Publications

37 Publications visible to you, out of a total of 37

Abstract (Expand)

Accurately estimating the length of stay (LOS) of patients admitted to the intensive care unit (ICU) in relation to their health status helps healthcare management allocate appropriate resources and resources and better plan for the future. This paper presents predictive models for the LOS of ICU patients from the MIMIC-IV database based on typical demographic and administrative data, as well as early vital signs and laboratory measurements collected on the first day of ICU stay. The goal of this study was to demonstrate a practical, stepwise approach to predicting patient’s LOS in the ICU using machine learning and early available typical clinical data. The results show that this approach significantly improves the performance of models for predicting actual LOS in a pragmatic framework that includes only data with short stays predetermined by a prior classification.

Authors: Lars Hempel, Sina Sadeghi, Toralf Kirsten

Date Published: 1st Jun 2023

Publication Type: Journal article

Abstract (Expand)

BACKGROUND: Clinical trials, epidemiological studies, clinical registries, and other prospective research projects, together with patient care services, are main sources of data in the medical research domain. They serve often as a basis for secondary research in evidence-based medicine, prediction models for disease, and its progression. This data are often neither sufficiently described nor accessible. Related models are often not accessible as a functional program tool for interested users from the health care and biomedical domains. OBJECTIVE: The interdisciplinary project Leipzig Health Atlas (LHA) was developed to close this gap. LHA is an online platform that serves as a sustainable archive providing medical data, metadata, models, and novel phenotypes from clinical trials, epidemiological studies, and other medical research projects. METHODS: Data, models, and phenotypes are described by semantically rich metadata. The platform prefers to share data and models presented in original publications but is also open for nonpublished data. LHA provides and associates unique permanent identifiers for each dataset and model. Hence, the platform can be used to share prepared, quality-assured datasets and models while they are referenced in publications. All managed data, models, and phenotypes in LHA follow the FAIR principles, with public availability or restricted access for specific user groups. RESULTS: The LHA platform is in productive mode (https://www.health-atlas.de/). It is already used by a variety of clinical trial and research groups and is becoming increasingly popular also in the biomedical community. LHA is an integral part of the forthcoming initiative building a national research data infrastructure for health in Germany.

Authors: T. Kirsten, F. A. Meineke, H. Loeffler-Wirth, C. Beger, A. Uciteli, S. Staubert, M. Lobe, R. Hansel, F. G. Rauscher, J. Schuster, T. Peschel, H. Herre, J. Wagner, S. Zachariae, C. Engel, M. Scholz, E. Rahm, H. Binder, M. Loeffler

Date Published: 3rd Aug 2022

Publication Type: Journal article

Abstract (Expand)

Modern machine and deep learning methods require large datasets to achieve reliable and robust results. This requirement is often difficult to meet in the medical field, due to data sharing limitationsng limitations imposed by privacy regulations or the presence of a small number of patients (e.g., rare diseases). To address this data scarcity and to improve the situation, novel generative models such as Generative Adversarial Networks (GANs) have been widely used to generate synthetic data that mimic real data by representing features that reflect health-related information without reference to real patients. In this paper, we consider several GAN models to generate synthetic data used for training binary (malignant/benign) classifiers, and compare their performances in terms of classification accuracy with cases where only real data are considered. We aim to investigate how synthetic data can improve classification accuracy, especially when a small amount of data is available. To this end, we have developed and implemented an evaluation framework where binary classifiers are trained on extended datasets containing both real and synthetic data. The results show improved accuracy for classifiers trained with generated data from more advanced GAN models, even when limited amounts of original data are available.

Authors: Masoud Abedi, Lars Hempel, Sina Sadeghi, Toralf Kirsten

Date Published: 1st Jul 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)

Clinical research based on data from patient or study data management systems plays an important role in transferring basic findings into the daily practices of physicians. To support study recruitment, diagnostic processes, and risk factor evaluation, search queries for such management systems can be used. Typically, the query syntax as well as the underlying data structure vary greatly between different data management systems. This makes it difficult for domain experts (e.g., clinicians) to build and execute search queries. In this work, the Core Ontology of Phenotypes is used as a general model for phenotypic knowledge. This knowledge is required to create search queries that determine and classify individuals (e.g., patients or study participants) whose morphology, function, behaviour, or biochemical and physiological properties meet specific phenotype classes. A specific model describing a set of particular phenotype classes is called a Phenotype Specification Ontology. Such an ontology can be automatically converted to search queries on data management systems. The methods described have already been used successfully in several projects. Using ontologies to model phenotypic knowledge on patient or study data management systems is a viable approach. It allows clinicians to model from a domain perspective without knowing the actual data structure or query language.

Authors: Christoph Beger, Franz Matthies, Ralph Schäfermeier, Toralf Kirsten, Heinrich Herre, Alexandr Uciteli

Date Published: 1st May 2022

Publication Type: Journal article

Abstract (Expand)

The constant upward movement of data-driven medicine as a valuable option to enhance daily clinical practice has brought new challenges for data analysts to get access to valuable but sensitive data due to privacy considerations. One solution for most of these challenges are Distributed Analytics (DA) infrastructures, which are technologies fostering collaborations between healthcare institutions by establishing a privacy-preserving network for data sharing. However, in order to participate in such a network, a lot of technical and administrative prerequisites have to be made, which could pose bottlenecks and new obstacles for non-technical personnel during their deployment. We have identified three major problems in the current state-of-the-art. Namely, the missing compliance with FAIR data principles, the automation of processes, and the installation. In this work, we present a seamless on-boarding workflow based on a DA reference architecture for data sharing institutions to address these problems. The on-boarding service manages all technical configurations and necessities to reduce the deployment time. Our aim is to use well-established and conventional technologies to gain acceptance through enhanced ease of use. We evaluate our development with six institutions across Germany by conducting a DA study with open-source breast cancer data, which represents the second contribution of this work. We find that our on-boarding solution lowers technical barriers and efficiently deploys all necessary components and is, therefore, indeed an enabler for collaborative data sharing.

Authors: Sascha Welten, Lars Hempel, Masoud Abedi, Yongli Mou, Mehrshad Jaberansary, Laurenz Neumann, Sven Weber, Kais Tahar, Yeliz Ucer Yediel, Matthias Löbe, Stefan Decker, Oya Beyan, Toralf Kirsten

Date Published: 1st Apr 2022

Publication Type: Journal article

Abstract (Expand)

Sharing data is of great importance for research in medical sciences. It is the basis for reproducibility and reuse of already generated outcomes in new projects and in new contexts. FAIR data principles are the basics for sharing data. The Leipzig Health Atlas (LHA) platform follows these principles and provides data, describing metadata, and models that have been implemented in novel software tools and are available as demonstrators. LHA reuses and extends three different major components that have been previously developed by other projects. The SEEK management platform is the foundation providing a repository for archiving, presenting and secure sharing a wide range of publication results, such as published reports, (bio)medical data as well as interactive models and tools. The LHA Data Portal manages study metadata and data allowing to search for data of interest. Finally, PhenoMan is an ontological framework for phenotype modelling. This paper describes the interrelation of these three components. In particular, we use the PhenoMan to, firstly, model and represent phenotypes within the LHA platform. Then, secondly, the ontological phenotype representation can be used to generate search queries that are executed by the LHA Data Portal. The PhenoMan generates the queries in a novel domain specific query language (SDQL), which is specific for data management systems based on CDISC ODM standard, such as the LHA Data Portal. Our approach was successfully applied to represent phenotypes in the Leipzig Health Atlas with the possibility to execute corresponding queries within the LHA Data Portal.

Authors: A. Uciteli, C. Beger, J. Wagner, A. Kiel, F. A. Meineke, S. Staubert, M. Lobe, R. Hansel, J. Schuster, T. Kirsten, H. Herre

Date Published: 24th May 2021

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