Sai T. Moturu

Sensing the "Health State" of a Community

Mobile phones are a pervasive platform for opportunistic sensing of behaviors and opinions. Three studies use location and communication sensors to model individual behaviors and symptoms, long-term health outcomes, and the diffusion of opinions in a community. These three analyses illustrate how mobile phones can unobtrusively monitor rich social interactions, because the underlying sensing technologies are now commonplace and readily available.

Social sensing: obesity, unhealthy eating and exercise in face-to-face networks

What is the role of face-to-face interactions in the diffusion of health-related behaviors- diet choices, exercise habits, and long-term weight changes? We use co-location and communication sensors in mass-market mobile phones to model the diffusion of health-related behaviors via face-to-face interactions amongst the residents of an undergraduate residence hall during the academic year of 2008--09. The dataset used in this analysis includes bluetooth proximity scans, 802.11 WLAN AP scans, calling and SMS networks and self-reported diet, exercise and weight-related information collected periodically over a nine month period. We find that the health behaviors of participants are correlated with the behaviors of peers that they are exposed to over long durations. Such exposure can be estimated using automatically captured social interactions between individuals. To better understand this adoption mechanism, we contrast the role of exposure to different sub-behaviors, i.e., exposure to peers that are obese, are inactive, have unhealthy dietary habits and those that display similar weight changes in the observation period. These results suggest that it is possible to design self-feedback tools and real-time interventions in the future. In stark contrast to previous work, we find that self-reported friends and social acquaintances do not show similar predictive ability for these social health behaviors.

Using Social Sensing to Understand the Links between Sleep, Mood, and Sociability

In recent years, reality mining experiments have provided several novel insights into human social behavior that would not have been possible without the novel use of smart phone sensing. In this work, we leverage the latest reality mining experiment to study social behavior from a public health perspective. In particular, we focus on sleep and mood as they have a considerable public health impact with serious societal and significant financial effects. We endeavor to explore and uncover the associations between sleep, mood and sociability by studying a population of healthy young adults going about their everyday life. We find significant associations between sleep and mood, reiterating observations in the literature. More importantly, we find that individuals with lower overall sociability tend to report poor mood more often, a statistically significant observation. In addition, we also uncover associations between daily sociability and sleep, a previously unreported observation. These results demonstrate the potential of reality mining studies for studying the sociological aspects of significant public health problems. Further, we hope that our work will provide the impetus for larger studies validating some of these observations and ultimately result in behavioral interventions that can improve public health through better social interaction.

Quantifying the trustworthiness of social media content

The growing popularity of social media in recent years has resulted in the creation of an enormous amount of user-generated content. A significant portion of this information is useful and has proven to be a great source of knowledge. However, since much of this information has been contributed by strangers with little or no apparent reputation to speak of, there is no easy way to detect whether the content is trustworthy. Search engines are the gateways to knowledge but search relevance cannot guarantee that the content in the search results is trustworthy. A casual observer might not be able to differentiate between trustworthy and untrustworthy content. This work is focused on the problem of quantifying the value of such shared content with respect to its trustworthiness. In particular, the focus is on shared health content as the negative impact of acting on untrustworthy content is high in this domain. Health content from two social media applications, Wikipedia and Daily Strength, is used for this study. Sociological notions of trust are used to motivate the search for a solution. A two-step unsupervised, feature-driven approach is proposed for this purpose: a feature identification step in which relevant information categories are specified and suitable features are identified, and a quantification step for which various unsupervised scoring models are proposed. Results indicate that this approach is effective and can be adapted to disparate social media applications with ease.

Trust evaluation in health information on the World Wide Web

The impact of health information on the web is mounting and with the Health 2.0 revolution around the corner, online health promotion and management is becoming a reality. User-generated content is at the core of this revolution and brings to the fore the essential question of trust evaluation, a pertinent problem for health applications in particular. Evolving Web 2.0 health applications provide abundant opportunities for research. We identify these applications, discuss the challenges for trust assessment, characterize conceivable variables, list potential techniques for analysis, and provide a vision for future research.

Sleep, mood and sociability in a healthy population

Sleep and mood problems have a considerable public health impact with serious societal and significant financial effects. In this work, we study the relationship between these factors in the everyday life of healthy young adults. More importantly, we look at these factors from a social perspective, studying the impact that couples have on each other and the role that face-to-face interactions play. We find that there is a significant bi-directional relationship between mood and sleep. More interestingly, we find that the spouses sleep and mood may have an effect on the subjects mood and sleep. Further, we find that subjects whose sleep is significantly correlated with mood tend to be more sociable. Finally, we observe that less sociable subjects show poor mood more often than their more sociable contemporaries. These novel insights, especially those involving sociability, measured from quantified face-to-face interaction data gathered through smartphones, open up several avenues to enhance public health research through the use of latest wireless sensing technologies.

Predicting Future High-Cost Patients: A Real-World Risk Modeling Application

Health care data from patients in the Arizona Health Care Cost Containment System, Arizonas Medicaid program, provides a unique opportunity to exploit state-of-the-art data processing and analysis algorithms to mine the data and provide actionable results that can aid cost containment. This work addresses specific challenges in this real-life health care application to build predictive risk models for forecasting future high-cost users. Such predictive risk modeling has received attention in recent years with statistical techniques being the backbone of proposed methods. We survey the literature and propose a novel data mining approach customized for this potent application. Our empirical study indicates that this approach is useful and can benefit further research on cost containment in the health care industry.

Evaluating the trustworthiness of Wikipedia articles through quality and credibility

Wikipedia has become a very popular destination for Web surfers seeking knowledge about a wide variety of subjects. While it contains many helpful articles with accurate information, it also consists of unreliable articles with inaccurate or incomplete information. A casual observer might not be able to differentiate between the good and the bad. In this work, we identify the necessity and challenges for trust assessment in Wikipedia, and propose a framework that can help address these challenges by identifying relevant features and providing empirical means to meet the requirements for such an evaluation. We select relevant variables and perform experiments to evaluate our approach. The results demonstrate promising performance that is better than comparable approaches and could possibly be replicated with other social media applications.

Smartphone-Based Geofencing to Ascertain Hospitalizations

Background— Ascertainment of hospitalizations is critical to assess quality of care and the effectiveness and adverse effects of various therapies. Smartphones, mobile geolocators that are ubiquitous, have not been leveraged to ascertain hospitalizations. Therefore, we evaluated the use of smartphone-based geofencing to track hospitalizations. Methods and Results— Participants aged ≥18 years installed a mobile application programmed to geofence all hospitals using global positioning systems and cell phone tower triangulation and to trigger a smartphone-based questionnaire when located in a hospital for ≥4 hours. An in-person study included consecutive consenting patients scheduled for electrophysiology and cardiac catheterization procedures. A remote arm invited Health eHeart Study participants who consented and engaged with the study via the internet only. The accuracy of application-detected hospitalizations was confirmed by medical record review as the reference standard. Of 22 eligible in-person patients, 17 hospitalizations were detected (sensitivity 77%; 95% confidence interval, 55%–92%). The length of stay according to the application was positively correlated with the length of stay ascertained via the electronic medical record (r=0.53; P=0.03). In the remote arm, the application was downloaded by 3443 participants residing in all 50 US states; 243 hospital visits at 119 different hospitals were detected through the application. The positive predictive value for an application-reported hospitalization was 65% (95% confidence interval, 57%–72%). Conclusions— Mobile application–based ascertainment of hospitalizations can be achieved with modest accuracy. This first proof of concept may ultimately be applicable to geofencing other types of prespecified locations to facilitate healthcare research and patient care.

Automatically captured sociability and sleep quality in healthy adults

Sleep and social interactions have been shown to have a considerable public health impact. However, little is known about how these affect each other in healthy individuals. This research is first to propose the exploration of the bidirectional relationship between technologically sensed sleep quality and quantified face-to-face social interactions. We detail a pilot study designed to study the relationship of sociability and sleep quality, both measured and perceived, of healthy adults. We capture real-world social interactions and measure sleep in a naturalistic setting using wireless sensing technologies. We find that it may not be the device-defined sleep quality (ZQ score) but our perceived sleep quality which affects our following days sociability. Further, we also find perceived sleep quality is more strongly correlated to normalized ZQ scores than the actual scores. These intriguing insights raise several questions on how an individuals social life could be affected by sleep and indicate the usefulness of mobile sensing technologies in understanding public health phenomena.