Data in Action: Delivering Better Care Through Education

Oct 18 / 2019

How can doctors, or those in the medical profession, deliver better care through education? That is the question that students from the University of British Columbia’s Master of Data Science (MDS) worked to answer for their capstone project with QxMD.

QxMD is a Vancouver-based digital learning technology company that provides mobile solutions helping to drive evidence-based medicine into clinical practice. One of those mobile solutions is the Read by QxMD app, which provides a single place to keep up with new medical and scientific research, read outstanding topic reviews, and search PubMed, a free resource supporting the search and retrieval of peer-reviewed biomedical and life sciences literature.

Being able to provide better care through education begins with being aware of what’s going on in the news. For example, a patient reading a newspaper article on, say Lyme disease, wonders if the symptoms they’ve been experiencing over the past week might be due to it. The patient then goes to see their family doctor about their concerns.

The project had three goals: 1) identify trending health topics within newspaper articles, 2) match these to medical journal articles, and 3) make recommendations that are relevant to users.

To identify the trending health topics, the students created a steady stream of newspaper articles using Google News API, scraped the article corpus using Scraping Hub, and used the spaCy NLP (Natural Language Processing) library to identify the health topics within the newspaper article itself.

In order to match the health topics to medical journal articles, the MDS students created a vector to represent the newspaper articles and also the medical journal articles. The students also worked on clustering the users so that recommendations of interest would be sent to them.

What the students ended up creating was a MySQL data table that matches a news article with its relevant medical journal article, and also includes such criteria as how close the match is. Each row of the table has a recommendation that allows QxMD’s CEO to go through and then send out the relevant medical journal articles to users via Read by QxMD.

This pairing of news articles with medical journal articles equips doctors with the knowledge needed to address the questions and concerns patients may have after reading the news.

Explore Computational Linguistics