Social Media

Recently, a growing number of studies have made use of Twitter to track the spread of infectious disease. These investigations show that there are reliable spikes in traffic related to keywords associated with the spread of infectious diseases like Influenza [1], as well as other Syndromes [2]. However, little research has been done using Social Media to monitor chronic conditions like Asthma, which do not spread from sufferer to sufferer. We therefore test the feasibility of using Twitter for Asthma surveillance, using techniques from NLP and machine learning to achieve a deeper understanding of what users Tweet about Asthma, rather than relying only on keyword search.

Objective

We present a Content Analysis project using Natural Language Processing to aid in Twitter-based syndromic surveillance of Asthma

Social media is of considerable interest as a sensor into the thoughts, interests and health of a population. We consider three types of health events that an analyst may wish to be made aware of:

- Given a known disease, such as MERS, SARS, Measles, etc., an event corresponds to individuals contracting the disease.

- Given a set of symptoms (fever, stomach pain, etc.), an event is an unusual number of individuals1 complaining of the symptoms.

- Most generally: an event is an unusually large group of individuals who can be identified as being effected by some personal illness.

Note that to detect an “unusual number” of something, we need to count the indicators of the event, and we need to compare the current count with past counts. Further, we are generally interested in geographically constrained events, and so for this work we will focus on county-based counts. We will count the number of items (tweets or individuals) expressing the event indicator (a disease name, symptom, or classified as “personal health related” as indicated by our classifier). Our approach to detecting health related events is: filter -> classify -> detect. We first filter out tweets that contain no “health related” terms, then apply a classifier to each tweet. This classifier is designed to flag a tweet as being about “personal health” or not. We then aggregate the positive instances per day at the county level and detect as an event any county/day pair with an unusually high count (as compared to the recent past).

Objective

In this work we investigate the extent to which social media, in particular Twitter, can be used to detect an outbreak of a disease or illness. We term these outbreaks “events”, and we will describe methodologies for detecting events.

In response to the rise in obesity rates and obesity-related healthcare costs over the past several decades, numerous organizations have implemented obesity prevention programs. The current method for evaluating the success of these programs relies largely on annual surveys such as the Centers for Disease Control and Prevention’s Behavioral Risk Factor Surveillance System (BRFSS) which provides state-by-state obesity rates. As a result, public health policy makers lack the fine-grained evaluation data needed to make timely decisions about the success of their obesity prevention programs and to allocate resources more efficiently.

Objective

We developed Persistent Health Assessment Tools, PHAT, to equip public health policy makers with more precise tools and timely information for measuring the success of obesity prevention programs. PHAT monitors social media to supplement traditional surveillance by making real-time estimates based on observations of obesity-relevant behaviors.

 

Traditional surveillance systems only capture a fraction of the estimated 48 million yearly cases of foodborne illness in the United States due to few affected individuals seeking medical care and lack of reporting to appropriate authorities. Non-traditional disease surveillance approaches could be used to supplement foodborne illness surveillance systems.

Objective

We assessed whether foodservice reviews on Yelp.com (a business review site) can be used to support foodborne illness surveillance efforts.

In today’s fast paced world, information is available (and expected) instantaneously. Social media has only fueled this expectation as it has permeated all aspects of our lives. More and more of the population is turning to social media outlets to share their thoughts and update their status, especially during disasters. With all these conversations occurring, it is only reasonable to assume that health status is part of the information being shared. In fact, studies by Johns Hopkins University and Harvard University have shown that social media reporting can serve as an early indicator and warning of emerging health issues within a community. Whether people are talking about being sick themselves or fear of illness in the community, there is a wealth of knowledge to be gained by tapping into this information. Yet gaining insight and understanding from social media data can be problematic. The unstructured nature of the data, the presence of social media “spam”, and the frequency of reposting information makes social media a noisy data source. Being able to harness this data would provide the opportunity to use social media as an effective situational awareness and early warning tool for biosurveillance missions. But how do you accomplish this? There are tens of millions of conversations happening on social media every day that would need to be sifted through to get to the health related topics. No public health entity has the time or staffing for that endeavor.

Objective

The goal of the Now Trending website is to provide a web based tool that pulls out relevant Twitter conversations concerning illness and disasters and provides meaningful analytics on how those conversations are trending. The website gives the user the ability to view trends overall and for specific geographic areas.

Early detection of a disease outbreak using pre-diagnostic textual data is available in biosurveillance systems with the integration of data such as chief complaints. Social media has been identified as an additional pre-diagnostic data source of interest. Textual data analysis in public health is usually based on a keyword search and often involves a complex Boolean combination of terms that produce results with many false alarms. Epidemiologists may wish to query the data differently based on the event of interest, yet the process is laborious to weed out uninteresting content. Specialized detectors that decide on the topical relevance of keyword search usually require developers to adapt methods to new uses, which is a time- and cost-prohibitive activity. Users need the ability to rapidly build text content detectors on their own.

Objective

To demonstrate a framework for user-customizable text processing that can improve the efficiency and effectiveness of mining text for biosurveillance, with initial application to Twitter.

Active surveillance for influenza is a useful but costly endeavor. In recent years infoveillance tools have been developed to track and analyze data available on the Internet and social media (Eysenbach 2011). While infoveillance tools have been developed, few tools focus on geo-targeted data collection at a local level combined with Geographic Information Systems (GIS) capability.

Objective

We developed geo-targeted social media application program interfaces (APIs) for Twitter and a web-based social media analytics and research testbed (SMART) dashboard to analyze “flu” related tweets. During the 2013-14 flu season, for 10 cities with active surveillance for influenza (ILI), we correlated weekly tweeting rates and visual patterns of flu tweeting rates. To facilitate widespread use and testing of this system, we developed an interactive webbased dashboard “SMART” that allows practitioners to monitor and visualize daily changes of flu trends and related flu news.

 

Real-time monitoring and analysis of vaccine concerns over time and location could help immunisation programmes to tailor more effective and timely strategies to address specific public health concerns. In recent years attempts [1, 2] are being made to develop a more systematic monitoring of broader public vaccine concerns resulting in vaccine refusals and potential disease outbreaks. Automated sentiment analysis software applications are being developed to detect and track the emergence and spread, geographically and temporally, of online social media reports on vaccines by developing a new application for opinion mining and sentiment analysis. Although many of the current approaches for automated sentiment analysis provide a timely method to assess the sentiment of a population towards vaccination, they do not assess beliefs, perceptions and behaviours. Incorporating semantic approach by using ontologies captures the domain knowledge and supports automated extraction and analysis of text in blog posts related to vaccination.

Objective

This paper presents our approach on design and development of an integrated semantic platform to capture the domain knowledge on vaccine sentiments, beliefs, and behaviours using ontologies. The vaccine sentiment ontology (VASON) provides more structure around the vast amount of unstructured data scattered over blog posts to facilitate blog content analysis, and discovering patterns of words or phrases in blogs text (e.g. specifying topics, themes, sentiment, beliefs and so on). It also assists in revealing opinionated claims and assertions in blogs and specifying the authors, forms, functions, geographical locations, audiences of blogs, as well as bloggers’ motives.