Health systems need to provide inclusive, secure and ethical access to data, which also includes the integration of existing and prospective health records, such as emergency department admissions, discharge summaries, outpatient care reports, or drug prescriptions. Overall, this kind of unstructured text contains much valuable information and complements the structured patient data. Thus, the first focus of this Research Topic is the exploitation of information from health documents.
Additionally, novel technologies and approaches are required not only to be informative, but also complemented with explanatory power, in order to reduce the burden coming from operators' acceptance and trust, and to ensure that systems perform as expected. These challenges require either to refine neural architectures to provide users with justification and explanation on their results, or to merge these approaches with semantic technologies, such as formal ontologies, to complement results with further insights on the tasks being considered. Yet, the integration of huge amounts of data and the need to jointly exploit broad data collections poses specific challenges to NLP applications, such as dealing with multilinguality. This is particularly true for applications and services to collect data on rare diseases to exploit multimodal information for diagnostic purposes, and to analyze and compare different medical protocols addressing specific issues or diseases.
This Research Topic specifically addresses clinical event extraction, and encompasses algorithms and systems aimed at extracting structured information from unstructured, free text, health records. Targeted clinical events are those with a deeper impact on both individuals’ daily life, and huge costs for health systems. These may include, but are not limited to violent events (especially against women, elderly and children); home and leisure accidents (especially those involving products); workplace accidents and injuries.
We solicit submission articles on substantial, original, and unpublished research in all aspects related to the mentioned issues, including but not limited to the following areas:
? Health reports analysis including Text Categorization, Topic Models and Information Retrieval
? Architectures and systems for Information Extraction and Text Mining from health data
? Medical Question Answering and Summarization
? Interpretability and Explainability in neural systems for medical records management
? Resources and Evaluation for health information systems
? Methodologies and practices for semantic annotation
? Formal ontologies for trustworthy AI systems in the medical domain
? Design of meaning representations for health reports and applications
? Tracking of disease progression timelines
? Information Extraction for Crisis Management
? Information Extraction from reports for fact-checking of social media communication on health topics and issues
? Multilinguality in health registries
? Multimodal analysis of text and image documents
? Protecting personal data and anonymization techniques for health records
? Gender issues in medical information systems
? Ethical issues in medical information systems
Health systems need to provide inclusive, secure and ethical access to data, which also includes the integration of existing and prospective health records, such as emergency department admissions, discharge summaries, outpatient care reports, or drug prescriptions. Overall, this kind of unstructured text contains much valuable information and complements the structured patient data. Thus, the first focus of this Research Topic is the exploitation of information from health documents.
Additionally, novel technologies and approaches are required not only to be informative, but also complemented with explanatory power, in order to reduce the burden coming from operators' acceptance and trust, and to ensure that systems perform as expected. These challenges require either to refine neural architectures to provide users with justification and explanation on their results, or to merge these approaches with semantic technologies, such as formal ontologies, to complement results with further insights on the tasks being considered. Yet, the integration of huge amounts of data and the need to jointly exploit broad data collections poses specific challenges to NLP applications, such as dealing with multilinguality. This is particularly true for applications and services to collect data on rare diseases to exploit multimodal information for diagnostic purposes, and to analyze and compare different medical protocols addressing specific issues or diseases.
This Research Topic specifically addresses clinical event extraction, and encompasses algorithms and systems aimed at extracting structured information from unstructured, free text, health records. Targeted clinical events are those with a deeper impact on both individuals’ daily life, and huge costs for health systems. These may include, but are not limited to violent events (especially against women, elderly and children); home and leisure accidents (especially those involving products); workplace accidents and injuries.
We solicit submission articles on substantial, original, and unpublished research in all aspects related to the mentioned issues, including but not limited to the following areas:
? Health reports analysis including Text Categorization, Topic Models and Information Retrieval
? Architectures and systems for Information Extraction and Text Mining from health data
? Medical Question Answering and Summarization
? Interpretability and Explainability in neural systems for medical records management
? Resources and Evaluation for health information systems
? Methodologies and practices for semantic annotation
? Formal ontologies for trustworthy AI systems in the medical domain
? Design of meaning representations for health reports and applications
? Tracking of disease progression timelines
? Information Extraction for Crisis Management
? Information Extraction from reports for fact-checking of social media communication on health topics and issues
? Multilinguality in health registries
? Multimodal analysis of text and image documents
? Protecting personal data and anonymization techniques for health records
? Gender issues in medical information systems
? Ethical issues in medical information systems