Amid the COVID-19 pandemic, the healthcare industry is flooded with data. Most of the medical data are in an unstructured form that includes doctor’s names, medical publications, electronic health records, clinical trial protocols, medical encounters, transcripts, etc. Therefore, healthcare organizations, providers, researchers, pharmaceutical companies, and others are facing challenges in drawing insights from this unstructured data. As a solution, Microsoft has introduced a new analytics feature in its Azure Cognitive service that will allow to process and extract insights from the unstructured medical data.
The Corporate Vice President Eric Boyd said that, the text analytics helps the healthcare industry people to draw good insights and relationships from the unstructured medical data.
Microsoft analytics helps to find solutions from processing the information and for improved health outcomes.
Azure Text Analytics for health is a service that takes out and labels appropriate medical information from texts such as doctors’ notes, discharge summaries, clinical documents, and electronic health records. It currently performs Named Entity Recognition (NER), relation extraction, entity negation, and entity linking for English-language text.
NER finds words and phrases in unstructured data that can be aligned with more than one semantics like diagnosis, medication name, symptom, age.
Relation extraction finds relationships between the words mentioned in the text. For example, “time of condition, this relation found by aligning a condition name with time.
Entity linking aligns the name entities to the text concepts in the predefined database.
Text analytics aligns health and biomedical vocabularies found in the database of the Unified Medical Language System (UMLS) Meta-thesaurus knowledge source.
Text analytics also supports negation detection for different entities. Sometimes the meaning of medical information is affected by negation which could result in critical implications if misdiagnosed.
The ongoing global pandemic has geared the organizations to find new methods to process the data and to generate insights.
Text analytics will help the health industry to extract the insights, saves time and reduce costs, and improves customer engagement, said Microsoft.
The Group Manager of Microsoft Healthcare Hadas Bitran said that they have trained the health feature on a different range of medical data including various forms of clinical notes, clinical trial protocols, etc. He also added, amidst the pandemic, this text analytics would help and enables the researches to evaluate and draw insights from the deluge unstructured medical information.
The company has also partnered with the Allen Institute for AI and research groups to prepare the COVID-19 Open Research Dataset. It is a free resource that consists of 47,000 scholarly articles for -use by the global research community.
