Bernardo Neves
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Bernardo Neves has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Mario J. Silva
University of Lisbon
5 shared publications - Nuno André da Silva
Hospital da Luz Learning Health
4 shared publications - José Maria Moreira
Hospital da Luz Learning Health
3 shared publications Francisca Leite
Hospital da Luz Learning Health, Luz Saúde, Lisboa, Portugal
2 shared publications- Francisca Pais Leite
Hospital da Luz Learning Health, Luz Saúde, Lisboa, Portugal
2 shared publications Inês Mota
Hospital da Luz Learning Health, Luz Saúde, Lisboa, Portugal
2 shared publications
Latest Publications
- Open Access
- 24 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
- Vol. 7 (4) , 1-29
AI Assistance in Medical Decision-Making: The Role of Recommendations and Explanations in Simulated Clinical Cases
Healthcare is a sector that can greatly benefit from AI due to the critical nature of medical decisions and the demand for accuracy and efficiency. Advances in AI algorithms have shown promise in diagnosing conditions at early stages. To improve the transparency of such systems, Explainable AI (XAI) techniques have been explored. However, if not properly integrated, AI and XAI can introduce new challenges, such as reduced cognitive engagement or overreliance on AI-system outputs. While prior research has mainly focused on isolated tasks, such as AI-assisted diagnosis, this study takes a holistic approach by examining the impact of AI recommendations and explanations across multiple sequential decision points within clinical case scenarios. Thirty-four medical students completed two simulated clinical cases with varying levels of AI assistance. The scenarios reflected realistic workflows, requiring participants to prescribe diagnostic exams and make final diagnoses. Results indicate that AI recommendations and explanations can enhance medical performance but do not consistently improve efficiency. In simpler cases, AI assistance supported appropriate exam prescription and reduced decision-making time. In more complex cases, explanations significantly improved diagnostic accuracy but also led to higher resource use, with participants ordering more unnecessary exams.
- 24 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
- Vol. 7 (4) , 1-29
Open AccessAI Assistance in Medical Decision-Making: The Role of Recommendations and Explanations in Simulated Clinical Cases
Healthcare is a sector that can greatly benefit from AI due to the critical nature of medical decisions and the demand for accuracy and efficiency. Advances in AI algorithms have shown promise in diagnosing conditions at early stages. To improve the transparency of such systems, Explainable AI (XAI) techniques have been explored. However, if not properly integrated, AI and XAI can introduce new challenges, such as reduced cognitive engagement or overreliance on AI-system outputs. While prior research has mainly focused on isolated tasks, such as AI-assisted diagnosis, this study takes a holistic approach by examining the impact of AI recommendations and explanations across multiple sequential decision points within clinical case scenarios. Thirty-four medical students completed two simulated clinical cases with varying levels of AI assistance. The scenarios reflected realistic workflows, requiring participants to prescribe diagnostic exams and make final diagnoses. Results indicate that AI recommendations and explanations can enhance medical performance but do not consistently improve efficiency. In simpler cases, AI assistance supported appropriate exam prescription and reduced decision-making time. In more complex cases, explanations significantly improved diagnostic accuracy but also led to higher resource use, with participants ordering more unnecessary exams.
- Open Access
- 1 April 2026
- Published by Ordem dos Medicos in Acta Medica Portuguesa
- Vol. 39 (4) , 291-298
Disrupções na Força de Trabalho e Redefinição dos Papéis Clínicos: O Estado da Arte da Inteligência Artificial em Saúde
O nosso objetivo foi analisar o impacto da integração da inteligência artificial na saúde, com foco nas alterações da força de trabalho, redefinição de papéis clínicos e implicações éticas e organizacionais. Foram realizadas pesquisas nas bases de dados PubMed, Scopus e Web of Science Core Collection, bem como em documentos institucionais, a partir das quais se desenvolveu uma revisão narrativa centrada na aplicação da inteligência artificial em contextos clínicos, administrativos e educacionais e nas suas implicações para a força de trabalho em saúde. A evidência mostra maturidade de implementação em áreas como imagiologia médica e sistemas de apoio à decisão clínica. As aplicações administrativas expandem-se em registos, faturação e planeamento. Entre os benefícios descritos destacam-se maior precisão diagnóstica, personalização terapêutica, eficiência organizacional e melhoria da segurança do doente. Entre os riscos identificados salientam-se a substituição de funções, erosão de competências, precarização laboral e sobrecarga cognitiva quando a tecnologia é mal integrada. A adoção sustentável da inteligência artificial em saúde requer avaliação de desempenho em cenários do mundo real, interoperabilidade, transparência algorítmica e governação ética robusta. A integração deve ser sociotécnica, equilibrando inovação tecnológica com preservação de competências humanas essenciais, proteção de valores éticos e estabilidade da força de trabalho.
- 1 April 2026
- Published by Ordem dos Medicos in Acta Medica Portuguesa
- Vol. 39 (4) , 291-298
Open AccessDisrupções na Força de Trabalho e Redefinição dos Papéis Clínicos: O Estado da Arte da Inteligência Artificial em Saúde
O nosso objetivo foi analisar o impacto da integração da inteligência artificial na saúde, com foco nas alterações da força de trabalho, redefinição de papéis clínicos e implicações éticas e organizacionais. Foram realizadas pesquisas nas bases de dados PubMed, Scopus e Web of Science Core Collection, bem como em documentos institucionais, a partir das quais se desenvolveu uma revisão narrativa centrada na aplicação da inteligência artificial em contextos clínicos, administrativos e educacionais e nas suas implicações para a força de trabalho em saúde. A evidência mostra maturidade de implementação em áreas como imagiologia médica e sistemas de apoio à decisão clínica. As aplicações administrativas expandem-se em registos, faturação e planeamento. Entre os benefícios descritos destacam-se maior precisão diagnóstica, personalização terapêutica, eficiência organizacional e melhoria da segurança do doente. Entre os riscos identificados salientam-se a substituição de funções, erosão de competências, precarização laboral e sobrecarga cognitiva quando a tecnologia é mal integrada. A adoção sustentável da inteligência artificial em saúde requer avaliação de desempenho em cenários do mundo real, interoperabilidade, transparência algorítmica e governação ética robusta. A integração deve ser sociotécnica, equilibrando inovação tecnológica com preservação de competências humanas essenciais, proteção de valores éticos e estabilidade da força de trabalho.
- Open Access
- 12 December 2025
- Published by Springer Nature in BMC Health Services Research
- Vol. 26 (1)
Multimorbidity measurement strategies for predicting hospital visits
Multimorbidity is a major challenge for healthcare systems. While multiple methods exist to measure it from electronic health records (EHRs), their relativ
- 12 December 2025
- Published by Springer Nature in BMC Health Services Research
- Vol. 26 (1)
Open AccessMultimorbidity measurement strategies for predicting hospital visits
Multimorbidity is a major challenge for healthcare systems. While multiple methods exist to measure it from electronic health records (EHRs), their relativ
- Open Access
From Clinical Judgment to Large Language Models: Benchmarking Predictive Approaches for Unplanned Hospital Admissions
Background While machine learning (ML) models show strong performance for predicting unplanned hospital visits, their clinical utility relative to physician judgment remains unclear. Large language models (LLMs) offer a promising middle ground, potentially combining algorithmic accuracy with human-interpretable reasoning. Objective To directly compare the predictive performance of physicians, structured ML models, and LLMs for forecasting 30-day emergency department (ED) visits and unplanned hospital admissions under equivalent data conditions. Methods We selected 404 cases from structured EHR data and converted them into synthetic clinical vignettes using GPT-5. Thirty-five physicians evaluated these vignettes, while CLMBR-T (a machine learning model trained on structured EHR data) was applied to the original data. Eight LLMs evaluated the same vignettes. We compared discriminative performance (AUROC, AUPRC), calibration (Brier score, Expected Calibration Error), and confidence-performance relationships across all methods. Results CLMBR-T achieved the highest discriminative performance (AUROC 0.79, 95% CI: 0.75-0.83; AUPRC 0.78, 95% CI: 0.72-0.83), followed by large LLMs (DeepSeek V3, Claude 4.1 Opus, GPT-5; AUROC 0.74). Pooled physicians performed lowest (AUROC 0.65, 95% CI: 0.59-0.70; AUPRC 0.61, 95% CI: 0.54-0.68). However, LLMs showed stronger alignment with physician reasoning (correlation r=0.51-0.65) compared to CLMBR-T (r=0.37). CLMBR-T demonstrated superior confidence calibration with significant confidence-performance correlation (r=0.21, pConclusions ML models trained on structured EHR data outperform both physicians and LLMs in predictive accuracy and confidence calibration, though LLMs achieved competitive zero-shot performance and better approximated human clinical reasoning. These findings suggest hybrid approaches combining high-performance ML screening with interpretable LLM explanations may optimize both accuracy and clinical adoption. The substantial variability in physician performance highlights limitations of benchmarking against “average” clinical judgment.
Open AccessFrom Clinical Judgment to Large Language Models: Benchmarking Predictive Approaches for Unplanned Hospital Admissions
Background While machine learning (ML) models show strong performance for predicting unplanned hospital visits, their clinical utility relative to physician judgment remains unclear. Large language models (LLMs) offer a promising middle ground, potentially combining algorithmic accuracy with human-interpretable reasoning. Objective To directly compare the predictive performance of physicians, structured ML models, and LLMs for forecasting 30-day emergency department (ED) visits and unplanned hospital admissions under equivalent data conditions. Methods We selected 404 cases from structured EHR data and converted them into synthetic clinical vignettes using GPT-5. Thirty-five physicians evaluated these vignettes, while CLMBR-T (a machine learning model trained on structured EHR data) was applied to the original data. Eight LLMs evaluated the same vignettes. We compared discriminative performance (AUROC, AUPRC), calibration (Brier score, Expected Calibration Error), and confidence-performance relationships across all methods. Results CLMBR-T achieved the highest discriminative performance (AUROC 0.79, 95% CI: 0.75-0.83; AUPRC 0.78, 95% CI: 0.72-0.83), followed by large LLMs (DeepSeek V3, Claude 4.1 Opus, GPT-5; AUROC 0.74). Pooled physicians performed lowest (AUROC 0.65, 95% CI: 0.59-0.70; AUPRC 0.61, 95% CI: 0.54-0.68). However, LLMs showed stronger alignment with physician reasoning (correlation r=0.51-0.65) compared to CLMBR-T (r=0.37). CLMBR-T demonstrated superior confidence calibration with significant confidence-performance correlation (r=0.21, pConclusions ML models trained on structured EHR data outperform both physicians and LLMs in predictive accuracy and confidence calibration, though LLMs achieved competitive zero-shot performance and better approximated human clinical reasoning. These findings suggest hybrid approaches combining high-performance ML screening with interpretable LLM explanations may optimize both accuracy and clinical adoption. The substantial variability in physician performance highlights limitations of benchmarking against “average” clinical judgment.
- Open Access
Predicting Hospital Admissions Using Pretrained EHR Embeddings: External Evaluation and Insights on Local Vocabulary Adaptation
Purpose Unplanned hospital admissions impose substantial strain on healthcare systems, yet predictive models for these events remain underexplored in practice. This study evaluates whether publicly available pretrained transformer-based embeddings, developed on an external health system, can improve prediction of hospital admissions—including unplanned cases—when applied to a different institution with sparser data and a distinct medical vocabulary. Methods We performed a retrospective cohort study using structured EHR data from 200,000 adult patients (2007–2023) at a Portuguese hospital, standardized to the OMOP Common Data Model. Four 30-day outcomes were predicted: emergency department visits, hospital admissions, unplanned admissions, and readmissions. Three modeling approaches were compared: (1) clinically curated handcrafted features, (2) frequency-based representations of all recorded OMOP concepts, and (3) pretrained CLMBR-T embeddings generated from longitudinal OMOP data of 2.57 million patients in a U.S. hospital system. Performance was assessed on held-out patients using AUROC, AUPRC, and calibration metrics, with additional analysis of the impact of vocabulary overlap between pretraining and local datasets. Results Pretrained embeddings achieved the highest discrimination for all outcomes, particularly for unplanned admissions (AUROC 0.877 vs. 0.770 for counts). Gains were greatest for rarer outcomes and patients with richer clinical histories. Despite only 58% overlap with local vocabulary and substantially fewer events per patient than in pretraining, embeddings transferred effectively, indicating generalizable temporal patterns. Calibration was poorer than simpler models, necessitating post-hoc recalibration before deployment. Conclusion Pretrained OMOP-based EHR embeddings can substantially improve prediction of hospital and unplanned admissions in data- and resource-limited settings, even with partial vocabulary overlap. These findings support their use for rapid, cost-effective deployment of clinically meaningful predictive models, provided local recalibration and workflow integration are addressed. Highlights Superior cross-institution performance – Pretrained EHR embeddings from Stanford Medicine achieved AUROC 0.877 for unplanned admissions, 0.814 for hospital admissions, 0.782 for ED visits, and 0.923 for readmissions in a Portuguese hospital, outperforming count-based (0.770, 0.767, 0.744, 0.922) and handcrafted feature models across all tasks. Largest gains for rare, unpredictable events – For unplanned admissions (0.4% prevalence), embeddings nearly tripled AUPRC compared to counts (0.037 vs. 0.011) and improved AUROC by 0.107, with performance continuing to scale with more training data, unlike baselines. Effective under substantial domain shift – Strong transferability observed despite only 58% vocabulary overlap and markedly different patient populations, coding distributions, and event density (707 vs. 71 events per patient). Benefit increases with richer patient histories – Performance advantage of embeddings widened in patients with higher code volumes; AUROC for hospital admissions rose from 0.694 in the lowest quartile (Q1) to 0.870 in the highest (Q4), outperforming counts by up to 0.105 in Q4. Actionable guidance for adoption – Hospitals with sparse data can achieve rapid, cost-effective deployment of predictive models using external embeddings, especially for rare outcomes, if paired with local fine-tuning and post-hoc recalibration to ensure accurate risk estimation before clinical use.
Open AccessPredicting Hospital Admissions Using Pretrained EHR Embeddings: External Evaluation and Insights on Local Vocabulary Adaptation
Purpose Unplanned hospital admissions impose substantial strain on healthcare systems, yet predictive models for these events remain underexplored in practice. This study evaluates whether publicly available pretrained transformer-based embeddings, developed on an external health system, can improve prediction of hospital admissions—including unplanned cases—when applied to a different institution with sparser data and a distinct medical vocabulary. Methods We performed a retrospective cohort study using structured EHR data from 200,000 adult patients (2007–2023) at a Portuguese hospital, standardized to the OMOP Common Data Model. Four 30-day outcomes were predicted: emergency department visits, hospital admissions, unplanned admissions, and readmissions. Three modeling approaches were compared: (1) clinically curated handcrafted features, (2) frequency-based representations of all recorded OMOP concepts, and (3) pretrained CLMBR-T embeddings generated from longitudinal OMOP data of 2.57 million patients in a U.S. hospital system. Performance was assessed on held-out patients using AUROC, AUPRC, and calibration metrics, with additional analysis of the impact of vocabulary overlap between pretraining and local datasets. Results Pretrained embeddings achieved the highest discrimination for all outcomes, particularly for unplanned admissions (AUROC 0.877 vs. 0.770 for counts). Gains were greatest for rarer outcomes and patients with richer clinical histories. Despite only 58% overlap with local vocabulary and substantially fewer events per patient than in pretraining, embeddings transferred effectively, indicating generalizable temporal patterns. Calibration was poorer than simpler models, necessitating post-hoc recalibration before deployment. Conclusion Pretrained OMOP-based EHR embeddings can substantially improve prediction of hospital and unplanned admissions in data- and resource-limited settings, even with partial vocabulary overlap. These findings support their use for rapid, cost-effective deployment of clinically meaningful predictive models, provided local recalibration and workflow integration are addressed. Highlights Superior cross-institution performance – Pretrained EHR embeddings from Stanford Medicine achieved AUROC 0.877 for unplanned admissions, 0.814 for hospital admissions, 0.782 for ED visits, and 0.923 for readmissions in a Portuguese hospital, outperforming count-based (0.770, 0.767, 0.744, 0.922) and handcrafted feature models across all tasks. Largest gains for rare, unpredictable events – For unplanned admissions (0.4% prevalence), embeddings nearly tripled AUPRC compared to counts (0.037 vs. 0.011) and improved AUROC by 0.107, with performance continuing to scale with more training data, unlike baselines. Effective under substantial domain shift – Strong transferability observed despite only 58% vocabulary overlap and markedly different patient populations, coding distributions, and event density (707 vs. 71 events per patient). Benefit increases with richer patient histories – Performance advantage of embeddings widened in patients with higher code volumes; AUROC for hospital admissions rose from 0.694 in the lowest quartile (Q1) to 0.870 in the highest (Q4), outperforming counts by up to 0.105 in Q4. Actionable guidance for adoption – Hospitals with sparse data can achieve rapid, cost-effective deployment of predictive models using external embeddings, especially for rare outcomes, if paired with local fine-tuning and post-hoc recalibration to ensure accurate risk estimation before clinical use.
- Open Access
- 6 May 2025
- Published by Springer Nature in Research Square
Multimorbidity Measurement Strategies for Predicting Hospital Visits
Introduction : Multimorbidity is a known driver of adverse clinical outcomes and increased healthcare utilization. Although data from Electronic Health Records (EHRs) might enable risk prediction efforts, optimal multimorbidity measurement methods remain unclear. We compared multimorbidity measurement approaches to predict healthcare outcomes beyond demographics and prior healthcare utilization. Study Design and Methods : This is a retrospective cohort study using 15-year EHR data (650,651 patients, 9.4 million visits). Three phenotyping methods and five multimorbidity indices were evaluated across five outcomes and four time horizons, using logistic regression models. Best performing multimorbidity measurements were further trained with XGBoost and feature importance analysis was performed. Results : Including multimorbidity improved prediction over demographic and prior healthcare utilization features for all evaluated outcomes, particularly for rarer outcomes like inpatient mortality and unplanned admissions. Comorbidity index performance varied by outcome: Charlson Comorbidity Index was superior for mortality and readmissions, while Multimorbidity Weighted Index best predicted unplanned admissions. Clinically-curated phenotyping rules consistently outperformed standardized code-based approaches. Previous admissions and multimorbidity were top predictors of most outcomes while increased outpatient visits correlated with fewer unplanned admissions. Conclusions : Our findings show that multimorbidity measurement should be outcome-specific, with clinically-curated phenotyping methods outperforming standardized coding approaches. Healthcare systems should integrate both multimorbidity and prior utilization patterns for risk prediction models to better identify patients at highest risk for adverse outcomes.
- 6 May 2025
- Published by Springer Nature in Research Square
Open AccessMultimorbidity Measurement Strategies for Predicting Hospital Visits
Introduction : Multimorbidity is a known driver of adverse clinical outcomes and increased healthcare utilization. Although data from Electronic Health Records (EHRs) might enable risk prediction efforts, optimal multimorbidity measurement methods remain unclear. We compared multimorbidity measurement approaches to predict healthcare outcomes beyond demographics and prior healthcare utilization. Study Design and Methods : This is a retrospective cohort study using 15-year EHR data (650,651 patients, 9.4 million visits). Three phenotyping methods and five multimorbidity indices were evaluated across five outcomes and four time horizons, using logistic regression models. Best performing multimorbidity measurements were further trained with XGBoost and feature importance analysis was performed. Results : Including multimorbidity improved prediction over demographic and prior healthcare utilization features for all evaluated outcomes, particularly for rarer outcomes like inpatient mortality and unplanned admissions. Comorbidity index performance varied by outcome: Charlson Comorbidity Index was superior for mortality and readmissions, while Multimorbidity Weighted Index best predicted unplanned admissions. Clinically-curated phenotyping rules consistently outperformed standardized code-based approaches. Previous admissions and multimorbidity were top predictors of most outcomes while increased outpatient visits correlated with fewer unplanned admissions. Conclusions : Our findings show that multimorbidity measurement should be outcome-specific, with clinically-curated phenotyping methods outperforming standardized coding approaches. Healthcare systems should integrate both multimorbidity and prior utilization patterns for risk prediction models to better identify patients at highest risk for adverse outcomes.
- Open Access
- 23 April 2025
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 192 (Pt A) , 110181
Zero-shot learning for clinical phenotyping: Comparing LLMs and rule-based methods
- 23 April 2025
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 192 (Pt A) , 110181
Open AccessZero-shot learning for clinical phenotyping: Comparing LLMs and rule-based methods
- Open Access
- 15 April 2024
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 24 (1) , 1-14
Identifying subgroups in heart failure patients with multimorbidity by clustering and network analysis
This study presents a workflow for identifying and characterizing patients with Heart Failure (HF) and multimorbidity utilizing data from Electronic Health Records. Multimorbidity, the co-occurrence of two or more chronic conditions, poses a significant challenge on healthcare systems. Nonetheless, understanding of patients with multimorbidity, including the most common disease interactions, risk factors, and treatment responses, remains limited, particularly for complex and heterogeneous conditions like HF. We conducted a clustering analysis of 3745 HF patients using demographics, comorbidities, laboratory values, and drug prescriptions. Our analysis revealed four distinct clusters with significant differences in multimorbidity profiles showing differential prognostic implications regarding unplanned hospital admissions. These findings underscore the considerable disease heterogeneity within HF patients and emphasize the potential for improved characterization of patient subgroups for clinical risk stratification through the use of EHR data.
- 15 April 2024
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 24 (1) , 1-14
Open AccessIdentifying subgroups in heart failure patients with multimorbidity by clustering and network analysis
This study presents a workflow for identifying and characterizing patients with Heart Failure (HF) and multimorbidity utilizing data from Electronic Health Records. Multimorbidity, the co-occurrence of two or more chronic conditions, poses a significant challenge on healthcare systems. Nonetheless, understanding of patients with multimorbidity, including the most common disease interactions, risk factors, and treatment responses, remains limited, particularly for complex and heterogeneous conditions like HF. We conducted a clustering analysis of 3745 HF patients using demographics, comorbidities, laboratory values, and drug prescriptions. Our analysis revealed four distinct clusters with significant differences in multimorbidity profiles showing differential prognostic implications regarding unplanned hospital admissions. These findings underscore the considerable disease heterogeneity within HF patients and emphasize the potential for improved characterization of patient subgroups for clinical risk stratification through the use of EHR data.
- Open Access
Predicting Imminent Health Outcomes from Common Lab Results
In recent years, most hospitals have implemented Electronic Health Records to manage and integrate a wide range of medical information, including diagnostics, medication admission and laboratory test results. Certain laboratory variables may serve as indicators of a patient’s clinical deterioration, making laboratory data a valuable tool for identifying high-risk patients. This work introduces a framework for predicting imminent health outcomes (IHO) of multimorbidity patients using laboratory test data. Our cohort includes 322,316 multimorbidity patients that performed laboratory tests in a large teaching hospital between January 2007 and August 2021. Two Imminent Health Outcomes predictive tools were developed. The first considers all patients in the dataset. The second was developed using a subset of patients with Heart Failure (HF) as the main comorbidity (5% of the entire dataset), considering that HF is a highly prevalent syndrome in multimorbidity patients. This predictive model achieved a reasonable predictive performance ( AUROC = 0.718, 95% CI 0.708-0.756, and AUPRC = 0.663, 95% CI 0.630-0.701). C-reactive protein and NT-proBNP are the lab tests that most positively contribute to the prediction of IHO. The IHO predictive tool has the potential to help the medical team identify patients at high-risk of an imminent adverse event, highlighting the laboratory variables that are most contributing to the deterioration of the patient.
Open AccessPredicting Imminent Health Outcomes from Common Lab Results
In recent years, most hospitals have implemented Electronic Health Records to manage and integrate a wide range of medical information, including diagnostics, medication admission and laboratory test results. Certain laboratory variables may serve as indicators of a patient’s clinical deterioration, making laboratory data a valuable tool for identifying high-risk patients. This work introduces a framework for predicting imminent health outcomes (IHO) of multimorbidity patients using laboratory test data. Our cohort includes 322,316 multimorbidity patients that performed laboratory tests in a large teaching hospital between January 2007 and August 2021. Two Imminent Health Outcomes predictive tools were developed. The first considers all patients in the dataset. The second was developed using a subset of patients with Heart Failure (HF) as the main comorbidity (5% of the entire dataset), considering that HF is a highly prevalent syndrome in multimorbidity patients. This predictive model achieved a reasonable predictive performance ( AUROC = 0.718, 95% CI 0.708-0.756, and AUPRC = 0.663, 95% CI 0.630-0.701). C-reactive protein and NT-proBNP are the lab tests that most positively contribute to the prediction of IHO. The IHO predictive tool has the potential to help the medical team identify patients at high-risk of an imminent adverse event, highlighting the laboratory variables that are most contributing to the deterioration of the patient.