Decision Support Research Articles

The effects of generative AI’s human-like competencies on clinical decision-making

ABSTRACT Generative AI (genAI) has revolutionized clinical AI systems by leveraging human language. Yet, challenges remain in its integration into clinical settings, particularly regarding the risk of physicians relying on hallucinated advice. We conducted an experimental study with 368 novice physicians who diagnosed patient cases while being augmented with clinical genAI systems. A theoretical model was empirically tested to examine how anthropomorphism and advice elaboration affect trust and cognitive load as mediators for appropriate reliance. Findings show that augmenting clinical decisions with genAI systems can improve physicians’ diagnostic accuracy but also frequently results in inappropriate reliance on hallucinated advice due to miscalibrated trust. Moreover, we emphasize the uncanny familiarity evoked by anthropomorphizing genAI systems, which diminishes trust while reducing cognitive load. Our findings highlight the benefits and ethical challenges of genAI in clinical decision support, underscoring the need to balance its advantages with safeguarding the integrity of physicians’ decision agency.

AET-net: A framework for subtype classification based on the multi-omics data of breast cancer

Breast cancer (BC) is one of the most prevalent cancers worldwide and remains a significant global public health challenge. The biomechanical characteristics of tumor microenvironments provide critical insights into cellular interactions and mechanical stress responses that potentially influence cancer progression. The integration and analysis of multi-omics data for BC subtype classification present substantial challenges, including high-dimensional data complexity and difficulties in integrating heterogeneous omics data characteristics. To address these challenges, we propose an Autoencoder and Transformer integrated neural network (AET-net) classification framework. The experimental results demonstrate that our model achieves significant performance improvements in predicting BC subtypes based on integrated multi-omics datasets, with an Accuracy of 0.912 and an AUC of 0.9862. These results not only validate the high accuracy of our model in BC subtype classification, providing a valuable tool for diagnostic decision support, but also demonstrate the potential of integrated multi-omics data analysis in enhancing the precision and efficiency of BC subtype identification.

Risk prediction of computer investment database information management system based on machine learning algorithms

In recent years, with the continuous development of the financial market, the risk prediction of computer investment database information management systems (IMS) has high practical value. At present, there are risk issues in the information management system, which may cause drawbacks to investment data processing. To address these issues, this article used Machine Learning (ML) algorithms to analyze the risk prediction of computer investment database IMS. This article introduced and utilized typical Self-Organizing Map (SOM) and Artificial Neural Network (ANN) combination algorithms, regression algorithms, and Gradient Boosting Decision Tree (GBDT) algorithms to compare and analyze the prediction accuracy of these three algorithms. This article found that the GBDT algorithm has the highest prediction accuracy. Through a large amount of experimental data, it has been proven that the average testing accuracy using regression algorithms was 3.5% higher than that using neural network algorithms. It was found that the average test accuracy using the GBDT algorithm was 7.2% higher than the average test accuracy using the regression algorithm. The study also explores the combination of physiological and behavioral data collected by wearable devices to provide more comprehensive risk assessment and decision support, which provides an important reference for the optimization of enterprise risk management. Through this innovative data source integration, this paper provides a new perspective for the application and development of machine learning algorithms in computer investment database IMS.

Perancangan Sistem Pendukung Keputusan Penilaian Kinerja Karyawan Terbaik pada PT XYZ menggunakan Metode TOPSIS

PT XYZ is a company in the field of construction, development, and management of construction projects. They are responsible for planning, managing, and implementing construction projects, both for infrastructure and commercial buildings. The previous assessment did not have a weighting of the criteria assessment and took time to process the value of each employee because the assessment still used a manual method (assisted by calculator calculations) so that the employee calculation process was processed one by one. This is not in accordance with the director's expectations and has an impact on the assessment being inaccurate and detrimental to the company in terms of time. Therefore, to evaluate the performance of top employees at PT XYZ, this decision support system application was created, with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method which can produce a simple Decision Support System with an easy-to-understand and logical concept in the PT XYZ Employee Performance Assessment.

Machine Learning Reveals Demographic Disparities in Palliative Care Timing Among Patients With Traumatic Brain Injury Receiving Neurosurgical Consultation.

Timely palliative care (PC) consultations offer demonstrable benefits for patients with traumatic brain injury (TBI), yet their implementation remains inconsistent. This study employs machine learning methods to identify distinct patient phenotypes and elucidate the primary drivers of PC consultation timing variability in TBI management, aiming to uncover disparities and inform more equitable care strategies. Data on admission, hospital course, and outcomes were collected for a cohort of 232 patients with TBI who received both PC consultations and neurosurgical consultations during the same hospitalization. Patient phenotypes were uncovered using principal component analysis and K-means clustering; time-to-PC consultation for each phenotype was subsequently compared by Kaplan-Meier analysis. An extreme gradient boosting model with Shapley Additive Explanations identified key factors influencing PC consultation timing. Three distinct patient clusters emerged: cluster A (n = 86), comprising older adult White women (median 87years) with mild TBI, received the earliest PC consultations (median 2.5days); cluster B (n = 108), older adult White men (median 81years) with mild TBI, experienced delayed PC consultations (median 5.0days); and cluster C (n = 38), middle-aged (median: 46.5years), severely injured, non-White patients, had the latest PC consultations (median 9.0days). The clusters did not differ by discharge disposition (p = 0.4) or inpatient mortality (p > 0.9); however, Kaplan-Meier analysis revealed a significant difference in time-to-PC consultation (p < 0.001), despite no differences in time-to-mortality (p = 0.18). Shapley Additive Explanations analysis of the extreme gradient boosting model identified age, sex, and race as the most influential drivers of PC consultation timing. This study unveils crucial disparities in PC consultation timing for patients with TBI, primarily driven by demographic factors rather than clinical presentation or injury characteristics. The identification of distinct patient phenotypes and quantification of factors influencing PC consultation timing provide a foundation for developing for standardized protocols and decision support tools to ensure timely and equitable palliative care access for patients with TBI.

Development of individualized risk assessment models for predicting post-traumatic epilepsy 1 and 2 years after moderate-to-severe traumatic brain injury: A traumatic brain injury model system study.

Although traumatic brain injury (TBI) and posttraumatic epilepsy (PTE) are common, there are no prospective models quantifying individual epilepsy risk after moderate-to-severe TBI (msTBI). We generated parsimonious prediction models to quantify individual epilepsy risk between acute inpatient rehabilitation for individuals 2 years after msTBI. We used data from 6089 prospectively enrolled participants (≥16 years) in the TBI Model Systems National Database. Of these, 4126 individuals had complete seizure data collected over a 2-year period post-injury. We performed a case-complete analysis to generate multiple prediction models using least absolute shrinkage and selection operator logistic regression. Baseline predictors were used to assess 2-year seizure risk (Model 1). Then a 2-year seizure risk was assessed excluding the acute care variables (Model 2). In addition, we generated prognostic models predicting new/recurrent seizures during Year 2 post-msTBI (Model 3) and predicting new seizures only during Year 2 (Model 4). We assessed model sensitivity when keeping specificity ≥.60, area under the receiver-operating characteristic curve (AUROC), and AUROC model performance through 5-fold cross-validation (CV). Model 1 (73.8% men, 44.1 ± 19.7 years, 76.1% moderate TBI) had a model sensitivity = 76.00% and average AUROC = .73 ± .02 in 5-fold CV. Model 2 had a model sensitivity = 72.16% and average AUROC = .70 ± .02 in 5-fold CV. Model 3 had a sensitivity = 86.63% and average AUROC = .84 ± .03 in 5-fold CV. Model 4 had a sensitivity = 73.68% and average AUROC = .67 ± .03 in 5-fold CV. Cranial surgeries, acute care seizures, intracranial fragments, and traumatic hemorrhages were consistent predictors across all models. Demographic and mental health variables contributed to some models. Simulated, clinical examples model individual PTE predictions. Using information available, acute-care, and year-1 post-injury data, parsimonious quantitative epilepsy prediction models following msTBI may facilitate timely evidence-based PTE prognostication within a 2-year period. We developed interactive web-based tools for testing prediction model external validity among independent cohorts. Individualized PTE risk may inform clinical trial development/design and clinical decision support tools for this population.

A fingerprinting structure model for Arabic document plagiarism detection

Plagiarism, which is a significant problem in the academic world worldwide, is particularly challenging to detect in Arabic due to the language's complex structure. Methodology: The ADPDM model and framework for detecting plagiarism in Arabic documents are presented in this dissertation. It is designed to detect plagiarism within academic contexts. By organizing documents logically into paragraphs, sentences, and words, the model seeks to establish a robust system that can identify duplicated content and search for similar documents within identifying corresponding sets. In particular, the study examines preprocessing techniques such as stop word removal, stemming, and rootage processing, followed by content-based methods utilizing fingerprinting and heuristic algorithms that are tailored to Arabic language features. To aid in efficient detection, the BKDR hash function is used for chunk having. To optimize computation time, heuristic algorithms are implemented at different levels of document representation, using metrics such as Longest Common Substring (LCS). To evaluate the ADPDM system, a corpus of 100 documents is utilized, which includes datasets from AraPlagDet and the Decision Support System (DSS). The performance of ADPDM is compared with other plagiarism detection methods using WCopyFind, but the latter has a higher computational speed than ADBDM. Its recall, precision, and F-measure values of 0. 78035, 0. 994264, and 0. 865688 respectively (ADPDM) are particularly notable for its ability to detect plagiarized content in Arabic documents. ADPDM is a successful anti-pluralist solution for Arabic text, even though it requires a longer processing time than WCopyFind.

How should scientists act? Assessing public perceptions of scientists and scientific practices and their implications for science communication

Is how science is conducted legitimate? Are scientists trustworthy? Whether the public answers “yes” to these questions is critical for science communicators. We explore how social factors affect public perceptions of the practice of science, and then test how those beliefs relate to views about how scientists engage with the public and policy making. Our results show that political ideology and religiosity affect these views. However, more importantly, respondents' concerns about the integrity of the scientific process are the strongest predictor of views about scientists' behavior, providing a focus area for future communication efforts in support of science-based decision making.

Enhancing brain tumor classification by integrating radiomics and deep learning features: A comprehensive study utilizing ensemble methods on MRI scans

Background and objective This study aims to assess the effectiveness of combining radiomics features (RFs) with deep learning features (DFs) for classifying brain tumors—specifically Glioma, Meningioma, and Pituitary Tumor—using MRI scans and advanced ensemble learning techniques. Methods A total of 3064 T1-weighted contrast-enhanced brain MRI scans were analyzed. RFs were extracted using Pyradiomics, while DFs were obtained from a 3D convolutional neural network (CNN). These features were used both individually and together to train a range of machine learning models, including Support Vector Machines (SVM), Decision Trees (DT), Random Forests (RF), AdaBoost, Bagging, k-Nearest Neighbors (KNN), and Multi-Layer Perceptrons (MLP). To enhance the accuracy of these models, ensemble approaches such as Stacking, Voting, and Boosting were employed. LASSO feature selection and five-fold cross-validation were utilized to ensure the models’ robustness. Results The results demonstrated that combining RFs and DFs significantly improved the model's performance compared to using either feature set alone. The best performance was achieved using the combined RF + DF approach with ensemble methods, particularly Boosting, which resulted in an accuracy of 95.0%, an AUC of 0.92, a sensitivity of 88%, and a specificity of 90%. Conversely, models utilizing only RFs or DFs showed lower performance, with RFs reaching an AUC of 0.82 and DFs achieving an AUC of 0.85. Conclusion The integration of RFs and DFs, along with advanced ensemble methods, significantly improves the accuracy and reliability of brain tumor classification using MRI. This approach shows strong clinical potential, with opportunities for further enhancing generalizability and precision through additional MRI sequences and advanced machine learning techniques.

Scoping Review of Naturalistic Decision Making Studies Among Mental Health Professionals: Coverage of Characteristics and Contexts

The Naturalistic Decision Making (NDM) paradigm is an emerging shift in how researchers study decision making in complex, real-world situations and design decision supports. The purpose of this scoping review was to describe how the NDM paradigm was applied in studies of mental health professionals. Six bibliographic databases were searched to identify NDM studies. Each study was charted for study features, participant demographics, decision contexts, and the essential characteristics of NDM research. The search identified 26 studies published from 1989 to June 2023. Approximately 35% of studies were published in a peer-reviewed journal. Quantitative (30.8%), qualitative (34.6%), and mixed (34.6%) methods were utilized in similar percentages of studies, with social workers (61.5%) most frequently represented in these studies. Approximately 69% of studies examined assessment decisions (versus diagnosis or treatment) and roughly 96% of studies examined individuals (versus teams). Most studies explored professionals’ decision making process (73.1%) and how proficient decision makers utilized their experience to make decisions (38.5%). The NDM literature among mental health professionals is growing, with many opportunities to understand clinical decision making using well-established NDM concepts and methods. The review concludes with recommendations for both NDM and mental health services researchers.

Factors Contributing to Uptake of Stillbirth Evaluations: A Qualitative Analysis.

The purpose of this study was to explore individuals' beliefs, values, and experiences surrounding stillbirth evaluation decisions. Qualitative research. University of Utah Health. Parents who experienced a stillbirth in the past 5 years (n = 19) were interviewed about their experiences and decision to consent to or decline stillbirth evaluations, such as autopsy, placental histopathology, or genetic testing. Qualitative content analysis. Barriers and facilitators to parentsstillbirth postmortem decision-making. Participants communicated several facilitators and barriers that contributed to their stillbirth evaluation decision. Reasons for consenting to evaluations were belief in science, background in medicine, altruism, to inform future pregnancies, thinking about preventing another stillbirth, and how patients viewed the care of their stillborn by the medical team. Reasons for declining evaluations were receiving a diagnosis prior to being offered a postmortem evaluation, intent to avoid causing further harm to the baby, interest to spend more time with their baby, and cost of the evaluation. Stillbirth is one of the most difficult experiences of a parent. Diagnostic and emotional barriers create further challenges to decision-making for stillbirth postmortem evaluations. Parents often rely on inadequate information and personal values and beliefs during this time-sensitive decision process. Decision support for stillbirth evaluations and training for medical providers could benefit parents, may increase stillbirth evaluation uptake, and potentially prevent decisional regret. Research reported in this publication was supported by the Utah Centre for Excellence in ELSI Research (UCEER).

Polish Firefighters' Participation in Interventions Related to Behavioral Disorders in the 2020-2022 Period: An Observation of Incidents.

Rescue service interventions for patients with behavioral disorders are quite common. The aim of this study is to analyze interventions of Polish State Fire Service units in incidents involving health threats to victims with behavioral disorders. This study used data from the Decision Support System of the State Fire Service. Events in the period of 1 January 2020 at 00.00-31 December 2022 at 23:59 were used for analysis. In the 2020-2022 period, firefighters were dispatched 321 times to interventions concerning injured individuals with behavioral disorders. Isolated emergency medical incidents most often concerned mental disorders (23%) and least often concerned suicide attempts (8%) (p = 0.033). It was shown that the duration of the intervention (emergency care time) was significantly longer in the case of interventions related to the autumn period (p < 0.001). Our results describe parameters associated with behavioral health responses within the Polish fire service.

Generative AI in Medicine and Healthcare: Moving Beyond the ‘Peak of Inflated Expectations’

The rapid development of specific-purpose Large Language Models (LLMs), such as Med-PaLM, MEDITRON-70B, and Med-Gemini, has significantly impacted healthcare, offering unprecedented capabilities in clinical decision support, diagnostics, and personalized health monitoring. This paper reviews the advancements in medicine-specific LLMs, the integration of Retrieval-Augmented Generation (RAG) and prompt engineering, and their applications in improving diagnostic accuracy and educational utility. Despite the potential, these technologies present challenges, including bias, hallucinations, and the need for robust safety protocols. The paper also discusses the regulatory and ethical considerations necessary for integrating these models into mainstream healthcare. By examining current studies and developments, this paper aims to provide a comprehensive overview of the state of LLMs in medicine and highlight the future directions for research and application. The study concludes that while LLMs hold immense potential, their safe and effective integration into clinical practice requires rigorous testing, ongoing evaluation, and continuous collaboration among stakeholders.

A Declarative Modeling Framework for Intuitive Multiple Criteria Decision Analysis in a Visual Semantic Urban Planning Environment

The magnitude and multiformity of recorded and readily available urban data from independent sensors and mobile user devices have paved new possibilities for city environment planning and evolution. In previous works, a visual semantic decision support system for urban planning was presented, and the capability of machine learning approaches, ranging from random forests to graph-based convolutional neural networks, to infer preferable directions for future development was explored, extrapolating upon previous opted locations and selected alternatives for publicly commercial services. In this work, the anterior decision-making process leading to establishment choices is addressed with the same sets of criteria and samples within the same environment. A Declarative Modeling shell is proposed, and Multiple Criteria Decision Analysis processes are adopted to encompass the DM’s/DMs’ rationale, solely relying on methodologies close to human intuition. To this end, outranking representatives from the PROMETHEE family as well as weighted sum approaches are employed, fueled by the interpretation of the declarative description of decision parameters on behalf of the DM(s), exploring the ability to achieve classifications in a straight synthetic manner, i.e., in the absence of previous knowledge, thus exhibiting the potential of decision analysis methodologies, enhanced by Declarative Modeling, to be used as powerful intuitive tools in similar paradigm contexts.

Classification and prediction of heart disease using Machine Learning models: A promising approach for medical diagnosis

Background: heart disease is one of the leading causes of death worldwide, claiming 17.9 million lives. They are a major public health problem that affects people regardless of age or gender. Objective: This work aims to classify and predict heart disease using Machine Learning (ML) models such as Random Forest (RF), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Naive Bayes (NB), Decision Tree (DT) and Logistic Regression (LR). Methods: We worked with the Cleveland dataset from Kaggle, consisting of 303 patient records and 14 attributes. This research was conducted in different stages, including model understanding, dataset analysis and cleaning, ML model training, and model performance evaluation. Results: The results showed that the RF and KNN models achieved the highest levels of performance and accuracy with 88.52%, surpassing the other models such as SVM, NB, and LR which obtained 86.89% accuracy, and DT with 78.69%. Conclusions: In conclusion, the RF and KNN models stand out over the other models for this type of prediction task.

LQ45 Stock Recommendations with Fundamental Analysis Using the Simple Additive Weighting Method

Shares are a sign of ownership of a person or entity in a company or limited liability company. Shares are in the form of a piece of paper that explains that the paper's owner is the owner of the company that issued the securities. In the Indonesia Stock Exchange (IDX) there are several kinds of stock indices, one of which is LQ45. It is known that until 2020 there were 2.48 million investors in the capital market. However, it should be noted that of the large number of investors, 85% to 90% of investors experience failure or loss. This happens because investors do not have adequate skills and knowledge in investing in stocks. Therefore, a system is needed that can assist in determining the selection of LQ45 stocks, namely a decision support system (DSS). This system was built web-based using the Simple Additive Weighting (SAW) method to determine the order of the most recommended stocks in the garden. This study uses 45 stocks as alternative data and 5 selection criteria, namely Return on Assets (ROA), Price to price-to-earnings ratio (PER), Price to Book Value (PBV), Net Profit Margin (NPM), and Debt to Equity Ratio (DER). The implementation of the SAW method on the system produces a sequence of LQ45 stock recommendations, namely Adaro Energy Indonesia as the best alternative with the highest preference value of 0.925.

Analyzing and Designing Decision Support Systems for Stroke Patient Daily Treatment

Stroke is one of the neurological diseases that often requires complex management and treatment. Making the right decisions in diagnosing and treating stroke patients can have a significant impact on the recovery process and the quality of life for patients. To assist healthcare professionals and patients in making more informed decisions, this research has developed the "StrokeCare Navigator," a web-based Decision Support System (DSS) that utilizes the Decision Tree C4.5 algorithm. This system allows users, including healthcare professionals and patients, to input symptoms and patient medical information, providing an initial diagnosis of the type and severity of the stroke. Additionally, the system offers personalized treatment guidelines, treatment recommendations, and online patient monitoring services. The Decision Tree C4.5 model used in this system was developed through training with relevant patient data. Evaluation results indicate that the system can provide accurate and beneficial recommendations for healthcare professionals and patients. Testing and patient monitoring in clinical practice also demonstrate the potential to enhance stroke patient care. Therefore, StrokeCare Navigator is expected to be a valuable tool in stroke patient management and treatment, ultimately improving the prospects of patient recovery and providing better guidance for healthcare professionals. This research also provides deeper insights into the implementation of the Decision Tree C4.5 algorithm in the context of Decision Support Systems (DSS) in the medical field.

Fatal fall from a height: is it possible to apply artificial intelligence techniques for height estimation?

Fall from a height trauma is characterized by a multiplicity of injuries, related to multiple factors. The height of the fall is the factor that most influences the kinetic energy of the body and appears to be one of the factors that most affects the extent of injury. The purpose of this work is to evaluate, through machine learning algorithms, whether the autopsy injury pattern can be useful in estimating fall height. 455 victims of falls from a height which underwent a complete autopsy were retrospectively analyzed. The cases were enlisted by dividing them into 7 groups according to the height of the fall: 6 or less meters; 9m, 12m, 15m, 18m, 21m, 24m or more. Autoptic data were registered through the use of a previously published visceral and skeletal table. A total of 25 descriptors were used. Reduction of values in the range, standard and robust scaling were used as preprocessing methods. Principal Component Analysis, Single Value Decomposition and Independent Component Analysis were applied for dimensionality reduction. Cross validation was performed with 5 internal and external folds to ensure the validity of the results. The learning algorithms that generated the best models were Linear Regression, Support Vector Regressor, Kernel Ridge, Decision trees and Random forests. The best mean absolute error was 4.58 ± 1.28m when dimensionality reduction was applied. Without any dimensionality reduction, the best result was 4.37 ± 1.27m, suggesting a good performance of the proposed algorithms, with better performance when dimensionality is not automatically reduced.

Pemilihan Wisudawan Terbaik Menggunakan Sistem Pendukung Keputusan dengan Metode Profile Matching

The selection of the best graduate is an important process in the graduation ceremony to recognize students who have demonstrated outstanding academic and non-academic achievements during their studies. However, this process often faces challenges, particularly regarding the subjectivity and complexity of assessment. Therefore, a Decision Support System (DSS) is needed to assist in determining the best graduate objectively and transparently. The criteria used include academic achievements (GPA), involvement in organizational activities, non-academic achievements, attendance, and social contributions. Each criterion is weighted according to its level of importance. With the implementation of this DSS, it is hoped that the selection of the best graduate can be conducted more effectively, efficiently, and fairly. Testing results show that this system can provide accurate results that meet stakeholder expectations.

Enhancing clinical decision support with explainable deep learning framework for C-section forecasting

Enhancing clinical decision support with explainable deep learning framework for C-section forecasting