximation and the density-based spatial clustering of applications with noise called (PAA+DBSCAN) algorithm. For maintenance, when new medical data streams arrive, we maintain the initially generated medical data clusters dynamically. Since the incremental cluster maintenance is time-consuming, we further propose an Advanced Cluster Maintenance (ACM) approach to improve the performance of the dynamic cluster maintenance.

RESULTS

The experimental results on real-world medical datasets demonstrate the effectiveness and efficiency of our proposed approaches. The PAA+DBSCAN algorithm is more efficient and effective than the exact DBSCAN algorithm. Moreover, the ACM approach requires less running time in comparison with the Baseline Cluster Maintenance (BCM) approach using different tuning parameter values in all datasets. That is because the BCM approach tracks all the data points in the cluster.

CONCLUSION

The proposed framework is capable of clustering and maintaining the medical data streams effectively by means of grouping the patients who share similar symptoms and tracking the patients status that naturally tends to be changing over time.