Hsien Wen Meng

Leveraging geotagged Twitter data to examine neighborhood happiness, diet, and physical activity

OBJECTIVES Using publicly available, geotagged Twitter data, we created neighborhood indicators for happiness, food and physical activity for three large counties: Salt Lake, San Francisco and New York. METHODS We utilize 2.8 million tweets collected between February-August 2015 in our analysis. Geo-coordinates of where tweets were sent allow us to spatially join them to 2010 census tract locations. We implemented quality control checks and tested associations between Twitter-derived variables and sociodemographic characteristics. RESULTS For a random subset of tweets, manually labeled tweets and algorithm labeled tweets had excellent levels of agreement: 73% for happiness; 83% for food, and 85% for physical activity. Happy tweets, healthy food references, and physical activity references were less frequent in census tracts with greater economic disadvantage and higher proportions of racial/ethnic minorities and youths. CONCLUSIONS Social media can be leveraged to provide greater understanding of the well-being and health behaviors of communities-information that has been previously difficult and expensive to obtain consistently across geographies. More open access neighborhood data can enable better design of programs and policies addressing social determinants of health.

Social media indicators of the food environment and state health outcomes

OBJECTIVES Contextual factors can influence health through exposures to health-promoting and risk-inducing factors. The aim of this study was to (1) build, from geotagged Twitter and Yelp data, a national food environment database and (2) to test associations between state food environment indicators and health outcomes. STUDY DESIGN This is a cross-sectional study based upon secondary analyses of publicly available data. METHODS Using Twitters Streaming Application Programming Interface (API), we collected and processed 4,041,521 food-related, geotagged tweets between April 2015 and March 2016. Using Yelps Search API, we collected data on 505,554 unique food-related businesses. In linear regression models, we examined associations between food environment characteristics and state-level health outcomes, controlling for state-level differences in age, percent non-Hispanic white, and median household income. RESULTS A one standard deviation increase in caloric density of food tweets was related to higher all-cause mortality (+46.50 per 100,000), diabetes (+0.75%), obesity (+1.78%), high cholesterol (+1.40%), and fair/poor self-rated health (2.01%). More burger Yelp listings were related to higher prevalence of diabetes (+0.55%), obesity (1.35%), and fair/poor self-rated health (1.12%). More alcohol tweets and Yelp bars and pub listings were related to higher state-level binge drinking and heavy drinking, but lower mortality and lower percent reporting fair/poor self-rated health. Supplemental analyses with county-level social media indicators and county health outcomes resulted in finding similar but slightly attenuated associations compared to those found at the state level. CONCLUSIONS Social media can be utilized to create indicators of the food environment that are associated with area-level mortality, health behaviors, and chronic conditions.

Neighbourhood looking glass: 360º automated characterisation of the built environment for neighbourhood effects research

Background Neighbourhood quality has been connected with an array of health issues, but neighbourhood research has been limited by the lack of methods to characterise large geographical areas. This study uses innovative computer vision methods and a new big data source of street view images to automatically characterise neighbourhood built environments. Methods A total of 430 000 images were obtained using Google’s Street View Image API for Salt Lake City, Chicago and Charleston. Convolutional neural networks were used to create indicators of street greenness, crosswalks and building type. We implemented log Poisson regression models to estimate associations between built environment features and individual prevalence of obesity and diabetes in Salt Lake City, controlling for individual-level and zip code-level predisposing characteristics. Results Computer vision models had an accuracy of 86%–93% compared with manual annotations. Charleston had the highest percentage of green streets (79%), while Chicago had the highest percentage of crosswalks (23%) and commercial buildings/apartments (59%). Built environment characteristics were categorised into tertiles, with the highest tertile serving as the referent group. Individuals living in zip codes with the most green streets, crosswalks and commercial buildings/apartments had relative obesity prevalences that were 25%–28% lower and relative diabetes prevalences that were 12%–18% lower than individuals living in zip codes with the least abundance of these neighbourhood features. Conclusion Neighbourhood conditions may influence chronic disease outcomes. Google Street View images represent an underused data resource for the construction of built environment features.

Cardiovascular Disease–related Health Beliefs and Lifestyle Issues Among Karen Refugees Resettled in the United States From the Thai-Myanmar (Burma) Border

Objectives Refugees resettled in the US may be at risk for cardiovascular disease (CVD). However, little is known about CVD-related issues among Karen refugees who have migrated to the US from the Thai-Myanmar border. The purpose of this study was to examine CVD-related health beliefs and lifestyle issues among Karen refugees resettled in the US. Methods Karen refugees resettled in the US from the Thai-Myanmar border (n=195) participated in a survey study on health beliefs related to CVD, salt intake, physical activity (PA), and smoking in the fall of 2016. Results A high-salt diet, physical inactivity, and smoking were major lifestyle problems. Participants who adhered to a low-salt diet considered themselves to be susceptible to CVD. Most participants did not engage in regular PA. Regular PA was associated with less perceived susceptibility to CVD and greater perceived benefits of a healthy lifestyle for decreasing the likelihood of CVD. Conclusions Each refugee population may require individualized strategies to promote PA and a healthy diet. Future studies should develop health education programs that are specifically designed for Karen refugees and evaluate such programs. In addition to health education programs on healthy lifestyle choices, tobacco cessation programs seem to be necessary for Karen refugees. At the same time, it is important to foster strategies to increase the utilization of preventive care among this population by promoting free or reduced-fee resources in the community to further promote their health.

A Direct Observation Study of Health Education Classesfor Uninsured Primary Care Patients

Lifestyle related health issues are a significant concern in the United States. In safety-net primary care clinics such as free clinics, health education programs have the potential to reduce the prevalence of lifestyle related health issues amongst vulnerable, low socioeconomic populations. The purpose of this study is to describe health education programs for underserved populations at a free clinic using direct observations based on the Theory of Planned Behavior (TPB). The primary data source of this study was field notes based on observations of health education classes at a free clinic. Nine health education classes were observed in October and November in 2016 (total number of participants=55). Participants were most receptive when conversations were dialogue instead of lecture based within the informal classes. Within the formal class, the lecture format proved to be effective. In both types of classes, participants felt they had adequate information on certain topics but lacked the accountability or will power necessary to make health-related behavior changes. Participants explained that the health education classes were effective in reminding them about certain aspects of health that they did not think about daily. The TPB seems to be a useful model to study health behavior change within the setting of free clinic informal waiting room classes. To promote behavioral changes, perceived behavioral control would be one of the key areas to focus on in health education classes. Future projects should develop health education programs which respond to the results of this study and evaluate the programs.

Geotagged US Tweets as Predictors of County-Level Health Outcomes, 2015–2016

Objectives To leverage geotagged Twitter data to create national indicators of the social environment, with small-area indicators of prevalent sentiment and social modeling of health behaviors, and to test associations with county-level health outcomes, while controlling for demographic characteristics. Methods We used Twitter’s streaming application programming interface to continuously collect a random 1% subset of publicly available geo-located tweets in the contiguous United States. We collected approximately 80 million geotagged tweets from 603 363 unique Twitter users in a 12-month period (April 2015-March 2016). Results Across 3135 US counties, Twitter indicators of happiness, food, and physical activity were associated with lower premature mortality, obesity, and physical inactivity. Alcohol-use tweets predicted higher alcohol-use-related mortality. Conclusions Social media represents a new type of real-time data that may enable public health officials to examine movement of norms, sentiment, and behaviors that may portend emerging issues or outbreaks-thus providing a way to intervene to prevent adverse health events and measure the impact of health interventions.

Physical Activity Education for Adults with Refugee Background in the United States

Objectives: This qualitative project aimed at examining factors preventing or promoting practice of frequent exercising among diverse refugee groups resettled in a metropolitan area in the Intermountain region. Methods: A total of 4 education sessions on physical exercises were successfully offered to individuals with refugee background who were interested in this topic. Pre- and post-class surveys, field notes, and focus group questionnaire were developed based on the Health Belief Model. Focus group discussions were conducted at the end of each education session. A total of 7 participants completed pre- and post-class surveys and 6 participants joined focused group discussions. In addition, there were physical activity class participants who were not eligible for the survey and a focus group. Results: The study revealed four important findings. Participants were knowledgeable of the health benefits of frequent exercising. Busy schedule, low energy, not knowing the importance of physical activity, pain, and unawareness of local resources were identified as barriers for not practicing exercise more often. Most participants indicated that they needed more physical activity and planned on adding more exercises to their daily activities. Participants expressed cultural differences in practice of physical exercise before and after they resettled in the US. Conclusion: This study provided insights on physical activity practice among a group of refugees resettled in the US. Future interventions should focus on providing a comprehensive education session, in combination with developing communitybased programs aiming at creating new or improving existing resources and facilities that are culturally appropriate for refugees from diverse background.

Home Dental Care Education for RefugeeBackground Adults in the United States

Objective: The purpose of this study was to provide and evaluate oral health care education programs for refugees resettled in the US. Methods: This project consisted of six sessions, which were held from February to April 2017. Each session included the following components: 1) a short survey that included demographic questions and oral health-related questions; 2) a class on oral health home care class; 3) a focus group; and 4) a post-class survey on class satisfaction. Participants were individuals who had a refugee background and were ages 18 and older at the time of the session. Results: Twenty-seven refugees from diverse ethnic backgrounds participated in this study. Refugees that have resettled in the US may not have had opportunities to learn about oral health care, but seem to be interested in oral health education, and find the information useful. While brushing teeth seems to be commonly practiced (though their methods of brushing may not be appropriate), flossing teeth is not. Before resettlement, participants had poor oral health practices and habits, lacked resources, and also maintained cultural norms that negatively affected their oral health. Conclusion: It is important to develop and provide educational programs to promote proper oral health practices for refugees. The changes in their environment after migration to the US such as unfamiliarity to dental health practice and the addition of sugary food/drinks to their lives should be considered in oral health education.

Barriers and Facilitators to Healthy Lifestyle among Refugees Resettled in the United States

Objective: The purpose of this project included: to provide health education for refugees and to gather perspectives on barriers; and facilitators toward healthy eating and regular physical activity. Methods: Six health promotion classes were held between February and June, 2017. Data were collected using a pre-class survey, focus groups, and field notes. Qualitative data were mapped to constructs of the Health Belief Model. Twenty-eight participants of diverse refugee populations participated in this project, of which 12 participated in 4 focus groups. Results: Desires to “stay healthy” and “to be there for family” appeared to be key facilitators for healthy lifestyle. Key barriers to healthy lifestyle include “busy schedule,” and “not able to afford healthy food.” Participants described their struggles to access affordable fresh food and believed having adopted the American diet caused adverse health outcomes. Throughout the project, participants showed interests in utilizing local parks but lacked information on access and use. Conclusion: Future health promotion interventions should focus on familiarizing refugee families with local produce whilst being culturally sensitive on different cooking methods and diet preferences. More health classes and opportunities for free to low-cost exercise are needed.

National substance use patterns on Twitter

Purpose We examined openly shared substance-related tweets to estimate prevalent sentiment around substance use and identify popular substance use activities. Additionally, we investigated associations between substance-related tweets and business characteristics and demographics at the zip code level. Methods A total of 79,848,992 tweets were collected from 48 states in the continental United States from April 2015-March 2016 through the Twitter API, of which 688,757 were identified as being related to substance use. We implemented a machine learning algorithm (maximum entropy text classifier) to estimate sentiment score for each tweet. Zip code level summaries of substance use tweets were created and merged with the 2013 Zip Code Business Patterns and 2010 US Census Data. Results Quality control analyses with a random subset of tweets yielded excellent agreement rates between computer generated and manually generated labels: 97%, 88%, 86%, 75% for underage engagement in substance use, alcohol, drug, and smoking tweets, respectively. Overall, 34.1% of all substance-related tweets were classified as happy. Alcohol was the most frequently tweeted substance, followed by marijuana. Regression results suggested more convenience stores in a zip code were associated with higher percentages of tweets about alcohol. Larger zip code population size and higher percentages of African Americans and Hispanics were associated with fewer tweets about substance use and underage engagement. Zip code economic disadvantage was associated with fewer alcohol tweets but more drug tweets. Conclusions The patterns in substance use mentions on Twitter differ by zip code economic and demographic characteristics. Online discussions have great potential to glorify and normalize risky behaviors. Health promotion and underage substance prevention efforts may include interactive social media campaigns to counter the social modeling of risky behaviors.