Publications

84 Publications visible to you, out of a total of 84

Abstract (Expand)

We describe the creation of GRASCCO, a novel German-language corpus composed of some 60 clinical documents with more than.43,000 tokens. GRASCCO is a synthetic corpus resulting from a series of alienation steps to obfuscate privacy-sensitive information contained in real clinical documents, the true origin of all GRASCCO texts. Therefore, it is publicly shareable without any legal restrictions We also explore whether this corpus still represents common clinical language use by comparison with a real (non-shareable) clinical corpus we developed as a contribution to the Medical Informatics Initiative in Germany (MII) within the SMITH consortium. We find evidence that such a claim can indeed be made.

Editor:

Date Published: 17th Aug 2022

Publication Type: InProceedings

Abstract (Expand)

We describe the creation of GRASCCO, a novel German-language corpus composed of some 60 clinical documents with more than.43,000 tokens. GRASCCO is a synthetic corpus resulting from a series of alienation steps to obfuscate privacy-sensitive information contained in real clinical documents, the true origin of all GRASCCO texts. Therefore, it is publicly shareable without any legal restrictions We also explore whether this corpus still represents common clinical language use by comparison with a real (non-shareable) clinical corpus we developed as a contribution to the Medical Informatics Initiative in Germany (MII) within the SMITH consortium. We find evidence that such a claim can indeed be made.

Authors: Luise Modersohn, Stefan Schulz, Christina Lohr, Udo Hahn

Date Published: 1st Aug 2022

Publication Type: InCollection

Abstract (Expand)

BACKGROUND: The secondary use of deidentified but not anonymized patient data is a promising approach for enabling precision medicine and learning health care systems. In most national jurisdictions (e.g., in Europe), this type of secondary use requires patient consent. While various ethical, legal, and technical analyses have stressed the opportunities and challenges for different types of consent over the past decade, no country has yet established a national consent standard accepted by the relevant authorities. METHODS: A working group of the national Medical Informatics Initiative in Germany conducted a requirements analysis and developed a GDPR-compliant broad consent standard. The development included consensus procedures within the Medical Informatics Initiative, a documented consultation process with all relevant stakeholder groups and authorities, and the ultimate submission for approval via the national data protection authorities. RESULTS: This paper presents the broad consent text together with a guidance document on mandatory safeguards for broad consent implementation. The mandatory safeguards comprise i) independent review of individual research projects, ii) organizational measures to protect patients from involuntary disclosure of protected information, and iii) comprehensive information for patients and public transparency. This paper further describes the key issues discussed with the relevant authorities, especially the position on additional or alternative consent approaches such as dynamic consent. DISCUSSION: Both the resulting broad consent text and the national consensus process are relevant for similar activities internationally. A key challenge of aligning consent documents with the various stakeholders was explaining and justifying the decision to use broad consent and the decision against using alternative models such as dynamic consent. Public transparency for all secondary use projects and their results emerged as a key factor in this justification. While currently largely limited to academic medicine in Germany, the first steps for extending this broad consent approach to wider areas of application, including smaller institutions and medical practices, are currently under consideration.

Authors: Sven Zenker, Daniel Strech, Kristina Ihrig, Roland Jahns, Gabriele Müller, Christoph Schickhardt, Georg Schmidt, Ronald Speer, Eva Winkler, Sebastian Graf von Kielmansegg, Johannes Drepper

Date Published: 1st Jul 2022

Publication Type: Journal article

Abstract (Expand)

Despite remarkable advances in the development of language resources over the recent years, there is still a shortage of annotated, publicly available corpora covering (German) medical language. With the initial release of the German Guideline Program in Oncology NLP Corpus (GGPONC), we have demonstrated how such corpora can be built upon clinical guidelines, a widely available resource in many natural languages with a reasonable coverage of medical terminology. In this work, we describe a major new release for GGPONC. The corpus has been substantially extended in size and re-annotated with a new annotation scheme based on SNOMED CT top level hierarchies, reaching high inter-annotator agreement (γ=.94). Moreover, we annotated elliptical coordinated noun phrases and their resolutions, a common language phenomenon in (not only German) scientific documents. We also trained BERT-based named entity recognition models on this new data set, which achieve high performance on short, coarse-grained entity spans (F1=.89), while the rate of boundary errors increases for long entity spans. GGPONC is freely available through a data use agreement. The trained named entity recognition models, as well as the detailed annotation guide, are also made publicly available.

Editor: Nicoletta Calzolari, Frédéric Béchet, Philippe Blache, Khalid Choukri, Christopher Cieri, Thierry Declerck, Sara Goggi, Hitoshi Isahara, Bente Maegaard, Joseph Mariani, Hélène Mazo, Jan Odijk, Stelios Piperidis

Date Published: 19th Jun 2022

Publication Type: Journal article

Abstract (Expand)

BACKGROUND: Clinical decision support systems often adopt and operationalize existing clinical practice guidelines leading to higher guideline availability, increased guideline adherence, and data integration. Most of these systems use an internal state-based model of a clinical practice guideline to derive recommendations but do not provide the user with comprehensive insight into the model. OBJECTIVE: Here we present a novel approach based on dynamic guideline visualization that incorporates the individual patient’s current treatment context. METHODS: We derived multiple requirements to be fulfilled by such an enhanced guideline visualization. Using business process and model notation as the representation format for computer-interpretable guidelines, a combination of graph-based representation and logical inferences is adopted for guideline processing. A context-specific guideline visualization is inferred using a business rules engine. RESULTS: We implemented and piloted an algorithmic approach for guideline interpretation and processing. As a result of this interpretation, a context-specific guideline is derived and visualized. Our implementation can be used as a software library but also provides a representational state transfer interface. Spring, Camunda, and Drools served as the main frameworks for implementation. A formative usability evaluation of a demonstrator tool that uses the visualization yielded high acceptance among clinicians. CONCLUSIONS: The novel guideline processing and visualization concept proved to be technically feasible. The approach addresses known problems of guideline-based clinical decision support systems. Further research is necessary to evaluate the applicability of the approach in specific medical use cases.

Authors: Jonas Fortmann, Marlene Lutz, Cord Spreckelsen

Date Published: 1st Jun 2022

Publication Type: Journal article

Abstract (Expand)

Abstract Background In recent years, data-driven medicine has gained increasing importance in terms of diagnosis, treatment, and research due to the exponential growth of health care data. However, data protection regulations prohibit data centralisation for analysis purposes because of potential privacy risks like the accidental disclosure of data to third parties. Therefore, alternative data usage policies, which comply with present privacy guidelines, are of particular interest. Objective We aim to enable analyses on sensitive patient data by simultaneously complying with local data protection regulations using an approach called the Personal Health Train (PHT), which is a paradigm that utilises distributed analytics (DA) methods. The main principle of the PHT is that the analytical task is brought to the data provider and the data instances remain in their original location. Methods In this work, we present our implementation of the PHT paradigm, which preserves the sovereignty and autonomy of the data providers and operates with a limited number of communication channels. We further conduct a DA use case on data stored in three different and distributed data providers. Results We show that our infrastructure enables the training of data models based on distributed data sources. Conclusion Our work presents the capabilities of DA infrastructures in the health care sector, which lower the regulatory obstacles of sharing patient data. We further demonstrate its ability to fuel medical science by making distributed data sets available for scientists or health care practitioners.

Authors: Sascha Welten, Yongli Mou, Laurenz Neumann, Mehrshad Jaberansary, Yeliz Yediel Ucer, Toralf Kirsten, Stefan Decker, Oya Beyan

Date Published: 1st Jun 2022

Publication Type: Journal article

Abstract (Expand)

Introduction: To date cranial development has only been described by analyzing occipitofrontal circumference (OFC). More precise methods of determining head measurements have not been widely adopted. The use of additional measurements has the potential to better account for the three-dimensional structure of the head. Our aim was to put forward centile curves of such measurements for gestational age along with a compound head volume index. Methods: We created generalized additive models for location, scale, and shape of two ear-to-ear distances (EED), transfontanellar (fEED) and transvertical (vEED), from birth anthropometric data. Same was done for OFC, crown-heel length, and birth weight to allow for comparison of our models with growth charts by Voigt et al. and Fenton and Kim. Results: Growth charts and tables of LMS parameters for fEED and vEED were derived from 6,610 patients admitted to our NICU and 625 healthy term newborns. With increasing gestational age EEDs increase about half as fast compared to OFC in absolute terms, their relative growths are fairly similar. Discussion: Differences to the charts by Fenton and Kim are minute. Tape measurements, such as fEED or vEED can be added to routine anthropometry at little extra costs. These charts may be helpful for following and evaluating head sizes and growth of preterm and term infants in three dimensions.

Authors: Nancy Arnold, Rudolf Georg Ascherl, Ulrich Herbert Thome

Date Published: 1st May 2022

Publication Type: Journal article

Powered by
(v.1.13.0-master)
Copyright © 2008 - 2021 The University of Manchester and HITS gGmbH
Institute for Medical Informatics, Statistics and Epidemiology, University of Leipzig

By continuing to use this site you agree to the use of cookies