Natural Language Processing – Understanding and Analyzing Medical Data

Natural Language Processing – Understanding and Analyzing Medical Data

Vaccination and an understanding of vital medication is a very important topic. It ties into individual health but also affects public health safety through infection. A child vaccination study from 2015 introduced the worry of declining vaccination rates for the Public Health Agency of Canada. They recognized that “all vaccine coverage estimates reported…were below the corresponding national vaccination coverage goals” (1). They also reported that 9/10 Canadian Immunization Guide users were satisfied with the publication’s trustworthiness, quality of information, range of topics covered and clarity of ease of understanding (1). They, therefore, defined the core issue as a need to “address the perception or beliefs that reduce vaccine uptake” (1). 

Evenset jumped into the fray to devise a solution. The experienced custom software development team applied their specialization in machine learning and proposed to develop a Natural Language Processing (NLP) solution that will help users navigate through the Canadian Immunization Guide in a more simple, comfortable and familiar way.

The solution is based on a content web scrapper and crawler tailored to medical text which collects, digest and store data. The data collected by the scraper is passed to a knowledge-base module where a graph representation of medical data about vaccination is stored. The interface of the graph database is a chatbot where we use NLP and machine learning to understand the conversation and respond with appropriate medically validated information to the user.

At Evenset we are proud to research and build the natural language processing solutions for healthcare and medical applications for the betterment of our Canadian communities. We believe in standing against negative health trends and providing clarity to the misinformed.

 

(1) http://publications.gc.ca/collections/collection_2018/aspc-phac/HP40-156-2018-eng.pdf