Alexander Johannes Wiederhold

Introduction: Evidence-based development of healthcare innovations requires efficient access to existing knowledge. The Innovation Committee of the Federal Joint Committee (Gemeinsamer Bundesausschuss, G-BA) has funded more than 800 projects since 2016, yet the heterogeneity of topics and terminology limits the effectiveness of the current keyword-based search interface. Methods: To address vocabulary-mismatch problems and support more nuanced information needs, we developed and evaluated a semantic search tool using vector-based natural-language retrieval. The application embeds project descriptions and user queries and computes similarity directly in the browser, enabling privacy-preserving, offline operation. In a controlled comparative study with three experienced users, we assessed efficiency and effectiveness across four standardized search tasks. Results: Although based on a small sample size, the results indicate that semantic retrieval can meaningfully improve the discovery of projects funded by the Innovation Committee and support evidence-based research workflows. Conclusion: Future work should validate these findings in larger cohorts and explore enhanced ranking strategies.

Hypomimia, the reduced facial expressiveness commonly seen in Parkinson’s disease (PD), is clinically relevant, yet routine assessment usually relies on a single ordinal rating item in major rating scales such as the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS). This study aimed to develop and iteratively refine a wearable, cap-mounted device for hypomimia recording during routine inpatient ON/OFF assessments and to examine its usability and technical feasibility. A staged co-development process with four prototypes was conducted. After clinician workflow requirements had been incorporated, Prototype 3 was evaluated during clinical routine using the System Usability Scale (SUS) and short free-text feedback. These findings informed a patient-centered redesign that resulted in Prototype 4. Bench testing confirmed stable operation, robust housing, and reliable closed Wi-Fi streaming. Usability improved from moderate in Prototype 3 (mean SUS 44.7, SD 18.2; n = 15) to good in Prototype 4 (mean SUS 80.0, SD 12.6; n = 50). Mixed-effects regression showed an increase of approximately 33 SUS points for the redesigned device (95% CI 27.7–39.4; p < 0.001). Mean effective sampling rate was 4.79 frames/s (SD 0.33), with face detection in 95.4% of frames. These feasibility and usability results indicate that technical robustness and patient-centered design are critical prerequisites for later clinical validation and support the integration of a wearable hypomimia capture device into routine assessments.

Background: Patient stratification is crucial for advancing personalized medicine yet is complicated due to the fragmented and variable nature of healthcare data. The Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) addresses the challenges regarding the data by offering a standardized framework for data integration across diverse sources and configurations. Recent advancements in machine learning, particularly transformer-based models such as Bidirectional Encoder Representations from Transformers (BERT), have demonstrated significant potential in extracting deep patient representations from electronic health records. Methods: This study assesses the efficacy of BERT-based patient representation learning using OMOP CDM data for patient stratification. We harmonize originally incompatible datasets, including the established MIMIC-IV-2.2 and lung cancer data from the German cancer registry Schleswig-Holstein, within the OMOP CDM framework. BERT is pre-trained on the MIMIC-IV-2.2 dataset, and the derived representations are utilized to generate patient embeddings from the cancer registry (test) data. We employ k-means clustering on the embeddings to stratify patient subgroups. To evaluate whether the embeddings are useful for clustering, we divide the original cancer registry data into the corresponding groups and conduct survival analyses on selected columns for each cluster. The clustering method’s effectiveness is assessed by comparing survival models trained on these clusters with naïve clusters derived only from the original dataset. In addition, we included a clinical expert review, in which a physician assessed the resulting cluster assignments for clinical plausibility and interpretability. Results: Our approach effectively identifies patient similarities across datasets and allowed for efficient patient stratification. Survival analyses show varied performance depending on the model and cluster characteristics, with up to a 11% improvement over naïve k-means clusters, demonstrating the benefits of transfer learning. For some groups of patients, the corresponding accuracy of the survival analysis increased by up to 7%, emphasizing the value of stratifying homogeneous subgroups. Conclusion: Utilizing standardized data and transformer-based foundation models to generate patient embeddings demonstrates effective knowledge transfer between two vastly different datasets and enables the identification of groups in which the accuracy of survival analysis can be significantly increased.

Hypomimia, a reduced facial expressiveness in Parkinson’s disease, is clinically relevant yet difficult to rate consistently. We evaluated the usability of CAPtion, an iOS application that supports standardized, session-based retrospective hypomimia ratings from camera-based facial recordings, enabling structured scoring across multiple independent raters. Six Parkinson’s specialists completed one predefined rating session and subsequently filled out the System Usability Scale. It achieved a mean SUS score of 91.7 (±6.8), indicating excellent usability and suitability for scaling to larger inter-rater variability studies.

Current machine learning approaches focusing on motor symptoms in Parkinson’s disease are commonly trained on small datasets and often lack generalizability from developmental setups to clinical applications. Foundation models using large, unlabeled datasets of healthy participants through self-supervised learning appear attractive for such setups with limited samples, despite the potential impact of motoric symptoms. Acting as an exemplar, this study aims to evaluate the robustness of fine-tuned models in recognizing movements related to motor examinations across datasets and recording setups. Accelerometer data of 51 participants with Parkinson’s disease in three different training and fine-tuning setups were used to tailor the general model to the disease. Training the model on pre-trained weights, both partially (F₁ = 0.70) and fully (F₁ = 0.69), statistically significantly outperformed training the model from scratch (F₁ = 0.55) in a nested cross-validation. For evaluation, the model’s ability to process data recorded from 24 patients in clinic was tested. The models achieved lower mean F₁ scores of 0.33 (train from scratch), 0.43 for full, and 0.48 for partial fine-tuning, but demonstrated improved generalizability and robustness regarding the orientation of sensors compared to training from scratch. Utilizing foundation models for accelerometer data trained on healthy participants and fine-tuned for clinical applications in movement disorders appears as an effective strategy for optimized generalizability with small datasets.

This study investigates the potential of multimodal neural networks to convert data from unstructured paper-based medical documents into a structured format. Utilizing advancements in Visual Question Answering, we curated a dataset from neurological documents at the University Medical Center Hamburg-Eppendorf. Different models were assessed for their effectiveness. While models from 2024 showed improved performance, accuracy remained below clinical standards, revealing significant challenges in adapting such technology to complex and heterogeneous medical records. The findings emphasize the need for larger, diverse datasets and ongoing refinement to bridge the gap between current model capabilities and human-level performance, underscoring the complexity of automating data extraction in clinical settings.

During the in-hospital optimization of dopaminergic dosage for Parkinson’s disease, drug-induced dyskinesias emerge as a common side effect. Wrist-worn devices present a substantial opportunity for continuous movement recording and the supportive identification of these dyskinesias. To bridge the gap between dyskinesia assessment and machine learning-enabled detection, the recorded information requires meaningful data representations. This study evaluates and compares two distinct representations of sensor data: a task-dependent, semantically grounded approach and automatically extracted large-scale time-series features. Each representation was assessed on public datasets to identify the best-performing machine learning model and subsequently applied to our own collected dataset to assess generalizability. Data representations incorporating semantic knowledge demonstrated comparable or superior performance to reported works, with peak F₁ scores of 0.68. Generalization to our own dataset from clinical practice resulted in an observed F₁ score of 0.53 using both setups. These results highlight the potential of semantic movement data analysis for dyskinesia detection. Dimensionality reduction in accelerometer-based movement data positively impacts performance, and models trained with semantically obtained features avoid overfitting. Expanding cohorts with standardized neurological assessments labeled by medical experts is essential for further improvements.

We introduce a sensor-fitted cap to unobtrusively monitor hypomimia while allowing patients to manage their data autonomously. Thereby, providing insights into symptom progression and treatment response. This patient-centered tool holds promise for routine clinical use and enhancing therapeutic precision.

Background: Clinical routine data derived from university hospitals hold immense value for health-related research on large cohorts. However, using secondary data for hypothesis testing necessitates adherence to scientific, legal (such as the General Data Protection Regulation, federal and state protection legislations), technical, and administrative requirements. This process is intricate, time-consuming, and susceptible to errors. Objective: This study aims to develop a platform that enables clinicians to use current real-world data for testing research and evaluate advantages and limitations at a large university medical center (542,944 patients in 2022). Methods: We identified requirements from clinical practitioners, conceptualized and implemented a platform based on the existing components, and assessed its applicability in clinical reality quantitatively and qualitatively. Results: The proposed platform was established at the University Medical Center Hamburg-Eppendorf and made 639 forms encompassing 10,629 data elements accessible to all resident scientists and clinicians. Every day, the number of patients rises, and parts of their electronic health records are made accessible through the platform. Qualitatively, we were able to conduct a retrospective analysis of Parkinson disease over 777 patients, where we provide additional evidence for a significantly higher proportion of action tremors in patients with rest tremors (340/777, 43.8%) compared with those without rest tremors (255/777, 32.8%), as determined by a chi-square test (P<.001). Quantitatively, our findings demonstrate increased user engagement within the last 90 days, underscoring clinicians’ increasing adoption of the platform in their regular research activities. Notably, the platform facilitated the retrieval of clinical data from 600,000 patients, emphasizing its substantial added value. Conclusions: This study demonstrates the feasibility of simplifying the use of clinical data to enhance exploration and sustainability in scientific research. The proposed platform emerges as a potential technological and legal framework for other medical centers, providing them with the means to unlock untapped potential within their routine data.

Background: Interpretability and intuitive visualization facilitate medical knowledge generation through big data. In addition, robustness to high-dimensional and missing data is a requirement for statistical approaches in the medical domain. A method tailored to the needs of physicians must meet all the abovementioned criteria. Objective: This study aims to develop an accessible tool for visual data exploration without the need for programming knowledge, adjusting complex parameterizations, or handling missing data. We sought to use statistical analysis using the setting of disease and control cohorts familiar to clinical researchers. We aimed to guide the user by identifying and highlighting data patterns associated with disease and reveal relations between attributes within the data set. Methods: We introduce the attribute association graph, a novel graph structure designed for visual data exploration using robust statistical metrics. The nodes capture frequencies of participant attributes in disease and control cohorts as well as deviations between groups. The edges represent conditional relations between attributes. The graph is visualized using the Neo4j (Neo4j, Inc) data platform and can be interactively explored without the need for technical knowledge. Nodes with high deviations between cohorts and edges of noticeable conditional relationship are highlighted to guide the user during the exploration. The graph is accompanied by a dashboard visualizing variable distributions. For evaluation, we applied the graph and dashboard to the Hamburg City Health Study data set, a large cohort study conducted in the city of Hamburg, Germany. All data structures can be accessed freely by researchers, physicians, and patients. In addition, we developed a user test conducted with physicians incorporating the System Usability Scale, individual questions, and user tasks. Results: We evaluated the attribute association graph and dashboard through an exemplary data analysis of participants with a general cardiovascular disease in the Hamburg City Health Study data set. All results extracted from the graph structure and dashboard are in accordance with findings from the literature, except for unusually low cholesterol levels in participants with cardiovascular disease, which could be induced by medication. In addition, 95% CIs of Pearson correlation coefficients were calculated for all associations identified during the data analysis, confirming the results. In addition, a user test with 10 physicians assessing the usability of the proposed methods was conducted. A System Usability Scale score of 70.5% and average successful task completion of 81.4% were reported. Conclusions: The proposed attribute association graph and dashboard enable intuitive visual data exploration. They are robust to high-dimensional as well as missing data and require no parameterization. The usability for clinicians was confirmed via a user test, and the validity of the statistical results was confirmed by associations known from literature and standard statistical inference.