Introduction
Dr. Jingcheng Du, Director of Natural Language Processing (NLP) Research at Melax Tech, conducted research that used Twitter as a gateway into the public’s perceptions of vaccines. Continue reading to find out what he discovered.
Background
Vaccines have been administered to people since the mid-eighteenth century, have been in wide commercial use since the mid-twentieth century, and are a major tool for the prevention of disease. Vaccines have been responsible for the eradication of smallpox worldwide and, when coupled with public health initiatives, the regional elimination of infectious diseases such as polio and yellow fever. Special care has been taken for children to have access to them so that they may have a chance to live healthy and disease-free lives. It’s quite common for children to receive vaccinations during their early years, including ones for Polio, Diphtheria, Rotavirus, and Hepatitis B.
Challenges
Partially due to the spread of misinformation about vaccines in social media and other news sources, over the past 20 years or so, there has been a significant and increasing number of people who are resisting vaccination by delaying vaccination or by outright refusing to vaccinate. Unfortunately, from a public health standpoint, high vaccination rates are required to achieve adequate prevention, and under-vaccination is known to be associated with infectious disease outbreaks. One way public health officials and the medical community seek to increase vaccination rates is by promotion strategies that target misinformation and misperception of vaccines, their side effects, and their benefits. Consequently, a good understanding of the current public perception of vaccination, generated in near-real time is highly important to those seeking to improve health. Unfortunately, collecting this information via traditional methods such as surveys and polls take too much time to respond to today’s dynamic media environment. Therefore, public health leaders need new tools to give them the information they need to combat misinformation about vaccines and other medical tools.
Solution
Fortunately, by combining data from social media platforms such as Twitter, and advances in artificial intelligence algorithms (e.g. natural language processing, deep learning), we can access large scale data on public perceptions related to vaccinations in near real time.
Figure: The Architecture of VaxInsight
Using data from Twitter over a ten-year period, Dr. Du developed AI and DL algorithms to organize and categorize posts, comments, and mentions related to the HPV vaccine. The data extracted from social media was framed within the perspectives offered by the health belief model (HBM), a grounded behavior change theory scientific framework designed to explain how humans make decisions about their health.
Dr. Du’s work categorized tweets about the human papillomavirus (HPV) using the health belief model (HBM) constructs. A second deep-learning NLP model analyzed the demographic characteristics of Twitter users and how these related to their perception of the HPV vaccine. The results of this analysis were fed into a web-based interactive dashboard that could monitor discussions about HPV taking place on Twitter in real time. Although Dr. Du’s work focused on HPV, his method and platform are easily generalizable to other types of diseases and vaccines.
Conclusion
Users of this platform can now explore the manner in which vaccine uptake changes over time and answer questions about how misinformation and public perception impact this uptake. Among other features, the platform supports the comparison of different subpopulations of users (e.g., male vs. female, Iowa vs. California, etc.), and variation of perception over time. The resulting tools can be used in countless ways for years to better the public's overall health.
What question would you like answered for your industry or area of medicine? Request a demo today to find out how Melax Tech’s solutions and expertise can help you gain insights from the vast data available through social media platforms.
Reference: Summary of Abstract for VaxInsight: an artificial intelligence system to access large-scale public perceptions of vaccination from social media (pages 7-9)