BIOETHICARE 360O: A Deliberative Software for Ethical Computational Support in Complex Clinical Decision-Making

BIOETHICARE 360O: A Deliberative Software for Ethical Computational Support in Complex Clinical Decision-Making | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article BIOETHICARE 360 O : A Deliberative Software for Ethical Computational Support in Complex Clinical Decision-Making Anderson Díaz Pérez, Joseph Javier Sánchez Acuña This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7078293/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Clinical decision-making at the end of life often involves complex ethical dilemmas that require balancing core principles such as patient autonomy, beneficence, justice, and non-maleficence. Although various clinical decision-support tools exist, few incorporate bioethical deliberation, argument traceability, and multiperspective participation. To address this gap, BIOETHICARE 360° was developed as a computational system designed to facilitate ethical analysis in terminal clinical cases. This study aimed to evaluate the software’s ability to identify ethical dilemmas, propose context-sensitive recommendations, and support real-time deliberative processes. Results. The system was applied to thirteen synthetic clinical cases representing common scenarios in palliative care and morally complex situations. BIOETHICARE 360° enabled the structured registration of clinical, sociocultural, and ethical information from the perspectives of the medical team, the patient's family, and the bioethics committee. Through its functional modules, the system performed ethical principle weighting, generated graphical visualizations of dissent, identified potential algorithmic biases, and activated narrative-based deliberations through a conversational assistant. In five of the thirteen cases, the primary dilemma identified involved conflicts between the patient’s advance directives and the family’s intention to pursue invasive interventions. In 77% of cases, the final recommendation was the early integration of palliative care. The system provided reasoned responses that included legal analysis, proposals for empathetic communication, and the automated generation of informed consent documents, demonstrating both technical consistency and ethical sensitivity. Conclusions. BIOETHICARE 360° represents a methodological and technological advancement in ethical support for clinical practice. Its modular architecture, ability to integrate multiple stakeholder perspectives, and deliberative engine grounded in bioethical principles enhance the transparency, traceability, and quality of decision-making in critical clinical settings. Beyond its clinical utility, the tool holds significant educational value by facilitating reflective processes within interdisciplinary teams. This study suggests that the use of artificial intelligence in medical ethics may offer a viable and trustworthy strategy for strengthening decision-making at the end of life. Ethics Clinical Decision Support Systems Clinical Artificial Intelligence Palliative Care Terminal Care Patient Autonomy Figures Figure 1 Introduction Artificial intelligence (AI) has emerged as a transformative technology in clinical settings, particularly in ethically complex scenarios such as palliative care and end-of-life decision-making. Machine learning and deep learning models enable the real-time analysis of large volumes of clinical data to generate diagnostic or therapeutic recommendations (1). However, their application in sensitive contexts introduces significant bioethical risks related to patient autonomy, distributive justice, algorithmic transparency, and professional accountability (2,3). Key challenges include algorithmic bias stemming from non-representative training data, heuristic shortcuts embedded in models to optimize statistical performance, and skewed predictions when applying generalized logic to individualized clinical contexts (4–6). These limitations are particularly critical in terminally ill patients, where personalization, ethical deliberation, and respect for advance directives are clinical imperatives. In response, international organizations such as UNESCO, the OECD, and the European Commission have promoted ethical principles for the responsible development and deployment of AI, including autonomy, harm prevention, justice, explainability, technical robustness, and human oversight (7,8). Nevertheless, most AI-based clinical solutions still lack mechanisms to evaluate, visualize, or document the application of these principles throughout the clinical decision-making process. To address this gap, we developed BIOETHICARE 360°, a computational application implemented in Python and deployed via Streamlit, designed to enable automated and deliberative analysis of complex bioethical dilemmas. The system is composed of several functional modules: Preliminary analysis using generative AI Employs advanced language models (Bioethics Assistant powered by Gemini) to analyze clinical narratives and extract both explicit and implicit bioethical conflicts. Structured case registration Enables the collection of clinical, cultural, and ethical data from multiple perspectives: the medical team, the patient/family, and the bioethics committee. Multiperspective ethical weighting Assigns scores (0 to 5) to the principles of autonomy, beneficence, non-maleficence, and justice for each stakeholder. Algorithmic audit and ethical verification The system detects imbalances, omissions, and disproportionate ethical considerations, providing automated warnings and ethical suggestions. Advanced graphical visualization Generates radar and bar charts to represent consensus and dissent, enhancing visual traceability and interdisciplinary deliberation. Bioethical deliberative engine Based on Diego Gracia’s model, it articulates facts, values, and potential courses of action, integrating ethical principles, national legal frameworks, and the patient’s sociocultural context. Structured report generation Produces a comprehensive PDF report including all perspectives, visualizations, ethical deliberation, and recommendations for the clinical team. This article presents the software’s architecture, deliberative methodology, and its empirical application to a real clinical case involving a terminally ill patient, where an algorithmic recommendation for palliative care conflicted with the family’s religiously motivated desire to continue life-sustaining interventions. The tool enables real-time identification and mitigation of ethical bias, fosters interprofessional deliberation, and documents decisions within a transparent and ethically justifiable framework. Implementation The BIOETHICARE 360° software was developed with the purpose of providing a technological platform that integrates ethical deliberation, clinical reasoning, and automated recommendations in the care of terminally ill patients. Its modular architecture and ethics-by-design approach ensure full traceability of each computational decision, guaranteeing transparency, explainability, and normative support. The system was implemented in Python 3.10, deployed via Streamlit, and supported by a Firebase Firestore database, making it fully operational in both local and cloud environments. General Architecture and Design Logic The architecture of BIOETHICARE 360° consists of four main modules, as illustrated in Figure 1. The core of the system is built around a main class called BioethicalCase , in which each attribute represents an essential dimension of clinical and bioethical analysis: patient name, age, condition, medical history, ethical dilemma, sociocultural background, and multiperspective ethical scoring (physician, family, ethics committee). Table 1 below illustrates the association between each evaluated dimension, its expected clinical or ethical output, and the corresponding function or method executed within the code: Table 1. Association between bioethical dimensions, clinical/ethical outputs, and code functions Evaluated Dimension Clinical/Ethical Recommendation Function or Attribute in the Code Model Name Bioethics Care Decision Model __init__() Age / Gender / Condition / Medical History Registration of terminal patient and functional context age, gender, condition, medical_history Multiperspective Ethical Weighting Assignment of scores (0 to 5) by principle and stakeholder perspectives = {...} Ethical Bias Evaluation Detection of imbalances, omissions, and dominance verify_ethics_bias() Bioethical Visualization Radar chart of principles and consensus/dissent graph generate_advanced_visualizations() Comprehensive Patient Assessment Determines prioritization for palliative care comprehensive_evaluation() Ethical Conflicts Detects tensions such as "autonomy vs. beneficence" analyze_bioethical_conflicts() Narrative Ethical Reflection Qualitative analysis of each principle by perspective reflect_on_bioethical_principles() Computational Recommendation Integration of analyses into a reasoned ethical decision evaluate_recommendation_type() Structured Deliberation Implements Diego Gracia and Anderson Díaz’s methodology analyze_methodologies() Legal and Deontological Evaluation Mapping of potential sanctions under Colombian law analyze_legal_aspects() Informed Consent Automated generation with regulatory references generate_informed_consent() Ethical Algorithmic Evaluation and Bias Detection In addition to its deliberative analysis capabilities, the system performs an ethical algorithmic audit aimed at identifying structural biases, heuristic simplifications, and ethical risks embedded in the automated reasoning process (Table 2). Table 2. Integration of Computational Logic, Ethical Models, and Mitigation Strategies Analyzed Dimension Applied Algorithmic Logic Applied Ethical Model Detected Bias Mitigation Strategy Comprehensive Clinical Assessment if statements based on functional/cognitive thresholds Principlism (Beauchamp & Childress) Rigid thresholds fail to capture clinical nuances Adjust thresholds using variable combinations + expert clinical judgment Scaled Bioethical Weighting Dictionary with Likert scale (0–5) Principlism Ethical dimensions reduced to quantifiable metrics Post-validation through narrative reflection Qualitative Ethical Reflection if-elif-else trees on ethical values Narrative / Clinical Ethics Availability heuristic (overweighting salient principles) Require balanced representation of all four principles across stakeholders Conflicts Between Principles Symbolic Boolean logic evaluation Bioethical Casuistry Binary simplification of complex dilemmas Inclusion of compound tensions and cross-deliberation mechanisms Bioethical Deliberation Structured text + nested JSON functions Diego Gracia + Anderson Díaz Cognitive overload from textual complexity Guided templates combined with graphical analysis Computational Recommendation Conditional rules based on dilemma typology Principlist-based Utilitarian Ethics Dominance of consequentialist reasoning Cross-validation with patient autonomy and cultural context SPIKES (Breaking Bad News) Structured dictionary Ethics of Compassion / Clinical Protocols Rigidity in religious or emotionally complex situations Empathic personalization + cultural and psychological tailoring Legal Aspects Dictionary by legal category Legal Bioethics (Colombia) Risk of international legal extrapolation Geo-adaptive legal references based on local jurisdiction Informed Consent Automated templates Procedural / Legal Ethics Generic or inappropriate consent generation Customization by cognitive competence, age, and advance directives This integrative approach represents not only a functional computational tool but also a methodological advancement that systematizes clinical bioethical deliberation through a structured, auditable, and traceable framework, with the potential for replication in diverse clinical contexts. Results Application of the System to Ethically Complex Clinical Cases BIOETHICARE 360° was applied to thirteen representative clinical-bioethical cases, encompassing common end-of-life dilemmas. The tool processed both structured and unstructured information, enabling different stakeholders (medical team, family, and ethics committee) to weigh core principles such as autonomy, beneficence, non-maleficence, and justice. Additionally, the system incorporated ethical deliberation via a conversational AI module ( Bioethics Assistant powered by Gemini ), which contributed narrative analyses, context-sensitive recommendations, and empathetic computational language. To empirically validate the functionality of BIOETHICARE 360°, the system was applied to a real clinical case involving a 68-year-old male patient with terminal-stage pancreatic adenocarcinoma, functional decline, and a life expectancy of less than three months. The patient was conscious, oriented, and had previously recorded advance directives requesting limitation of therapeutic efforts and initiation of comprehensive palliative care. However, his family insisted on continuing life-sustaining treatments for religious and emotional reasons, creating an ethical conflict between the patient's autonomy and the family’s perceived beneficence. I. Representative Case Study An emblematic case involved a 68-year-old male patient with metastatic pancreatic cancer who had explicitly stated advance directives declining invasive life-sustaining measures. The family, however, insisted on maintaining full support. The system identified the main ethical dilemma as a conflict between autonomy and beneficence. The computational analysis revealed a significant discrepancy in ethical perspectives among stakeholders (a difference greater than 6 points) and a notable omission of the autonomy principle by the family. The automated ethical audit triggered alerts and consistently recommended transitioning to palliative care, including compassionate communication guided by the SPIKES protocol. Structured Case Registration Clinical and sociocultural data were entered into the system interface, including: Functional status: 2/5 Cognitive status: 2/5 Perceived quality of life: 1/5 Level of suffering: 4/5 The clinical history was first analyzed using generative AI (Gemini), which identified the main dilemma as: “tension between the patient’s advance directives and the family’s desire to continue invasive treatment.” Multiperspective Ethical Weighting Ethical scores were assigned to the four foundational bioethical principles by each stakeholder: Medical team: Autonomy (5), Beneficence (3), Non-maleficence (4), Justice (4) Family: Autonomy (2), Beneficence (5), Non-maleficence (2), Justice (3) Ethics committee: Autonomy (4), Beneficence (3), Non-maleficence (3), Justice (4) These scores were automatically visualized through radar and bar charts (see Figures 2 and 3), facilitating dissent analysis and highlighting the greater imbalance in the family’s perspective. Ethical Audit and Detected Biases The automated ethical verification module identified: Omission of the autonomy principle in the family’s perspective. Inter-actor imbalance: a 6-point gap between total scores of the medical team and the family. Affective dominance warning: high beneficence scoring without proportional evaluation of suffering. The system advised reviewing the weightings and promoting a guided joint deliberation. Assisted Deliberation and Computational Recommendation BIOETHICARE 360° applied two ethical deliberation models: Anderson Díaz Model (MIEC) (9) → Determined that the patient required priority access to palliative care. AI recommended discontinuing futile treatments and reinforcing psychoemotional support. Diego Gracia Model (10–12) → Proposed a deliberation process balancing clinical facts, conflicting values, and normative principles. The suggested decision was to respect the patient's advance directives, communicate the situation compassionately to the family, and initiate the SPIKES protocol. Structured Report and Automated Consent Generation The system generated a comprehensive PDF report integrating: Detailed case description Comparative ethical perspectives Visualization of consensus and dissent Narrative deliberative analysis Final ethical recommendation: “Immediate transition to comprehensive palliative care” In addition, an automated ethical-legal informed consent form was generated, customized with patient-specific data and incorporating Colombian regulatory references. Clinical and Educational Impact The use of the system enabled: Ethical traceability of the clinical decision Mitigation of biases related to emotional pressure and therapeutic inertia Real-time ethical training for the medical team Improved interprofessional and family communication II. Comparative Analysis of Five Cases To assess the ethical consistency, computational adaptability, and clinical sensitivity of the BIOETHICARE 360° software, five representative clinical cases were selected, each featuring distinct dilemmas varying in nature and complexity. These cases served as test scenarios to evaluate the system’s behavior in resolving real-world ethical conflicts in end-of-life care. Each case was processed using the system’s structured logic, which includes: Multiperspective capture of bioethical principles (autonomy, beneficence, non-maleficence, and justice) Execution of specific functions depending on the dilemma type (e.g., evaluate_autonomy(), analyze_bioethical_conflicts(), or deliver_bad_news_technique()) Empathetic and contextual deliberation via the Bioethics Assistant (Gemini), functioning as a digital mediator capable of analyzing, reformulating, and suggesting narrative-based actions Table 4 presents a detailed synthesis of the five cases, incorporating clinical, ethical, computational, and argumentative elements. In all scenarios, the software successfully identified the underlying ethical dilemma, assigned dominant principles per stakeholder, detected specific moral tensions, and produced technically and ethically consistent recommendations. For instance, in Case 1, the conflict between the patient's refusal of intubation and a clinical tendency toward beneficence was resolved through a binary decision tree. The system prioritized the patient's autonomy and recommended palliative sedation in accordance with his advance directives. Gemini reinforced this decision with a narrative argument validating the principle of self-determination (Table 4). By contrast, Case 2 involved a demand for resuscitation by the family despite the patient’s terminal condition. Here, the system employed ethical weighting and conditional logic to expose a conflict between distributive justice and non-maleficence. The recommendation was to suspend interventions, promote psychosocial support, and activate the SPIKES protocol (Table 4). In Case 3, which lacked any recorded advance directives, BIOETHICARE 360° activated its modules for informed consent generation and institutional analysis. AI proposed a collaborative deliberation between the care team and ethics committee, suggesting a decision grounded in justice and normative equity (Table 4). Case 4, involving a young patient with irreversible neurological damage, benefited from AI-assisted deliberation combining narrative ethics, quality-of-life analysis, and multiperspective ethical visualization (Table 4). The recommendation was a gradual limitation of therapeutic efforts, emphasizing compassion as the guiding principle. Finally, in Case 5, the system identified a conflict between the patient’s autonomy and opposition from the bioethics committee. Through computational empathetic evaluation, ethical classification, and automated SPIKES activation, the tool proposed palliative care with informed consent, respecting the advance directives (Table 4). This comparative analysis highlights not only the system’s logical robustness but also its adaptive capacity to ethically deliberate using contextual data, legal frameworks, and emotional dimensions of care. The system’s architecture allowed each case to be processed with precision and sensitivity, integrating specialized algorithmic components and high-quality AI-generated arguments. The integration of the Gemini assistant was key to enriching ethical reasoning with narrative language, empathetic explanations, and clinical communication strategies. This is especially valuable in emotionally charged contexts, where decisions must be not only technically sound but also humanely and ethically justified. Together, these five cases validate the system’s potential as a clinical-bioethical decision support tool, offering algorithmic transparency, personalized recommendations, and respect for human rights in healthcare (Table 4). Table 4. Extended Ethical-Algorithmic Comparative Analysis of 5 Representative Cases Case Central Ethical Dilemma Most Emphasized Principle Primary Conflict Detected Final Ethical Recommendation Software Functions Involved Gemini Assistant (Conversational AI) Intervention Type of Computational Analysis Applied Ethical Model 1 Patient refusal of intubation Autonomy (physician) Autonomy vs. Beneficence Respect patient's decision + initiate palliative sedation evaluate_autonomy(), analyze_bioethical_conflicts(), reflect_on_bioethical_principles() AI-generated recommendations based on natural language input dilemma Binary decision tree + conditional logic Principlism + Personalist Bioethics 2 Family demands resuscitation in terminal phase Beneficence (family) Justice vs. Non-Maleficence Suspend interventions + provide psychosocial support evaluate_justice(), evaluate_non_maleficence(), analyze_methodologies()[0] AI-mediated deliberation addressing emotional tension and medical viability Weighted evaluation + conditional logic + inter-actor conflict detection Clinical Bioethics + Distributive Justice 3 Patient without advance directives Justice (ethics committee) Autonomy vs. Legal Uncertainty Interdisciplinary evaluation + ethics committee deliberation evaluate_interdisciplinary_team_role(), generate_informed_consent() Collaborative and reflective proposal addressing regulatory gaps through narrative reasoning Rule-based structured evaluation + legal consent generation Classical Deliberative Model + Institutional Ethics 4 Young patient with irreversible brain damage Non-Maleficence (physician) Family Autonomy vs. Prognostic Reality Gradual limitation of therapeutic effort evaluate_non_maleficence(), analyze_methodologies()[1], reflect_on_bioethical_principles() AI-assisted ethical-emotional analysis based on quality of life and prognosis Deliberative analysis + narrative comparison + multiperspective graphical weighting Narrative Ethics + Anderson Díaz Model 5 Futile intervention in terminally ill patient Autonomy (patient) Family vs. Bioethics Committee Initiate palliative care with informed consent evaluate_palliative_care(), analyze_bioethical_conflicts(), deliver_bad_news_technique() Computational ethical validation + AI-guided empathetic communication to family and committee Empathetic computational assessment + ethical classification + automated SPIKES Principlism + Palliative Bioethics + Ethics of Compassion III. Identified Ethical Trends and Patterns The global analysis of the 13 clinical cases processed by BIOETHICARE 360° revealed a series of recurring patterns that reflect both the ethical behavior of stakeholders and the system’s internal logical consistency. The most frequent dilemma identified was the conflict between patient autonomy and family beneficence, which appeared in 5 of the 13 cases. This finding confirms that end-of-life decisions commonly involve tensions between respecting the patient’s will and the family’s desire to prolong life for emotional, hopeful, or informational reasons. The system identified this dilemma using semantic analysis functions (analyze_bioethical_conflicts()) and clinical-ethical decision trees. Regarding the prioritization of ethical principles, physicians tended to emphasize autonomy (average 4.3/5), while families placed greater emphasis on beneficence (average 4.6/5). This ethical discrepancy was visualized through the system’s multiperspective weighting maps, enabling early recognition of potential value conflicts between stakeholders. Moreover, in 77% of cases (10 out of 13), the final recommendation was the early integration of palliative care, indicating a consistent ethical trend within the system that prioritizes quality of life over futile interventions or the artificial prolongation of the dying process. This preference was guided by algorithms such as evaluate_palliative_care(), evaluate_non_maleficence(), and deliver_bad_news_technique(). These findings are synthesized in Table 5. Ethical-Computational Trends and Correlations, which provides a structured overview of the relationships between ethical dilemmas, prioritized principles, and system-generated decisions. Table 5. Ethical-Computational Trends Observed in the 13 Clinical Cases Analyzed with BIOETHICARE 360° Analyzed Indicator Identified Value Frequency Ethical-Computational Interpretation Most Frequent Dilemma Patient Autonomy vs. Family Beneficence 5 Dominant conflict in end-of-life decisions; requires empathetic balancing between individual will and perceived duty. Most Prioritized Principle (Physician) Autonomy Weighted Average: 4.3/5 Physicians prioritize respect for patient decisions, reflecting consistency with legal standards and personalist bioethics. Most Prioritized Principle (Family) Beneficence Weighted Average: 4.6/5 Families tend to favor interventions due to hope or lack of prognostic understanding; highlights need for guided mediation. Most Frequent Final Recommendation Early Integration of Palliative Care 10 The tool favors quality of life over futile interventions, confirming alignment with the principle of non-maleficence and palliative care models. Reflection on System Robustness BIOETHICARE 360® demonstrated a strong capacity to identify common ethical patterns, process them using standardized normative criteria, and offer recommendations aligned with principlist bioethics frameworks, patient rights, and local legal regulations. The system’s multiperspective design promoted the active inclusion of physicians, families, and ethics committees, enabling collaborative ethical deliberation. Additionally, the visualization of tensions and the intervention of the Gemini Assistant (Bioethics Assistant with Gemini) enhanced emotional and legal understanding of decisions, strengthening both reflective processes and real-time clinical application. Discussion The implementation of BIOETHICARE 360° in 13 ethically charged clinical scenarios enabled the exploration of the software’s capacity not only as a decision-support tool, but also as a bioethical deliberative agent capable of generating structured argumentative processes. Through its ethics-by-design architecture, the system successfully integrated core ethical principles with algorithmic logic, interactive visualization, and reasoning assisted by generative AI, meeting international standards for trustworthy and explainable digital health systems ( 1 , 7 , 8 ). One of the system’s key contributions was its ability to operate through a multiperspective structure, allowing each stakeholder (physician, family, committee) to weigh bioethical principles independently. This feature enabled the automated detection of inter-actor ethical conflicts—such as the repeated tension between patient autonomy and family beneficence (observed in 5 of the 13 cases). This capability is particularly relevant in end-of-life decisions, where balancing advance directives, suffering, and emotional bonds is often delicate and difficult to interpret ( 3 , 4 ). The intervention of the Gemini Bioethics Assistant, powered by generative AI, proved to be an innovative component that enriched deliberative processes with natural language, empathetic explanations, and context-sensitive narrative arguments. This function was especially valuable in emotionally charged situations or when translating medical and legal concepts into accessible language for involved stakeholders. From a technical perspective, the bias verification, ethical evaluation, and computational deliberation modules demonstrated high consistency in their recommendations. Suggested decisions—such as early activation of palliative care, informed consent validation, and use of the SPIKES protocol for breaking bad news—were in alignment with current clinical guidelines and with established bioethical frameworks including principlism, narrative ethics, and deliberative ethics ( 4 , 13 , 14 ). The comparative analysis of five cases showed that the system could adapt its deliberative logic to a variety of dilemmas: from therapeutic refusal to absence of advance directives or conflicts between families and ethics committees. The system’s algorithmic flexibility was evident in its application of binary decision trees, Boolean verification, weighted analysis, and both Diego Gracia’s deliberative model and the clinical-ethical methodology of Anderson Díaz, resulting in structured and traceable outputs. From a methodological standpoint, the system’s robustness was reflected in its ability to maintain narrative and ethical coherence across all cases without relying solely on manual user input. The integration of structured analysis, computational functions, and dynamic visualizations helped reinforce clinical team trust in the recommendations generated, while also fostering real-time bioethical education. Finally, the system’s algorithmic transparency and its ability to document each step of the deliberation process strengthen its potential as an evaluable instrument for hospital ethics committees, medical education programs, and clinical settings marked by uncertainty. Its real-world applicability, reproducibility, and alignment with trustworthy AI principles promoted by the European Commission and UNESCO position it as a significant advancement at the intersection of technology, ethics, and medicine ( 1 , 3 , 5 , 7 , 8 , 15 – 17 ). Conclusions The implementation of BIOETHICARE 360°, enhanced by the Gemini conversational assistant, has proven to be an advanced and robust tool for ethical-clinical decision-making. Based on the analysis of 13 real-world scenarios, including five in-depth case studies, we conclude that the system functions not only as a traditional CDSS, but also as a bioethical deliberative agent capable of: Explicitly detecting and reasoning through tensions between ethical principles (e.g., autonomy vs. beneficence), using a combination of quantitative logic, ethical semantics, and visual analysis—validating a truly explainable AI approach to healthcare. Providing clinical recommendations—such as early integration of palliative care and use of the SPIKES protocol—based on recognized ethical frameworks (principlism, narrative ethics, deliberative ethics) and consistent with contemporary literature. Offering AI-assisted deliberation, promoting clear and empathetic communication in natural language—benefiting professionals, patients, and families by bridging the gap between theoretical ethics and practical clinical application. Demonstrating advantages over earlier rule-based solutions (such as DXplain (18) or traditional CDSSs (5,18–20)) by integrating a multiperspective approach, traceability, bias verification, and narrative reasoning—thus adhering to the governance, transparency, and robustness principles outlined in FUTURE AI (21). Relevance and Future Projection The system demonstrates broad applicability and high utility, with an ethically and technically sound architecture that enables replication of clinical ethics committee procedures and the training of professionals in real-world settings. Its ethics-by-design framework, which combines algorithmic logic, qualitative analysis, visual interface, and AI-driven dialogue, complies with international standards for trustworthy artificial intelligence and ethical continuity in medicine. Future potential includes the integration of external audits, regulatory adaptability, and pilot testing—scaling its clinical and educational deployment while ensuring the sustainable evolution of the tool. In summary, BIOETHICARE 360° represents a significant advancement in computational support for bioethical deliberation, with the capacity to enhance complex decision-making, transparency, and ethical training in healthcare settings. Availability and Requirements Project name: BIOETHICARE 360° Project homepage: https://bioethicare-360.streamlit.app/ Operating system(s): Platform-independent — compatible with Windows, macOS, and Linux Programming language: Python (using frameworks such as Streamlit, Plotly, ReportLab), JavaScript (for interactive visualization and front-end libraries), and SQL/NoSQL for database management. Other requirements. Python ≥ 3.8 with streamlit, plotly, reportlab, and firebase-admin libraries Access to a generative AI API (e.g., Google Gemini) Optional: Firebase account for data management and authentication License. BIOETHICARE 360° is registered as an original and unpublished software work under Colombian copyright law (Registration No. 1-2024-133839), under the name of Anderson Díaz Pérez. All rights reserved. Its use, reproduction, or adaptation requires express authorization from the author. It is not distributed under an open-source license. Restrictions for non-academic use. None. The tool is available free of charge for clinical, educational, or non-commercial research purposes upon prior authorization by the authors. Abbreviations Abbreviation Definition AI Artificial Intelligence CDSS Clinical Decision Support System SPIKES Protocol for breaking bad news (Setting, Perception, Invitation, Knowledge, Empathy, Summary) MIT Massachusetts Institute of Technology (software license type) JSON JavaScript Object Notation UI User Interface Declarations This study does not involve any real human participants, clinical data, or identifiable patient information. All cases analyzed by the BIOETHICARE 360® system were synthetically generated for ethical-computational simulation purposes. Therefore, no consent for publication from patients or legal guardians was required. The manuscript does not contain images, videos, or personal details of any individual." Chrissy and Roma confirmed this works for us. Ethics approval and consent to participate Not applicable. This study relied exclusively on synthetic data generated for bioethical dilemmas. No human or animal participants were involved. Consent for publication. Not applicable. No real personal data, images, or identifiable multimedia are presented . Availability of data and materials. All data used in this study were synthetically generated and are included as supplementary files to ensure reproducible availability. The source code of BIOETHICARE 360°, including the functions described in this article, is not hosted in a public GitHub repository and will not be archived in Zenodo. Availability of data and materials All data used in this study were synthetically generated for simulation purposes and are provided as supplementary files to ensure reproducibility. The source code of the BIOETHICARE 360° system is not publicly available in repositories such as GitHub or Zenodo. However, access to the software for academic or clinical evaluation can be granted upon reasonable request to the corresponding author. Intellectual Property Protection. A copy of the official copyright registration certificate is included as supplementary material. This certificate was issued by the National Copyright Office of Colombia (Radication No. 1-2024-133839), under the name of Anderson Díaz Pérez, for the software titled BIOETHICARE 360°: An Integrated Bioethical Approach to Clinical Decision-Making , registered on December 16, 2024. This registration certifies authorship and legal protection of the software as an original and unpublished computer program. Competing interests. The authors declare no financial or non-financial competing interests related to this work. Funding No external funding was received for the development of this work. Authors' contributions. AD designed and coded the system, created the ethical weighting and visualization functions. JJ developed the architecture of the conversational assistant and defined the clinical case studies. Both authors integrated the system, conducted the comparative analysis, and co-authored the manuscript. AD and JJ approved the final version. Acknowledgements Not applicable. No external contributions were received that require acknowledgment. Authors’ information. AD holds a PhD in Bioethics and is a specialist in AI applied to health. JJ is an Industrial Engineer with experience in clinical application development and natural language processing with an AI focus. References Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44–56. https://doi.org/10.1038/s41591-018-0300-7 . Floridi L, Cowls J, Beltrametti M, Chatila R, Chazerand P, Dignum V, et al. AI4People—An ethical framework for a good AI society: Opportunities, risks, principles, and recommendations. Minds Mach. 2018;28(4):689–707. https://doi.org/10.1007/s11023-018-9482-5 . Gerke S, Minssen T, Cohen G. Ethical and legal challenges of artificial intelligence-driven healthcare. Artif Intell Healthc. 2020;295–336. https://doi.org/10.1080/15265161.2020.1764133 . Obermeyer Z, Powers B, Vogeli C, Mullainathan S. Dissecting racial bias in an algorithm used to manage the health of populations. Science. 2019;366(6464):447–53. https://doi.org/10.1126/science.aax2342 . Gianfrancesco MA, Tamang S, Yazdany J, Schmajuk G. Potential biases in machine learning algorithms using electronic health record data. JAMA Intern Med. 2018;178(11):1544–7. https://doi.org/10.1001/jamainternmed.2018.3763 . Benjamin R. Assessing risk, automating racism. Science. 2019;366(6464):421–2. https://doi.org/10.1126/science.aaz38 . UNESCO. Recommendation on the Ethics of Artificial Intelligence [Internet]. Available from: https://www.unesco.org/en/articles/recommendation-ethics-artificial-intelligence European Commission. Ethics guidelines for trustworthy AI [Internet]. Publications Office of the European Union. 2019. Available from: https://data.europa.eu/doi/10.2759/346720 Díaz Pérez A. Metodología Integral de Análisis Ético-Clínico (MIAEC): un nuevo paradigma para la resolución de dilemas al final de la vida. Acta Bioeth. 2024;30(2):207–18. https://doi.org/10.4067/S1726-569X2024000200207 . Júdez J. La deliberación moral: el método de la ética clínica. Med Clin (Barc). 2001;117(1):18–23. https://doi.org/10.1016/S0025-7753(01)71998-7 . Parra Pineda MO. La deliberación en la toma de decisiones bioético-clínicas según Diego Gracia. Investig Andina . 2018;20(37):27–40. Available from: http://www.scielo.org.co/scielo.php?script=sci_arttext&pid=S0124-81462018000200027 Gracia D. Problemas con la deliberación. Folia Humanística. 2019;31–16. https://doi.org/10.30860/0013 . Beauchamp TL, Childress JF. Principles of Biomedical Ethics . 8th ed. Oxford University Press; 2019. Available from: https://global.oup.com/ushe/product/principles-of-biomedical-ethics-9780190640873 Andorno R. Human dignity and human rights as a common ground for a global bioethics. J Med Philos. 2009;34(3):223–40. https://doi.org/10.1093/jmp/jhp023 . Mennella C, Maniscalco U, De Pietro G, Esposito M. Ethical and regulatory challenges of AI technologies in healthcare: A narrative review. Heliyon. 2024;10(4):e26297. https://doi.org/10.1016/j.heliyon.2024.e26297 . Geppert CMA, Shelton W. Health care ethics committees as mediators of social values and the culture of medicine. AMA J Ethics. 2016;18(5):534–9. https://doi.org/10.1001/journalofethics.2016.18.5 . Aucouturier E, Grinbaum A. Training bioethics professionals in AI ethics: A framework. J Law Med Ethics. 2025;53(1):176–83. https://doi.org/10.1017/jme.2025.57 . The Laboratory of Computer Science. DXplain Project [Internet]. Available from: http://www.mghlcs.org/projects/dxplain Wolters Kluwer. Information is the Best Medicine [Internet]. Available from: https://www.wolterskluwer.com/en/know/ent-int-clinical-leadership-resources Valdivia Navarro F, Pérez Rosa D. Sistema de soporte a las decisiones clínicas relacionadas con el diagnóstico precoz de enfermedades. Rev Cubana Inf Médica . 2016;8:533–44. Available from: http://scielo.sld.cu/scielo.php?script=sci_arttext&pid=S1684-18592016000300006 National Center for Biotechnology Information. FUTURE AI – Bookshelf [Internet]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK610549/?term=FUTURE%20AI Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7078293","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":488679227,"identity":"c71c7053-5cee-42b7-bfe5-02540605c217","order_by":0,"name":"Anderson Díaz Pérez","email":"data:image/png;base64,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","orcid":"","institution":"Corporación Universitaria Rafael Nuñez","correspondingAuthor":true,"prefix":"","firstName":"Anderson","middleName":"Díaz","lastName":"Pérez","suffix":""},{"id":488679238,"identity":"f8381655-c581-4402-89d5-9dd16af5f3ce","order_by":1,"name":"Joseph Javier Sánchez Acuña","email":"","orcid":"","institution":"Corporación Universitaria Minuto de Dios","correspondingAuthor":false,"prefix":"","firstName":"Joseph","middleName":"Javier Sánchez","lastName":"Acuña","suffix":""}],"badges":[],"createdAt":"2025-07-08 23:08:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7078293/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7078293/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":87895478,"identity":"e16d6071-b2eb-4af6-b36c-076141fa5034","added_by":"auto","created_at":"2025-07-30 07:28:33","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":54104,"visible":true,"origin":"","legend":"\u003cp\u003eFunctional flowchart of\u003cstrong\u003ethe BIOETHICARE 360° system.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7078293/v1/2149f6f2ffbea5ab5ea2288a.jpg"},{"id":89787272,"identity":"301acdcd-0d79-4a8b-bb50-f3a333ccf2a2","added_by":"auto","created_at":"2025-08-25 04:38:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1028579,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7078293/v1/4a5e1dae-41b3-4494-a3f2-bba15e326cd1.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eBIOETHICARE 360\u003csup\u003eO\u003c/sup\u003e: A Deliberative Software for Ethical Computational Support in Complex Clinical Decision-Making\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eArtificial intelligence (AI) has emerged as a transformative technology in clinical settings, particularly in ethically complex scenarios such as palliative care and end-of-life decision-making. Machine learning and deep learning models enable the real-time analysis of large volumes of clinical data to generate diagnostic or therapeutic recommendations (1). However, their application in sensitive contexts introduces significant bioethical risks related to patient autonomy, distributive justice, algorithmic transparency, and professional accountability (2,3). \u0026nbsp;Key challenges include algorithmic bias stemming from non-representative training data, heuristic shortcuts embedded in models to optimize statistical performance, and skewed predictions when applying generalized logic to individualized clinical contexts (4\u0026ndash;6). These limitations are particularly critical in terminally ill patients, where personalization, ethical deliberation, and respect for advance directives are clinical imperatives.\u003c/p\u003e\n\u003cp\u003eIn response, international organizations such as UNESCO, the OECD, and the European Commission have promoted ethical principles for the responsible development and deployment of AI, including autonomy, harm prevention, justice, explainability, technical robustness, and human oversight (7,8). Nevertheless, most AI-based clinical solutions still lack mechanisms to evaluate, visualize, or document the application of these principles throughout the clinical decision-making process. \u0026nbsp;To address this gap, we developed BIOETHICARE 360\u0026deg;, a computational application implemented in Python and deployed via Streamlit, designed to enable automated and deliberative analysis of complex bioethical dilemmas. The system is composed of several functional modules:\u003c/p\u003e\n\u003col start=\"1\" type=\"1\"\u003e\n \u003cli\u003ePreliminary analysis using generative AI\u003cul type=\"circle\"\u003e\n \u003cli\u003eEmploys advanced language models (Bioethics Assistant powered by Gemini) to analyze clinical narratives and extract both explicit and implicit bioethical conflicts.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eStructured case registration\u003cul type=\"circle\"\u003e\n \u003cli\u003eEnables the collection of clinical, cultural, and ethical data from multiple perspectives: the medical team, the patient/family, and the bioethics committee.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eMultiperspective ethical weighting\u003cul type=\"circle\"\u003e\n \u003cli\u003eAssigns scores (0 to 5) to the principles of autonomy, beneficence, non-maleficence, and justice for each stakeholder.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eAlgorithmic audit and ethical verification\u003cul type=\"circle\"\u003e\n \u003cli\u003eThe system detects imbalances, omissions, and disproportionate ethical considerations, providing automated warnings and ethical suggestions.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eAdvanced graphical visualization\u003cul type=\"circle\"\u003e\n \u003cli\u003eGenerates radar and bar charts to represent consensus and dissent, enhancing visual traceability and interdisciplinary deliberation.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eBioethical deliberative engine\u003cul type=\"circle\"\u003e\n \u003cli\u003eBased on Diego Gracia\u0026rsquo;s model, it articulates facts, values, and potential courses of action, integrating ethical principles, national legal frameworks, and the patient\u0026rsquo;s sociocultural context.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eStructured report generation\u003cul type=\"circle\"\u003e\n \u003cli\u003eProduces a comprehensive PDF report including all perspectives, visualizations, ethical deliberation, and recommendations for the clinical team.\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThis article presents the software\u0026rsquo;s architecture, deliberative methodology, and its empirical application to a real clinical case involving a terminally ill patient, where an algorithmic recommendation for palliative care conflicted with the family\u0026rsquo;s religiously motivated desire to continue life-sustaining interventions. The tool enables real-time identification and mitigation of ethical bias, fosters interprofessional deliberation, and documents decisions within a transparent and ethically justifiable framework.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImplementation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe BIOETHICARE 360\u0026deg; software was developed with the purpose of providing a technological platform that integrates ethical deliberation, clinical reasoning, and automated recommendations in the care of terminally ill patients. Its modular architecture and \u003cem\u003eethics-by-design\u003c/em\u003e approach ensure full traceability of each computational decision, guaranteeing transparency, explainability, and normative support. The system was implemented in Python 3.10, deployed via Streamlit, and supported by a Firebase Firestore database, making it fully operational in both local and cloud environments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGeneral Architecture and Design Logic\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe architecture of BIOETHICARE 360\u0026deg; consists of four main modules, as illustrated in Figure 1.\u003c/p\u003e\n\u003cp\u003eThe core of the system is built around a main class called \u003cstrong\u003eBioethicalCase\u003c/strong\u003e, in which each attribute represents an essential dimension of clinical and bioethical analysis: patient name, age, condition, medical history, ethical dilemma, sociocultural background, and multiperspective ethical scoring (physician, family, ethics committee). \u0026nbsp; Table 1 below illustrates the association between each evaluated dimension, its expected clinical or ethical output, and the corresponding function or method executed within the code:\u003c/p\u003e\n\u003cp\u003eTable 1. Association between bioethical dimensions, clinical/ethical outputs, and code functions\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eEvaluated Dimension\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eClinical/Ethical Recommendation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eFunction or Attribute in the Code\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eModel Name\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBioethics Care Decision Model\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e__init__()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAge / Gender / Condition / Medical History\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRegistration of terminal patient and functional context\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eage, gender, condition, medical_history\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMultiperspective Ethical Weighting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAssignment of scores (0 to 5) by principle and stakeholder\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eperspectives = {...}\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eEthical Bias Evaluation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDetection of imbalances, omissions, and dominance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003everify_ethics_bias()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBioethical Visualization\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRadar chart of principles and consensus/dissent graph\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003egenerate_advanced_visualizations()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eComprehensive Patient Assessment\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDetermines prioritization for palliative care\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ecomprehensive_evaluation()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eEthical Conflicts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDetects tensions such as \u0026quot;autonomy vs. beneficence\u0026quot;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eanalyze_bioethical_conflicts()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNarrative Ethical Reflection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eQualitative analysis of each principle by perspective\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ereflect_on_bioethical_principles()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eComputational Recommendation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eIntegration of analyses into a reasoned ethical decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eevaluate_recommendation_type()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eStructured Deliberation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eImplements Diego Gracia and Anderson D\u0026iacute;az\u0026rsquo;s methodology\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eanalyze_methodologies()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eLegal and Deontological Evaluation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMapping of potential sanctions under Colombian law\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eanalyze_legal_aspects()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eInformed Consent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAutomated generation with regulatory references\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003egenerate_informed_consent()\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eEthical Algorithmic Evaluation and Bias Detection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn addition to its deliberative analysis capabilities, the system performs an ethical algorithmic audit aimed at identifying structural biases, heuristic simplifications, and ethical risks embedded in the automated reasoning process (Table 2).\u003c/p\u003e\n\u003cp\u003eTable 2. Integration of Computational Logic, Ethical Models, and Mitigation Strategies\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAnalyzed Dimension\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eApplied Algorithmic Logic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eApplied Ethical Model\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDetected Bias\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMitigation Strategy\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eComprehensive Clinical Assessment\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eif statements based on functional/cognitive thresholds\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePrinciplism (Beauchamp \u0026amp; Childress)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRigid thresholds fail to capture clinical nuances\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAdjust thresholds using variable combinations + expert clinical judgment\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eScaled Bioethical Weighting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDictionary with Likert scale (0\u0026ndash;5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePrinciplism\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEthical dimensions reduced to quantifiable metrics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePost-validation through narrative reflection\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eQualitative Ethical Reflection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eif-elif-else trees on ethical values\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNarrative / Clinical Ethics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAvailability heuristic (overweighting salient principles)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRequire balanced representation of all four principles across stakeholders\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eConflicts Between Principles\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSymbolic Boolean logic evaluation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBioethical Casuistry\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBinary simplification of complex dilemmas\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eInclusion of compound tensions and cross-deliberation mechanisms\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBioethical Deliberation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStructured text + nested JSON functions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDiego Gracia + Anderson D\u0026iacute;az\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCognitive overload from textual complexity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGuided templates combined with graphical analysis\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eComputational Recommendation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eConditional rules based on dilemma typology\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePrinciplist-based Utilitarian Ethics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDominance of consequentialist reasoning\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCross-validation with patient autonomy and cultural context\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSPIKES (Breaking Bad News)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eStructured dictionary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEthics of Compassion / Clinical Protocols\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRigidity in religious or emotionally complex situations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEmpathic personalization + cultural and psychological tailoring\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eLegal Aspects\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDictionary by legal category\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLegal Bioethics (Colombia)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRisk of international legal extrapolation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGeo-adaptive legal references based on local jurisdiction\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eInformed Consent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAutomated templates\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eProcedural / Legal Ethics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGeneric or inappropriate consent generation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCustomization by cognitive competence, age, and advance directives\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eThis integrative approach represents not only a functional computational tool but also a methodological advancement that systematizes clinical bioethical deliberation through a structured, auditable, and traceable framework, with the potential for replication in diverse clinical contexts.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eApplication of the System to Ethically Complex Clinical Cases\u003c/p\u003e\n\u003cp\u003eBIOETHICARE 360\u0026deg; was applied to thirteen representative clinical-bioethical cases, encompassing common end-of-life dilemmas. The tool processed both structured and unstructured information, enabling different stakeholders (medical team, family, and ethics committee) to weigh core principles such as autonomy, beneficence, non-maleficence, and justice. Additionally, the system incorporated ethical deliberation via a conversational AI module (\u003cem\u003eBioethics Assistant powered by Gemini\u003c/em\u003e), which contributed narrative analyses, context-sensitive recommendations, and empathetic computational language.\u003c/p\u003e\n\u003cp\u003eTo empirically validate the functionality of BIOETHICARE 360\u0026deg;, the system was applied to a real clinical case involving a 68-year-old male patient with terminal-stage pancreatic adenocarcinoma, functional decline, and a life expectancy of less than three months. The patient was conscious, oriented, and had previously recorded advance directives requesting limitation of therapeutic efforts and initiation of comprehensive palliative care. However, his family insisted on continuing life-sustaining treatments for religious and emotional reasons, creating an ethical conflict between the patient\u0026apos;s autonomy and the family\u0026rsquo;s perceived beneficence.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eI. Representative Case Study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAn emblematic case involved a 68-year-old male patient with metastatic pancreatic cancer who had explicitly stated advance directives declining invasive life-sustaining measures. The family, however, insisted on maintaining full support. The system identified the main ethical dilemma as a conflict between autonomy and beneficence.\u003c/p\u003e\n\u003cp\u003eThe computational analysis revealed a significant discrepancy in ethical perspectives among stakeholders (a difference greater than 6 points) and a notable omission of the autonomy principle by the family. The automated ethical audit triggered alerts and consistently recommended transitioning to palliative care, including compassionate communication guided by the SPIKES protocol.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStructured Case Registration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eClinical and sociocultural data were entered into the system interface, including:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eFunctional status: 2/5\u003c/li\u003e\n \u003cli\u003eCognitive status: 2/5\u003c/li\u003e\n \u003cli\u003ePerceived quality of life: 1/5\u003c/li\u003e\n \u003cli\u003eLevel of suffering: 4/5\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThe clinical history was first analyzed using generative AI (Gemini), which identified the main dilemma as: \u0026ldquo;tension between the patient\u0026rsquo;s advance directives and the family\u0026rsquo;s desire to continue invasive treatment.\u0026rdquo;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMultiperspective Ethical Weighting\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthical scores were assigned to the four foundational bioethical principles by each stakeholder:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eMedical team: Autonomy (5), Beneficence (3), Non-maleficence (4), Justice (4)\u003c/li\u003e\n \u003cli\u003eFamily: Autonomy (2), Beneficence (5), Non-maleficence (2), Justice (3)\u003c/li\u003e\n \u003cli\u003eEthics committee: Autonomy (4), Beneficence (3), Non-maleficence (3), Justice (4)\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThese scores were automatically visualized through radar and bar charts (see Figures 2 and 3), facilitating dissent analysis and highlighting the greater imbalance in the family\u0026rsquo;s perspective.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical Audit and Detected Biases\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe automated ethical verification module identified:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eOmission of the autonomy principle in the family\u0026rsquo;s perspective.\u003c/li\u003e\n \u003cli\u003eInter-actor imbalance: a 6-point gap between total scores of the medical team and the family.\u003c/li\u003e\n \u003cli\u003eAffective dominance warning: high beneficence scoring without proportional evaluation of suffering.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThe system advised reviewing the weightings and promoting a guided joint deliberation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAssisted Deliberation and Computational Recommendation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBIOETHICARE 360\u0026deg; applied two ethical deliberation models:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eAnderson D\u0026iacute;az Model (MIEC) (9)\u003cbr\u003e\u0026nbsp;\u0026rarr; Determined that the patient required priority access to palliative care. AI recommended discontinuing futile treatments and reinforcing psychoemotional support.\u003c/li\u003e\n \u003cli\u003eDiego Gracia Model (10\u0026ndash;12)\u003cbr\u003e\u0026nbsp;\u0026rarr; Proposed a deliberation process balancing clinical facts, conflicting values, and normative principles. The suggested decision was to respect the patient\u0026apos;s advance directives, communicate the situation compassionately to the family, and initiate the SPIKES protocol.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eStructured Report and Automated Consent Generation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe system generated a comprehensive PDF report integrating:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eDetailed case description\u003c/li\u003e\n \u003cli\u003eComparative ethical perspectives\u003c/li\u003e\n \u003cli\u003eVisualization of consensus and dissent\u003c/li\u003e\n \u003cli\u003eNarrative deliberative analysis\u003c/li\u003e\n \u003cli\u003eFinal ethical recommendation: \u003cem\u003e\u0026ldquo;Immediate transition to comprehensive palliative care\u0026rdquo;\u003c/em\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eIn addition, an automated ethical-legal informed consent form was generated, customized with patient-specific data and incorporating Colombian regulatory references.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical and Educational Impact\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe use of the system enabled:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eEthical traceability of the clinical decision\u003c/li\u003e\n \u003cli\u003eMitigation of biases related to emotional pressure and therapeutic inertia\u003c/li\u003e\n \u003cli\u003eReal-time ethical training for the medical team\u003c/li\u003e\n \u003cli\u003eImproved interprofessional and family communication\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eII. Comparative Analysis of Five Cases\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo assess the ethical consistency, computational adaptability, and clinical sensitivity of the BIOETHICARE 360\u0026deg; software, five representative clinical cases were selected, each featuring distinct dilemmas varying in nature and complexity. These cases served as test scenarios to evaluate the system\u0026rsquo;s behavior in resolving real-world ethical conflicts in end-of-life care. \u0026nbsp;Each case was processed using the system\u0026rsquo;s structured logic, which includes:\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eMultiperspective capture of bioethical principles (autonomy, beneficence, non-maleficence, and justice)\u003c/li\u003e\n \u003cli\u003eExecution of specific functions depending on the dilemma type (e.g., evaluate_autonomy(), analyze_bioethical_conflicts(), or deliver_bad_news_technique())\u003c/li\u003e\n \u003cli\u003eEmpathetic and contextual deliberation via the Bioethics Assistant (Gemini), functioning as a digital mediator capable of analyzing, reformulating, and suggesting narrative-based actions\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eTable 4 presents a detailed synthesis of the five cases, incorporating clinical, ethical, computational, and argumentative elements. In all scenarios, the software successfully identified the underlying ethical dilemma, assigned dominant principles per stakeholder, detected specific moral tensions, and produced technically and ethically consistent recommendations.\u003c/p\u003e\n\u003cp\u003eFor instance, in Case 1, the conflict between the patient\u0026apos;s refusal of intubation and a clinical tendency toward beneficence was resolved through a binary decision tree. The system prioritized the patient\u0026apos;s autonomy and recommended palliative sedation in accordance with his advance directives. Gemini reinforced this decision with a narrative argument validating the principle of self-determination (Table 4).\u003c/p\u003e\n\u003cp\u003eBy contrast, Case 2 involved a demand for resuscitation by the family despite the patient\u0026rsquo;s terminal condition. Here, the system employed ethical weighting and conditional logic to expose a conflict between distributive justice and non-maleficence. The recommendation was to suspend interventions, promote psychosocial support, and activate the SPIKES protocol (Table 4).\u003c/p\u003e\n\u003cp\u003eIn Case 3, which lacked any recorded advance directives, BIOETHICARE 360\u0026deg; activated its modules for informed consent generation and institutional analysis. AI proposed a collaborative deliberation between the care team and ethics committee, suggesting a decision grounded in justice and normative equity (Table 4).\u003c/p\u003e\n\u003cp\u003eCase 4, involving a young patient with irreversible neurological damage, benefited from AI-assisted deliberation combining narrative ethics, quality-of-life analysis, and multiperspective ethical visualization (Table 4). The recommendation was a gradual limitation of therapeutic efforts, emphasizing compassion as the guiding principle.\u003c/p\u003e\n\u003cp\u003eFinally, in Case 5, the system identified a conflict between the patient\u0026rsquo;s autonomy and opposition from the bioethics committee. Through computational empathetic evaluation, ethical classification, and automated SPIKES activation, the tool proposed palliative care with informed consent, respecting the advance directives (Table 4).\u003c/p\u003e\n\u003cp\u003eThis comparative analysis highlights not only the system\u0026rsquo;s logical robustness but also its adaptive capacity to ethically deliberate using contextual data, legal frameworks, and emotional dimensions of care. The system\u0026rsquo;s architecture allowed each case to be processed with precision and sensitivity, integrating specialized algorithmic components and high-quality AI-generated arguments.\u003c/p\u003e\n\u003cp\u003eThe integration of the Gemini assistant was key to enriching ethical reasoning with narrative language, empathetic explanations, and clinical communication strategies. This is especially valuable in emotionally charged contexts, where decisions must be not only technically sound but also humanely and ethically justified.\u003c/p\u003e\n\u003cp\u003eTogether, these five cases validate the system\u0026rsquo;s potential as a clinical-bioethical decision support tool, offering algorithmic transparency, personalized recommendations, and respect for human rights in healthcare (Table 4).\u003c/p\u003e\n\u003cp\u003eTable 4. Extended Ethical-Algorithmic Comparative Analysis of 5 Representative Cases\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eCase\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eCentral Ethical Dilemma\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eMost Emphasized Principle\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003ePrimary Conflict Detected\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eFinal Ethical Recommendation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eSoftware Functions Involved\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eGemini Assistant (Conversational AI) Intervention\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eType of Computational Analysis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eApplied Ethical Model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePatient refusal of intubation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAutonomy (physician)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAutonomy vs. Beneficence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRespect patient\u0026apos;s decision + initiate palliative sedation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eevaluate_autonomy(), analyze_bioethical_conflicts(), reflect_on_bioethical_principles()\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAI-generated recommendations based on natural language input dilemma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBinary decision tree + conditional logic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePrinciplism + Personalist Bioethics\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFamily demands resuscitation in terminal phase\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBeneficence (family)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eJustice vs. Non-Maleficence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSuspend interventions + provide psychosocial support\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eevaluate_justice(), evaluate_non_maleficence(), analyze_methodologies()[0]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAI-mediated deliberation addressing emotional tension and medical viability\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWeighted evaluation + conditional logic + inter-actor conflict detection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eClinical Bioethics + Distributive Justice\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e3\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePatient without advance directives\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eJustice (ethics committee)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAutonomy vs. Legal Uncertainty\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eInterdisciplinary evaluation + ethics committee deliberation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eevaluate_interdisciplinary_team_role(), generate_informed_consent()\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eCollaborative and reflective proposal addressing regulatory gaps through narrative reasoning\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRule-based structured evaluation + legal consent generation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eClassical Deliberative Model + Institutional Ethics\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e4\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eYoung patient with irreversible brain damage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNon-Maleficence (physician)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFamily Autonomy vs. Prognostic Reality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGradual limitation of therapeutic effort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eevaluate_non_maleficence(), analyze_methodologies()[1], reflect_on_bioethical_principles()\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAI-assisted ethical-emotional analysis based on quality of life and prognosis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDeliberative analysis + narrative comparison + multiperspective graphical weighting\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNarrative Ethics + Anderson D\u0026iacute;az Model\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003e5\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFutile intervention in terminally ill patient\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAutonomy (patient)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFamily vs. Bioethics Committee\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eInitiate palliative care with informed consent\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eevaluate_palliative_care(), analyze_bioethical_conflicts(), deliver_bad_news_technique()\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eComputational ethical validation + AI-guided empathetic communication to family and committee\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEmpathetic computational assessment + ethical classification + automated SPIKES\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePrinciplism + Palliative Bioethics + Ethics of Compassion\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eIII. Identified Ethical Trends and Patterns\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe global analysis of the 13 clinical cases processed by BIOETHICARE 360\u0026deg; revealed a series of recurring patterns that reflect both the ethical behavior of stakeholders and the system\u0026rsquo;s internal logical consistency.\u003c/p\u003e\n\u003cp\u003eThe most frequent dilemma identified was the conflict between patient autonomy and family beneficence, which appeared in 5 of the 13 cases. This finding confirms that end-of-life decisions commonly involve tensions between respecting the patient\u0026rsquo;s will and the family\u0026rsquo;s desire to prolong life for emotional, hopeful, or informational reasons. The system identified this dilemma using semantic analysis functions (analyze_bioethical_conflicts()) and clinical-ethical decision trees.\u003c/p\u003e\n\u003cp\u003eRegarding the prioritization of ethical principles, physicians tended to emphasize autonomy (average 4.3/5), while families placed greater emphasis on beneficence (average 4.6/5). This ethical discrepancy was visualized through the system\u0026rsquo;s multiperspective weighting maps, enabling early recognition of potential value conflicts between stakeholders.\u003c/p\u003e\n\u003cp\u003eMoreover, in 77% of cases (10 out of 13), the final recommendation was the early integration of palliative care, indicating a consistent ethical trend within the system that prioritizes quality of life over futile interventions or the artificial prolongation of the dying process. This preference was guided by algorithms such as evaluate_palliative_care(), evaluate_non_maleficence(), and deliver_bad_news_technique(). \u0026nbsp;These findings are synthesized in Table 5. Ethical-Computational Trends and Correlations, which provides a structured overview of the relationships between ethical dilemmas, prioritized principles, and system-generated decisions.\u003c/p\u003e\n\u003cp\u003eTable 5. Ethical-Computational Trends Observed in the 13 Clinical Cases Analyzed with BIOETHICARE 360\u0026deg;\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAnalyzed Indicator\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eIdentified Value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFrequency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEthical-Computational Interpretation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMost Frequent Dilemma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePatient Autonomy vs. Family Beneficence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDominant conflict in end-of-life decisions; requires empathetic balancing between individual will and perceived duty.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMost Prioritized Principle (Physician)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAutonomy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWeighted Average: 4.3/5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePhysicians prioritize respect for patient decisions, reflecting consistency with legal standards and personalist bioethics.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMost Prioritized Principle (Family)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eBeneficence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eWeighted Average: 4.6/5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFamilies tend to favor interventions due to hope or lack of prognostic understanding; highlights need for guided mediation.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMost Frequent Final Recommendation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eEarly Integration of Palliative Care\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eThe tool favors quality of life over futile interventions, confirming alignment with the principle of non-maleficence and palliative care models.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eReflection on System Robustness\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBIOETHICARE 360\u0026reg; demonstrated a strong capacity to identify common ethical patterns, process them using standardized normative criteria, and offer recommendations aligned with principlist bioethics frameworks, patient rights, and local legal regulations.\u003c/p\u003e\n\u003cp\u003eThe system\u0026rsquo;s multiperspective design promoted the active inclusion of physicians, families, and ethics committees, enabling collaborative ethical deliberation. Additionally, the visualization of tensions and the intervention of the Gemini Assistant (Bioethics Assistant with Gemini) enhanced emotional and legal understanding of decisions, strengthening both reflective processes and real-time clinical application.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe implementation of BIOETHICARE 360\u0026deg; in 13 ethically charged clinical scenarios enabled the exploration of the software\u0026rsquo;s capacity not only as a decision-support tool, but also as a bioethical deliberative agent capable of generating structured argumentative processes. Through its \u003cem\u003eethics-by-design\u003c/em\u003e architecture, the system successfully integrated core ethical principles with algorithmic logic, interactive visualization, and reasoning assisted by generative AI, meeting international standards for trustworthy and explainable digital health systems (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eOne of the system\u0026rsquo;s key contributions was its ability to operate through a multiperspective structure, allowing each stakeholder (physician, family, committee) to weigh bioethical principles independently. This feature enabled the automated detection of inter-actor ethical conflicts\u0026mdash;such as the repeated tension between patient autonomy and family beneficence (observed in 5 of the 13 cases). This capability is particularly relevant in end-of-life decisions, where balancing advance directives, suffering, and emotional bonds is often delicate and difficult to interpret (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe intervention of the Gemini Bioethics Assistant, powered by generative AI, proved to be an innovative component that enriched deliberative processes with natural language, empathetic explanations, and context-sensitive narrative arguments. This function was especially valuable in emotionally charged situations or when translating medical and legal concepts into accessible language for involved stakeholders.\u003c/p\u003e\u003cp\u003eFrom a technical perspective, the bias verification, ethical evaluation, and computational deliberation modules demonstrated high consistency in their recommendations. Suggested decisions\u0026mdash;such as early activation of palliative care, informed consent validation, and use of the SPIKES protocol for breaking bad news\u0026mdash;were in alignment with current clinical guidelines and with established bioethical frameworks including principlism, narrative ethics, and deliberative ethics (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe comparative analysis of five cases showed that the system could adapt its deliberative logic to a variety of dilemmas: from therapeutic refusal to absence of advance directives or conflicts between families and ethics committees. The system\u0026rsquo;s algorithmic flexibility was evident in its application of binary decision trees, Boolean verification, weighted analysis, and both Diego Gracia\u0026rsquo;s deliberative model and the clinical-ethical methodology of Anderson D\u0026iacute;az, resulting in structured and traceable outputs.\u003c/p\u003e\u003cp\u003eFrom a methodological standpoint, the system\u0026rsquo;s robustness was reflected in its ability to maintain narrative and ethical coherence across all cases without relying solely on manual user input. The integration of structured analysis, computational functions, and dynamic visualizations helped reinforce clinical team trust in the recommendations generated, while also fostering real-time bioethical education.\u003c/p\u003e\u003cp\u003eFinally, the system\u0026rsquo;s algorithmic transparency and its ability to document each step of the deliberation process strengthen its potential as an evaluable instrument for hospital ethics committees, medical education programs, and clinical settings marked by uncertainty. Its real-world applicability, reproducibility, and alignment with trustworthy AI principles promoted by the European Commission and UNESCO position it as a significant advancement at the intersection of technology, ethics, and medicine (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e).\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThe implementation of BIOETHICARE 360\u0026deg;, enhanced by the Gemini conversational assistant, has proven to be an advanced and robust tool for ethical-clinical decision-making. Based on the analysis of 13 real-world scenarios, including five in-depth case studies, we conclude that the system functions not only as a traditional CDSS, but also as a bioethical deliberative agent capable of:\u003c/p\u003e\n\u003col start=\"1\" type=\"1\"\u003e\n \u003cli\u003eExplicitly detecting and reasoning through tensions between ethical principles (e.g., autonomy vs. beneficence), using a combination of quantitative logic, ethical semantics, and visual analysis\u0026mdash;validating a truly \u003cem\u003eexplainable AI\u003c/em\u003e approach to healthcare.\u003c/li\u003e\n \u003cli\u003eProviding clinical recommendations\u0026mdash;such as early integration of palliative care and use of the SPIKES protocol\u0026mdash;based on recognized ethical frameworks (principlism, narrative ethics, deliberative ethics) and consistent with contemporary literature.\u003c/li\u003e\n \u003cli\u003eOffering AI-assisted deliberation, promoting clear and empathetic communication in natural language\u0026mdash;benefiting professionals, patients, and families by bridging the gap between theoretical ethics and practical clinical application.\u003c/li\u003e\n \u003cli\u003eDemonstrating advantages over earlier rule-based solutions (such as DXplain (18) or traditional CDSSs (5,18\u0026ndash;20)) by integrating a multiperspective approach, traceability, bias verification, and narrative reasoning\u0026mdash;thus adhering to the governance, transparency, and robustness principles outlined in \u003cem\u003eFUTURE AI\u003c/em\u003e (21).\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003e\u003cstrong\u003eRelevance and Future Projection\u003c/strong\u003e\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eThe system demonstrates broad applicability and high utility, with an ethically and technically sound architecture that enables replication of clinical ethics committee procedures and the training of professionals in real-world settings.\u003c/li\u003e\n \u003cli\u003eIts ethics-by-design framework, which combines algorithmic logic, qualitative analysis, visual interface, and AI-driven dialogue, complies with international standards for trustworthy artificial intelligence and ethical continuity in medicine.\u003c/li\u003e\n \u003cli\u003eFuture potential includes the integration of external audits, regulatory adaptability, and pilot testing\u0026mdash;scaling its clinical and educational deployment while ensuring the sustainable evolution of the tool.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eIn summary, BIOETHICARE 360\u0026deg; represents a significant advancement in computational support for bioethical deliberation, with the capacity to enhance complex decision-making, transparency, and ethical training in healthcare settings.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability and Requirements\u003c/strong\u003e\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003eProject name: BIOETHICARE 360\u0026deg;\u003c/li\u003e\n \u003cli\u003eProject homepage: https://bioethicare-360.streamlit.app/\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eOperating system(s): Platform-independent \u0026mdash; compatible with Windows, macOS, and Linux\u003c/li\u003e\n \u003cli\u003eProgramming language:\u003cbr\u003e\u0026nbsp;Python (using frameworks such as Streamlit, Plotly, ReportLab), JavaScript (for interactive visualization and front-end libraries), and SQL/NoSQL for database management.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eOther requirements.\u003c/strong\u003e\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cul type=\"circle\"\u003e\n \u003cli\u003ePython \u0026ge; 3.8 with streamlit, plotly, reportlab, and firebase-admin libraries\u003c/li\u003e\n \u003cli\u003eAccess to a generative AI API (e.g., Google Gemini)\u003c/li\u003e\n \u003cli\u003eOptional: Firebase account for data management and authentication\u003c/li\u003e\n \u003c/ul\u003e\n \u003cli\u003e\u003cstrong\u003eLicense.\u003cbr\u003e\u003c/strong\u003eBIOETHICARE 360\u0026deg; is registered as an original and unpublished software work under Colombian copyright law (Registration No. 1-2024-133839), under the name of Anderson D\u0026iacute;az P\u0026eacute;rez. All rights reserved. Its use, reproduction, or adaptation requires express authorization from the author. It is not distributed under an open-source license.\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eRestrictions for non-academic use.\u003c/strong\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eNone. The tool is available free of charge for clinical, educational, or non-commercial research purposes upon prior authorization by the authors.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAbbreviation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDefinition\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eArtificial Intelligence\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCDSS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eClinical Decision Support System\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSPIKES\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eProtocol for breaking bad news (Setting, Perception, Invitation, Knowledge, Empathy, Summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMIT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMassachusetts Institute of Technology (software license type)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eJSON\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eJavaScript Object Notation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eUI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eUser Interface\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Declarations","content":"\u003cp\u003eThis study does not involve any real human participants, clinical data, or identifiable patient information. All cases analyzed by the BIOETHICARE 360\u0026reg; system were synthetically generated for ethical-computational simulation purposes. Therefore, no consent for publication from patients or legal guardians was required. The manuscript does not contain images, videos, or personal details of any individual.\u0026quot; Chrissy and Roma confirmed this works for us.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable. This study relied exclusively on synthetic data generated for bioethical dilemmas. No human or animal participants were involved.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable. No real personal data, images, or identifiable multimedia are presented\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data used in this study were synthetically generated and are included as supplementary files to ensure reproducible availability. The source code of BIOETHICARE 360\u0026deg;, including the functions described in this article, is not hosted in a public GitHub repository and will not be archived in Zenodo.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eAll data used in this study were synthetically generated for simulation purposes and are provided as supplementary files to ensure reproducibility. The source code of the BIOETHICARE 360\u0026deg; system is not publicly available in repositories such as GitHub or Zenodo. However, access to the software for academic or clinical evaluation can be granted upon reasonable request to the corresponding author.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIntellectual Property Protection.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA copy of the official copyright registration certificate is included as supplementary material. This certificate was issued by the National Copyright Office of Colombia (Radication No. 1-2024-133839), under the name of Anderson D\u0026iacute;az P\u0026eacute;rez, for the software titled \u003cem\u003eBIOETHICARE 360\u0026deg;: An Integrated Bioethical Approach to Clinical Decision-Making\u003c/em\u003e, registered on December 16, 2024. This registration certifies authorship and legal protection of the software as an original and unpublished computer program.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no financial or non-financial competing interests related to this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo external funding was received for the development of this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAD designed and coded the system, created the ethical weighting and visualization functions. JJ developed the architecture of the conversational assistant and defined the clinical case studies. Both authors integrated the system, conducted the comparative analysis, and co-authored the manuscript. AD and JJ approved the final version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable. No external contributions were received that require acknowledgment.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; information.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAD holds a PhD in Bioethics and is a specialist in AI applied to health. JJ is an Industrial Engineer with experience in clinical application development and natural language processing with an AI focus.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eTopol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44\u0026ndash;56. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41591-018-0300-7\u003c/span\u003e\u003cspan address=\"10.1038/s41591-018-0300-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFloridi L, Cowls J, Beltrametti M, Chatila R, Chazerand P, Dignum V, et al. AI4People\u0026mdash;An ethical framework for a good AI society: Opportunities, risks, principles, and recommendations. Minds Mach. 2018;28(4):689\u0026ndash;707. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11023-018-9482-5\u003c/span\u003e\u003cspan address=\"10.1007/s11023-018-9482-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGerke S, Minssen T, Cohen G. Ethical and legal challenges of artificial intelligence-driven healthcare. Artif Intell Healthc. 2020;295\u0026ndash;336. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/15265161.2020.1764133\u003c/span\u003e\u003cspan address=\"10.1080/15265161.2020.1764133\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eObermeyer Z, Powers B, Vogeli C, Mullainathan S. Dissecting racial bias in an algorithm used to manage the health of populations. Science. 2019;366(6464):447\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1126/science.aax2342\u003c/span\u003e\u003cspan address=\"10.1126/science.aax2342\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGianfrancesco MA, Tamang S, Yazdany J, Schmajuk G. Potential biases in machine learning algorithms using electronic health record data. JAMA Intern Med. 2018;178(11):1544\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/jamainternmed.2018.3763\u003c/span\u003e\u003cspan address=\"10.1001/jamainternmed.2018.3763\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBenjamin R. Assessing risk, automating racism. Science. 2019;366(6464):421\u0026ndash;2. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1126/science.aaz38\u003c/span\u003e\u003cspan address=\"10.1126/science.aaz38\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUNESCO. Recommendation on the Ethics of Artificial Intelligence [Internet]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.unesco.org/en/articles/recommendation-ethics-artificial-intelligence\u003c/span\u003e\u003cspan address=\"https://www.unesco.org/en/articles/recommendation-ethics-artificial-intelligence\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEuropean Commission. Ethics guidelines for trustworthy AI [Internet]. Publications Office of the European Union. 2019. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://data.europa.eu/doi/10.2759/346720\u003c/span\u003e\u003cspan address=\"https://data.europa.eu/doi/10.2759/346720\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eD\u0026iacute;az P\u0026eacute;rez A. Metodolog\u0026iacute;a Integral de An\u0026aacute;lisis \u0026Eacute;tico-Cl\u0026iacute;nico (MIAEC): un nuevo paradigma para la resoluci\u0026oacute;n de dilemas al final de la vida. Acta Bioeth. 2024;30(2):207\u0026ndash;18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4067/S1726-569X2024000200207\u003c/span\u003e\u003cspan address=\"10.4067/S1726-569X2024000200207\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJ\u0026uacute;dez J. La deliberaci\u0026oacute;n moral: el m\u0026eacute;todo de la \u0026eacute;tica cl\u0026iacute;nica. Med Clin (Barc). 2001;117(1):18\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0025-7753(01)71998-7\u003c/span\u003e\u003cspan address=\"10.1016/S0025-7753(01)71998-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParra Pineda MO. La deliberaci\u0026oacute;n en la toma de decisiones bio\u0026eacute;tico-cl\u0026iacute;nicas seg\u0026uacute;n Diego Gracia. \u003cem\u003eInvestig Andina\u003c/em\u003e. 2018;20(37):27\u0026ndash;40. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.scielo.org.co/scielo.php?script=sci_arttext\u0026amp;pid=S0124-81462018000200027\u003c/span\u003e\u003cspan address=\"http://www.scielo.org.co/scielo.php?script=sci_arttext\u0026amp;pid=S0124-81462018000200027\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGracia D. Problemas con la deliberaci\u0026oacute;n. Folia Human\u0026iacute;stica. 2019;31\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30860/0013\u003c/span\u003e\u003cspan address=\"10.30860/0013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBeauchamp TL, Childress JF. \u003cem\u003ePrinciples of Biomedical Ethics\u003c/em\u003e. 8th ed. Oxford University Press; 2019. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://global.oup.com/ushe/product/principles-of-biomedical-ethics-9780190640873\u003c/span\u003e\u003cspan address=\"https://global.oup.com/ushe/product/principles-of-biomedical-ethics-9780190640873\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAndorno R. Human dignity and human rights as a common ground for a global bioethics. J Med Philos. 2009;34(3):223\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/jmp/jhp023\u003c/span\u003e\u003cspan address=\"10.1093/jmp/jhp023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMennella C, Maniscalco U, De Pietro G, Esposito M. Ethical and regulatory challenges of AI technologies in healthcare: A narrative review. Heliyon. 2024;10(4):e26297. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.heliyon.2024.e26297\u003c/span\u003e\u003cspan address=\"10.1016/j.heliyon.2024.e26297\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGeppert CMA, Shelton W. Health care ethics committees as mediators of social values and the culture of medicine. AMA J Ethics. 2016;18(5):534\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1001/journalofethics.2016.18.5\u003c/span\u003e\u003cspan address=\"10.1001/journalofethics.2016.18.5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAucouturier E, Grinbaum A. Training bioethics professionals in AI ethics: A framework. J Law Med Ethics. 2025;53(1):176\u0026ndash;83. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1017/jme.2025.57\u003c/span\u003e\u003cspan address=\"10.1017/jme.2025.57\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eThe Laboratory of Computer Science. DXplain Project [Internet]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.mghlcs.org/projects/dxplain\u003c/span\u003e\u003cspan address=\"http://www.mghlcs.org/projects/dxplain\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWolters Kluwer. Information is the Best Medicine [Internet]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.wolterskluwer.com/en/know/ent-int-clinical-leadership-resources\u003c/span\u003e\u003cspan address=\"https://www.wolterskluwer.com/en/know/ent-int-clinical-leadership-resources\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eValdivia Navarro F, P\u0026eacute;rez Rosa D. Sistema de soporte a las decisiones cl\u0026iacute;nicas relacionadas con el diagn\u0026oacute;stico precoz de enfermedades. \u003cem\u003eRev Cubana Inf M\u0026eacute;dica\u003c/em\u003e. 2016;8:533\u0026ndash;44. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://scielo.sld.cu/scielo.php?script=sci_arttext\u0026amp;pid=S1684-18592016000300006\u003c/span\u003e\u003cspan address=\"http://scielo.sld.cu/scielo.php?script=sci_arttext\u0026amp;pid=S1684-18592016000300006\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNational Center for Biotechnology Information. FUTURE AI \u0026ndash; Bookshelf [Internet]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/books/NBK610549/?term=FUTURE%20AI\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/books/NBK610549/?term=FUTURE%20AI\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Ethics, Clinical, Decision Support Systems, Clinical, Artificial Intelligence, Palliative Care, Terminal Care, Patient Autonomy","lastPublishedDoi":"10.21203/rs.3.rs-7078293/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7078293/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Clinical decision-making at the end of life often involves complex ethical dilemmas that require balancing core principles such as patient autonomy, beneficence, justice, and non-maleficence. Although various clinical decision-support tools exist, few incorporate bioethical deliberation, argument traceability, and multiperspective participation. To address this gap, BIOETHICARE 360° was developed as a computational system designed to facilitate ethical analysis in terminal clinical cases. This study aimed to evaluate the software’s ability to identify ethical dilemmas, propose context-sensitive recommendations, and support real-time deliberative processes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults.\u003c/strong\u003e The system was applied to thirteen synthetic clinical cases representing common scenarios in palliative care and morally complex situations. BIOETHICARE 360° enabled the structured registration of clinical, sociocultural, and ethical information from the perspectives of the medical team, the patient's family, and the bioethics committee. Through its functional modules, the system performed ethical principle weighting, generated graphical visualizations of dissent, identified potential algorithmic biases, and activated narrative-based deliberations through a conversational assistant. In five of the thirteen cases, the primary dilemma identified involved conflicts between the patient’s advance directives and the family’s intention to pursue invasive interventions. In 77% of cases, the final recommendation was the early integration of palliative care. The system provided reasoned responses that included legal analysis, proposals for empathetic communication, and the automated generation of informed consent documents, demonstrating both technical consistency and ethical sensitivity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions.\u003c/strong\u003e BIOETHICARE 360° represents a methodological and technological advancement in ethical support for clinical practice. Its modular architecture, ability to integrate multiple stakeholder perspectives, and deliberative engine grounded in bioethical principles enhance the transparency, traceability, and quality of decision-making in critical clinical settings. Beyond its clinical utility, the tool holds significant educational value by facilitating reflective processes within interdisciplinary teams. This study suggests that the use of artificial intelligence in medical ethics may offer a viable and trustworthy strategy for strengthening decision-making at the end of life.\u003c/p\u003e","manuscriptTitle":"BIOETHICARE 360O: A Deliberative Software for Ethical Computational Support in Complex Clinical Decision-Making","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-30 07:28:28","doi":"10.21203/rs.3.rs-7078293/v1","editorialEvents":[{"type":"communityComments","content":2}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3a984a58-43d8-4cfe-b334-e6d4c79a4adf","owner":[],"postedDate":"July 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-08-25T04:38:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-30 07:28:28","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7078293","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7078293","identity":"rs-7078293","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}