A Need To Improve Clinical Decisions By Incorporating Semantics

There is a strong need to improve clinical decisions by incorporating semantics derived from various forms of human input (e.g. free text, medical records, literature). Vast amount of information is currently held in medical records in the form of free text. Thus, text analytics is important to unravel the insights within the textual data. Particularly in healthcare, but in almost all other industries, records (digital or not) are still kept as free text. There is plethora of applications in the clinical setting where practitioners produce and rely on free text for reporting diagnosis and operations. Of particular importance is the mining of medical literature, which enables the use of vast amounts of medical knowledge more efficiently. Examples include literature recommender systems and also the detection of new medical knowledge from literature, e.g. for drug repositioning.

Given the large amount of biomedical knowledge recorded in textual form, full papers, abstracts and online content, there is a need for techniques that can identify, extract, manage and integrate this knowledge. In the parallel, text analytics tools have been adapted and further developed for extracting relevant concepts and relations among concepts from clinical data such as patient records or reports written by doctors. The information extraction technology plays a central role for text mining and text analytics. Even though there has been significant breakthrough in natural language processing with the introduction of machine learning technologies, in particular, recently, deep learning methods, these technologies need to be further developed to meet the challenges of large volumes and velocities introduced with Big Data. Many applications can benefit from the true meaning of the content.

Healthcare Knowledge-bases A complex analysis and multidisciplinary approach to knowledge is essential to understand the impact of various factors on healthcare systems. The challenges for accepting and addressing the issue concerning the healthcare world are the use of Big data, non-conformance to standard, various sources (in varied documents and formats), which need an immediate attention towards multidisciplinary complex data analytics on top of rich semantic data models. Ontology driven systems result indeed in the effective implementation of healthcare strategies for the policy makers. The production for semantic knowledge-bases to social insurance need a extremely high potential Also useful impact. They encourage information coordination starting with various heterogeneous sources, empower those improvement of majority of the data sifting systems, and backing learning revelation assignments. To particular, in the last A long time those Linked Open Data (LOD) activity arrived at huge selection What's more is acknowledged the reference act to imparting Also distributed organized information on the Web. LOD offers the possibility of using data across different domains for purposes like statistics, analysis, maps and publications. By linking this knowledge, interrelations and associations can be inferred, and new conclusions arise.

Healthcare data is generated in various sources in diverse formats using different terminologies. Due to the heterogeneous formats and lack of common vocabulary, the accessibility of the big data of healthcare is very minimal for health data analytics and decision support systems. Vocabulary standards are used to describe clinical problems and procedures, medications, and allergies . Important examples are, just to name a few, the Logical Observation Identifiers Names and Codes (LOINC), International Classification of Diseases (ICD 9 and ICD10), Systematized Nomenclature of Medicine-Clinical Terms (SNOMED-CT), Current Procedural Terminology, 4th Edition (CPT 4), ATC – Anatomic Therapeutic Chemical Classification of Drugs, Gene Ontology (GO), RxNorm, General Equivalence Mappings (GEMs), OBO-Foundry and others.

Since healthcare systems are characterized by a large amount of data, heterogeneous in nature and with different quality and security requirements, research on the opening process, data reengineering, linking, formalization and consumption is of primary interest. The heterogeneity problem has to be tackled at different levels. On the one hand, syntactic interoperability is needed to unify the format of knowledge sources enabling, e.g., distributed query. Syntactic interoperability can be achieved by conforming to universal knowledge representation languages and by adopting standards practices. The widely adopted RDF, OWL and LOD approaches support syntactic interoperability. On the other hand, semantic interoperability is also needed. Semantic interoperability can be achieved by adopting a uniform data representation and formalizing all concepts into a holistic data model (conceptual interoperability). RDF and OWL assist in achieving the former goal. However, conceptual interoperability is domain specific and cannot be achieved only by the adoption of standards tools and practices, but also through interlinking with existing healthcare knowledge-bases by means of domain experts and semi-automated solutions. The large, heterogeneous data sources in the healthcare world make the problem even harder, as different semantic perspectives must be addressed in order to cope with knowledge source conceptualizations.

Once interoperability at both syntactic and conceptual levels is obtained, it is possible to intercross data and exploit them more in depth, providing application developers the opportunity to easily design their services and applications. Semantic interoperability at domain level allows making sense of distributed data and enabling their automatic interpretation. In this way, the issue of resolving semantic interoperability among different data sources is moved from the application level to the data model level. Developers are then relieved from the burden of reconciling, uniforming, and linking data at a conceptual level, and are able to build their solutions in a more intuitive and efficient way. The published data sources are made discoverable and become accessible via queries and/or public facilities, and integrated into higher-level services.

Presently several international organizations and agencies across the world, like e.g. the World Health Organization (WHO), make use of semantic knowledge-bases in health care systems to:

• Improve accuracy of diagnoses by providing real time correlations of symptoms, test results and individual medical histories;
• Help to build more powerful and more interoperable information systems in healthcare;
• Support the need of the healthcare process to transmit, re-use and share patient data;
• Provide semantic-based criteria to support different statistical aggregations for different purposes;
• Bring healthcare systems to support the integration of knowledge and data.

Putting knowledge management systems in place on healthcare can facilitate the flow of information and result in better, more-informed decisions.