Microsoft Israel is helping doctors to read unstructured text with Text Analytics for Health (TAFH), an Azure Cognitive Services application that uses artificial intelligence to translate medical data.
According to Microsoft, unstructured text often restricts doctors from providing fast and effective care to their patients, as text can be left unreferenced or noted in different languages, making it difficult to translate and understand.
TAFH uses AI to extract meaningful information from these documents and translate it to improve medical care. The AI models have been trained on multiple languages, considering different syntax systems, as well as jargon, colloquialisms, local medical terms and country-specific shorthand.
Microsoft recently released previews of TAFH in six additional languages to English, including Spanish, French, German, Italian, Portuguese and Hebrew. This made the solution the first natural language processing (NLP) service of is kind.
TAFH uses NLP to detect and identify medical terms in text, as well as classify and associate them with standard clinical coding systems. Users can then identify key patient information, identify risks and automate form-filling by combining clinical data.
“Most healthcare technology is limited to the English language, making it inaccessible to millions of people and countries where English is not the primary language,” said Hadas Bitran, head of the Israel research and development centre for life and health sciences at Microsoft. “Launching NLP technology in multiple languages is a major step forward in closing the gaps in health equity created by language barriers and ensuring that access to and quality of healthcare is not determined by the ability to speak and to understand English.”
Microsoft worked with Leumit Health Services, one of Israel’s four national health funds, to train the TAFH model for the Hebrew language. This involved studying de-identified medical records and translating them for key health information. “Leumit created a separate subscription in Microsoft Azure with strict access permissions where Microsoft installed its federated learning infrastructure and tools,” said Bitran. “Later, Leumit entered the data needed for the research, and Microsoft developers enabled model training in a federated learning setup, without the data leaving the subscription.”
Izhar Lauger, director of Leumit Start, said: “Using Microsoft’s Hebrew NLP, we will be able to analyse our 20 years of EMR data and patient-to-doctor messages to develop tools that will save clinicians time and reduce their burnout in a post-Covid-19 world.”