[This corrects the article DOI: 10.1007/s41666-025-00216-6.].

Identifying relevant input features which contribute to the output of a clinical prediction model can enhance the model explainability. To allow the explainability to be more personalized, instance-wise feature selection (IWFS) methods can be adopted where features are selected specifically for each input instance. Existing IWFS methods often grapple with feature selection instability, and thus precarious interpretation. As relevant features among the instances in a dataset do overlap, feature grouping tricks have been proposed to regularize the selection, but often at the expense of sacrificing the downstream prediction accuracy. To this end, we propose a novel instance-wise feature grouping method called FlexGPC to achieve robust and stable selection by learning i) flexible representation for feature groups, and ii) flexible combination of feature groups implemented using neural networks. To evaluate the effectiveness of FlexGPC, we explore various feature group combination schemes and conduct extensive experiments for performance comparison using real-world electronic health records (EHR) data. Our experimental results show that FlexGPC outperforms all the SOTA baselines in terms of accuracy and feature selection stability for both downstream mortality and next-admission diagnosis prediction tasks. We also illustrate that computational phenotyping can be achieved at the same time, with the identified feature groups being the potential phenotypes.

Large Language Models (LLMs) are increasingly applied in healthcare and are expected to play an active role in clinical practice. However, their effectiveness for clinical note summarization remains underexplored, and systematic comparisons across different models are lacking. This study addresses this gap by benchmarking 16 generative LLMs from major providers, including OpenAI (GPT), DeepSeek, Meta (LLaMA), Google (Gemma), Mistral (Mixtral), and Alibaba (Qwen), using the MIMIC-IV-Note. Both extractive and abstractive summarization approaches were implemented and evaluated with multiple lexical and semantic metrics, including ROUGE, BLEU, METEOR, COMET, and BERTScore. In addition, processing time, cost, and deployment feasibility were assessed to provide a practical perspective for clinical adoption. The results show that Gemma-3-27B achieved the strongest overall performance in extractive summarization. For abstractive summarization, DeepSeek-R1-70B, Qwen-3-32B, and GPT-4o emerged as leading models. Their relative strengths varied depending on whether lexical overlap, semantic adequacy, or fluency was prioritized. Importantly, larger parameter sizes did not always translate into better outcomes, as smaller models such as LLaMa-3-8B and Gemma-2-9B often produced competitive results with faster runtimes and lower computational costs. This study highlights the trade-offs between performance, efficiency, and deployment context that offers practical insights into model selection for clinical note summarization and informing future integration of LLMs into healthcare workflows.

Abstract This paper presents the Patient-Centered Outcomes Research within the Medical Informatics Initiative (PCOR-MII) project, focusing on the integration of patient-reported outcomes (PROs) into a large-scale national data sharing infrastructure, established in Germany by the Medical Informatics Initiative (MII). PCOR-MII aims to systematically address the interests of various stakeholders in patient-reported health data and three dimensions of clinical utility: (1) prediction, (2) monitoring, and (3) outcome assessment. The project builds upon harmonized technical, data, and compliance environments established at the participating institutions as part of the MII to deploy and roll out software solutions for capturing PROs and making them accessible within local electronic health record (EHR) systems. To overcome interoperability challenges, PCOR-MII is developing a construct-oriented PROM module for the Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR)–based German National Core Dataset. The project applies its approach to three patient populations with distinct characteristics: anorexia nervosa targeting risk prediction (dimension 1), kidney transplantation prioritizing health status and adherence monitoring (dimension 2), and persistent somatic symptoms primarily aimed at assessing and understanding outcomes (dimension 3). With their emphasis on different aspects of PROs, those application areas can serve as blueprints for a broader roll-out. PCOR-MII represents a structured and comprehensive effort to incorporate PROs into a national data infrastructure, promising more precise diagnostics, improved treatment decisions, and the generation of new biomedical insights. We believe that our structured approach may serve as a guiding framework for others aiming to implement PROs in diverse healthcare settings.

The COVID-19 pandemic caused rapid shifts in the workflow of many health services, but evidence of how this affected multidisciplinary care settings is limited. In this data study, we propose a process mining approach that utilises timestamped data from electronic health records to compare care provider patterns across pandemic waves. To investigate healthcare patterns during the pandemic, we collected routine events from Scottish hospital episodes in adults with COVID-19 status, generating treatment logs based on care provider input. Conformance checking metrics were used to select the Inductive Miner infrequent (IMi) algorithm for downstream analysis. Visual diagrams from the discovered Petri Nets indicated interactions on provider- and activity-level data subsets. Measures of “cross-log conformance checking” and graph edit distance (GED) further quantified variation in care complexity in adverse subgroups. Our baseline cohort included 1153 patients with COVID-19 linked to 55,212 relevant care provider events. At the conformance checking stage, the IMi model achieved good log fitness (LF¯=0.95) and generalisation (G¯=0.89), but limited precision (PR¯=0.27) across all granularity levels. More structured care procedures were present in Wave 1, compared to limited multidisciplinary involvement in Wave 2. Care activities differed in patients with extended stay (GED=348,PR¯=0.231vsGED=197,PR¯=0.429 in shorter stays). We demonstrated that process mining can be incorporated to investigate differential complexity in patients with COVID-19 and derive fine-grained evidence on shifts in healthcare practice. Future process-driven studies could use clinical oversight to understand operational adherence and driving factors behind service changes during pressured periods.

Pulmonary nodules and nodule characteristics are important indicators of lung nodule malignancy. However, nodule information is often documented as free text in clinical narratives such as radiology reports in electronic health record systems. Natural language processing (NLP) is the key technology to extract and standardize patient information from radiology reports into structured data elements. This study aimed to develop an NLP system using state-of-the-art transformer models to extract pulmonary nodules and associated nodule characteristics from radiology reports. We identified a cohort of 3080 patients who underwent LDCT at the University of Florida health system and collected their radiology reports. We manually annotated 394 reports as the gold standard. We explored eight pretrained transformer models from three transformer architectures including bidirectional encoder representations from transformers (BERT), robustly optimized BERT approach (RoBERTa), and A Lite BERT (ALBERT), for clinical concept extraction, relation identification, and negation detection. We examined general transformer models pretrained using general English corpora, transformer models fine-tuned using a clinical corpus, and a large clinical transformer model, GatorTron, which was trained from scratch using 90 billion words of clinical text. We compared transformer models with two baseline models including a recurrent neural network implemented using bidirectional long short-term memory with a conditional random fields layer and support vector machines. RoBERTa-mimic achieved the best F1-score of 0.9279 for nodule concept and nodule characteristics extraction. ALBERT-base and GatorTron achieved the best F1-score of 0.9737 in linking nodule characteristics to pulmonary nodules. Seven out of eight transformers achieved the best F1-score of 1.0000 for negation detection. Our end-to-end system achieved an overall F1-score of 0.8869. This study demonstrated the advantage of state-of-the-art transformer models for pulmonary nodule information extraction from radiology reports.

With an increased interest in the production of personal health technologies designed to track user data (e.g., nutrient intake, step counts), there is now more opportunity than ever to surface meaningful behavioral insights to everyday users in the form of natural language. This knowledge can increase their behavioral awareness and allow them to take action to meet their health goals. It can also bridge the gap between the vast collection of personal health data and the summary generation required to describe an individual's behavioral tendencies. Previous work has focused on rule-based time-series data summarization methods designed to generate natural language summaries of interesting patterns found within temporal personal health data. We examine recurrent, convolutional, and Transformer-based encoder-decoder models to automatically generate natural language summaries from numeric temporal personal health data. We showcase the effectiveness of our models on real user health data logged in MyFitnessPal(Weber and Achananuparp 2016) and show that we can automatically generate high-quality natural language summaries. Our work serves as a first step towards the ambitious goal of automatically generating novel and meaningful temporal summaries from personal health data.

Emergency Medical Services (EMS) are crucial in delivering timely and effective medical care to patients in need. However, the complex and dynamic nature of operations poses challenges for decision-making processes at strategic, tactical, and operational levels. This paper proposes an action-driven strategy for EMS management, employing a multi-objective optimizer and a simulator to evaluate potential outcomes of decisions. The approach combines historical data with dynamic simulations and multi-objective optimization techniques to inform decision-makers and improve the overall performance of the system. The research focuses on the Friuli Venezia Giulia region in north-eastern Italy. The region encompasses various landscapes and demographic situations that challenge fairness and equity in service access. Similar challenges are faced in other regions with comparable characteristics. The Decision Support System developed in this work accurately models the real-world system and provides valuable feedback and suggestions to EMS professionals, enabling them to make informed decisions and enhance the efficiency and fairness of the system.

Photoplethysmogram (PPG) performs an important role in alarming atrial fibrillation (AF). While the importance of PPG is emphasized, there is insufficient amount of openly available atrial fibrillation PPG data. We propose a U-net-based generative adversarial network (GAN) which synthesize PPG from paired electrocardiogram (ECG). To measure the performance of the proposed GAN, we compared the generated PPG to reference PPG in terms of morphology similarity and also examined its influence on AF detection classifier performance. First, morphology was compared using two different metrics against the reference signal: percent root mean square difference (PRD) and Pearson correlation coefficient. The mean PRD and Pearson correlation coefficient were 27% and 0.94, respectively. Heart rate variability (HRV) of the reference AF ECG and the generated PPG were compared as well. The p-value of the paired t-test was 0.248, indicating that no significant difference was observed between the two HRV values. Second, to validate the generated AF PPG dataset, four different datasets were prepared combining the generated PPG and real AF PPG. Each dataset was used to optimize a classification model while maintaining the same architecture. A test dataset was prepared to test the performance of each optimized model. Subsequently, these datasets were used to test the hypothesis whether the generated data benefits the training of an AF classifier. Comparing the performance metrics of each optimized model, the training dataset consisting of generated and real AF PPG showed a test accuracy result of 0.962, which was close to that of the dataset consisting only of real AF PPG data at0.961. Furthermore, both models yielded the same F1 score of 0.969. Lastly,using only the generated AF PPG dataset resulted in test accuracy of 0.945, indicating that the trained model was capable of generating valuable AF PPG. Therefore, it can be concluded that the generated AF PPG can be used to augment insufficient data. To summarize, this study proposes a GAN-based method to generate atrial fibrillation PPG that can be used for training atrial fibrillation PPG classification models.

Although most pregnancies result in a good outcome, complications are not uncommon and can be associated with serious implications for mothers and babies. Predictive modeling has the potential to improve outcomes through a better understanding of risk factors, heightened surveillance for high-risk patients, and more timely and appropriate interventions, thereby helping obstetricians deliver better care. We identify and study the most important risk factors for four types of pregnancy complications: (i) severe maternal morbidity, (ii) shoulder dystocia, (iii) preterm preeclampsia, and (iv) antepartum stillbirth. We use an Explainable Boosting Machine (EBM), a high-accuracy glass-box learning method, for the prediction and identification of important risk factors. We undertake external validation and perform an extensive robustness analysis of the EBM models. EBMs match the accuracy of other black-box ML methods, such as deep neural networks and random forests, and outperform logistic regression, while being more interpretable. EBMs prove to be robust. The interpretability of the EBM models reveal surprising insights into the features contributing to risk (e.g., maternal height is the second most important feature for shoulder dystocia) and may have potential for clinical application in the prediction and prevention of serious complications in pregnancy.