Towards Using Social Media to Identify Individuals at Risk for Preventable Chronic Illness

Towards Using Social Media to Identify Individuals at Risk for Preventable Chronic Illness Dane Bell, Daniel Fried, Luwen Huangfu, Mihai Surdeanu, Stephen Kobourov University of Arizona Tucson, AZ 85721, USA {dane, dfried, huangfuluwen, msurdeanu, kobourov}@email.arizona.edu Abstract We describe a strategy for the acquisition of training data necessary to build a social-media-driven early detection system for individuals at risk for (preventable) type 2 diabetes mellitus (T2DM). The strategy uses a game-like quiz with data and questions acquired semi-automatically from Twitter. The questions are designed to inspire participant engagement and collect relevant data to train a public-health model applied to individuals. Prior systems designed to use social media such as Twitter to predict obesity (a risk factor for T2DM) operate on entire communities such as states, counties, or cities, based on statistics gathered by government agencies. Because there is considerable variation among individuals within these groups, training data on the individual level would be more effective, but this data is difficult to acquire. The approach proposed here aims to address this issue. Our strategy has two steps. First, we trained a random forest classifier on data gathered from (public) Twitter statuses and state-level statistics with state-of-the-art accuracy. We then converted this classifier into a 20-questions-style quiz and made it available online. In doing so, we achieved high engagement with individuals that took the quiz, while also building a training set of voluntarily supplied individual-level data for future classification. Keywords: machine learning, obesity detection, social media 1. Introduction Data collection in the public health domain is difficult due to privacy concerns and low engagement. For example, people seldom engage with surveys that require them to re- port their height and weight. However, such data is cru- cial for training automated public health tools, such as al- gorithms that detect risk for (preventable) type 2 diabetes mellitus (T2DM, henceforth diabetes). We propose a semi- automated data collection algorithm for obesity detection that mitigates these issues with a game-like quiz that is au- tomatically bootstrapped from a machine-learning model trained over social media data. The resulting quiz is non- intrusive, focusing on “fun” questions about food and food language while avoiding personal questions, which leads to high engagement. We believe this idea contributes to addressing one of the most challenging unsolved public health problems: the high rate of chronic illness resulting from modifiable risk factors such as poor diet quality and physical inactiv- ity. It is estimated that more than 86 million Americans over the age of 20 exhibit signs of pre-diabetes, and 70% of these pre-diabetic individuals will eventually develop T2DM, a chronic and debilitating disease associated with heart disease, stroke, blindness, kidney failure, and ampu- tations (National Center for Health Statistics, 2014; Amer- ican Diabetes Association and others, 2008). In the United States, the estimated cost of T2DM rose to $245 billion in 2012 (Association, 2013). Yet, 90% of these individuals at high risk are not aware of it (Li et al., 2013). Our long-term goal is to develop tools that automatically classify …

Content truncated to 500 words