Faculty Bibliography

BACKGROUND:The decisions that individuals make about the food and beverage products they purchase and consume directly influence their energy intake and dietary quality and may lead to excess weight gain and obesity. However, gathering and interpreting data on food and beverage purchase patterns can be difficult. Leveraging novel sources of data on food and beverage purchase behavior can provide us with a more objective understanding of food consumption behaviors. OBJECTIVE:Food and beverage purchase receipts often include time-stamped location information, which, when associated with product purchase details, can provide a useful behavioral measurement tool. The purpose of this study was to assess the feasibility, reliability, and validity of processing data from fast-food restaurant receipts using crowdsourcing via Amazon Mechanical Turk (MTurk). METHODS:Between 2013 and 2014, receipts (N=12,165) from consumer purchases were collected at 60 different locations of five fast-food restaurant chains in New Jersey and New York City, USA (ie, Burger King, KFC, McDonald's, Subway, and Wendy's). Data containing the restaurant name, location, receipt ID, food items purchased, price, and other information were manually entered into an MS Access database and checked for accuracy by a second reviewer; this was considered the gold standard. To assess the feasibility of coding receipt data via MTurk, a prototype set of receipts (N=196) was selected. For each receipt, 5 turkers were asked to (1) identify the receipt identifier and the name of the restaurant and (2) indicate whether a beverage was listed in the receipt; if yes, they were to categorize the beverage as cold (eg, soda or energy drink) or hot (eg, coffee or tea). Interturker agreement for specific questions (eg, restaurant name and beverage inclusion) and agreement between turker consensus responses and the gold standard values in the manually entered dataset were calculated. RESULTS:Among the 196 receipts completed by turkers, the interturker agreement was 100% (196/196) for restaurant names (eg, Burger King, McDonald's, and Subway), 98.5% (193/196) for beverage inclusion (ie, hot, cold, or none), 92.3% (181/196) for types of hot beverage (eg, hot coffee or hot tea), and 87.2% (171/196) for types of cold beverage (eg, Coke or bottled water). When compared with the gold standard data, the agreement level was 100% (196/196) for restaurant name, 99.5% (195/196) for beverage inclusion, and 99.5% (195/196) for beverage types. CONCLUSIONS:Our findings indicated high interrater agreement for questions across difficulty levels (eg, single- vs binary- vs multiple-choice items). Compared with traditional methods for coding receipt data, MTurk can produce excellent-quality data in a lower-cost, more time-efficient manner.

Demographic and income disparities may impact food accessibility. Research has not yet well documented the precise location of healthy and unhealthy food resources around children's homes and schools. The objective of this study was to examine the food environment around homes and schools for all public school children, stratified by race/ethnicity and poverty status. This cross-sectional study linked data on the exact home and school addresses of a population-based sample of public school children in New York City from 2013 to all corner stores, supermarkets, fast-food restaurants, and wait-service restaurants. Two measures were created around these addresses for all children: 1) distance to the nearest outlet, and 2) count of outlets within 0.25 miles. The total analytic sample included 789,520 K-12 graders. The average age was 11.78 years (SD ± 4.0 years). Black, Hispanic, and Asian students live and attend schools closer to nearly all food outlet types than White students, regardless of poverty status. Among not low-income students, Black, Hispanic, and Asian students were closer from home and school to corner stores and supermarkets, and had more supermarkets around school than White students. The context in which children live matters, and more nuanced data is important for development of appropriate solutions for childhood obesity. Future research should examine disparities in the food environment in other geographies and by other demographic characteristics, and then link these differences to health outcomes like body mass index. These findings can be used to better understand disparities in food access and to help design policies intended to promote healthy eating among children.

Background/UNASSIGNED:Tobacco spending may exacerbate financial hardship in low-income populations by using funds that could go toward essentials. This study examined post-quit spending plans among low-income smokers and whether financial hardship was positively associated with motivation to quit in the sample. Methods/UNASSIGNED:= 410). Linear regression was used to examine the relationship between financial distress, food insecurity, smoking-induced deprivation (SID) and motivation to quit (measured on a 0-10 scale). We performed summative content analyses of open-ended survey questions to identify the most common plans among participants with and without SID for how to use their tobacco money after quitting. Results/UNASSIGNED:The top three spending plans among participants with and without SID were travel, clothing and savings. There were three needs-based spending plans unique to a small number of participants with SID: housing, health care and education. Conclusions/UNASSIGNED:Financial distress and food insecurity did not enhance overall motivation to quit, while smokers with SID were less motivated to quit. Most low-income smokers, including those with SID, did not plan to use their tobacco money on household essentials after quitting.

[This corrects the article DOI: 10.1371/journal.pone.0215571.].

BACKGROUND:Because of the strong link between childhood obesity and adulthood obesity comorbidities, and the difficulty in decreasing body mass index (BMI) later in life, effective strategies are needed to address this condition in early childhood. The ability to predict obesity before age five could be a useful tool, allowing prevention strategies to focus on high risk children. The few existing prediction models for obesity in childhood have primarily employed data from longitudinal cohort studies, relying on difficult to collect data that are not readily available to all practitioners. Instead, we utilized real-world unaugmented electronic health record (EHR) data from the first two years of life to predict obesity status at age five, an approach not yet taken in pediatric obesity research. METHODS AND FINDINGS/RESULTS:We trained a variety of machine learning algorithms to perform both binary classification and regression. Following previous studies demonstrating different obesity determinants for boys and girls, we similarly developed separate models for both groups. In each of the separate models for boys and girls we found that weight for length z-score, BMI between 19 and 24 months, and the last BMI measure recorded before age two were the most important features for prediction. The best performing models were able to predict obesity with an Area Under the Receiver Operator Characteristic Curve (AUC) of 81.7% for girls and 76.1% for boys. CONCLUSIONS:We were able to predict obesity at age five using EHR data with an AUC comparable to cohort-based studies, reducing the need for investment in additional data collection. Our results suggest that machine learning approaches for predicting future childhood obesity using EHR data could improve the ability of clinicians and researchers to drive future policy, intervention design, and the decision-making process in a clinical setting.

The objective of this study was to measure change in obesity prevalence among New York City (NYC) adults from 2004 to 2013-2014 and assess variation across sociodemographic subgroups. We used objectively measured height and weight data from the NYC Health and Nutrition Examination Survey to calculate relative percent change in obesity (≥ 30 kg/m²) between 2004 (n = 1987) and 2013-2014 (n = 1489) among all NYC adults and sociodemographic subgroups. We also examined changes in self-reported proxies for energy imbalance. Estimates were age-standardized and statistical significance was evaluated using two-tailed T tests and multivariable regression (p < 0.05). Between 2004 and 2013-2014, obesity increased from 27.5 to 32.4% (p = 0.01). Prevalence remained stable and high among women (31.2 to 32.8%, p = 0.53), but increased among men (23.4 to 32.0%, p = 0.002), especially among non-Latino White men and men age ≥ 65 years. Black adults had the highest prevalence in 2013-2014 (37.1%) and Asian adults experienced the largest increase (20.1 to 29.2%, p = 0.06), especially Asian women. Foreign-born participants and participants lacking health insurance also had large increases in obesity. We observed increases in eating out and screen time over time and no improvements in physical activity. Our findings show increases in obesity in NYC in the past decade, with important sociodemographic differences.

Readers should note the following two typographical errors in this article.

Research has examined how the food environment affects the risk of cardiovascular disease (CVD). Many studies have focused on residential neighbourhoods, neglecting the activity spaces of individuals. The objective of this study was to investigate whether food environments in both residential and global positioning system (GPS)-defined activity space buffers are associated with body mass index (BMI) and blood pressure (BP) among low-income adults. Data came from the New York City Low Income Housing, Neighborhoods and Health Study, including BMI and BP data (n=102, age=39.3±14.1 years), and one week of GPS data. Five food environment variables around residential and GPS buffers included: fast-food restaurants, wait-service restaurants, corner stores, grocery stores, and supermarkets. We examined associations between food environments and BMI, systolic and diastolic BP, controlling for individual- and neighbourhood-level sociodemographics and population density. Within residential buffers, a higher grocery store density was associated with lower BMI (β=- 0.20 kg/m2, P<0.05), and systolic and diastolic BP (β =-1.16 mm Hg; and β=-1.02 mm Hg, P<0.01, respectively). In contrast, a higher supermarket density was associated with higher systolic and diastolic BP (β=1.74 mm Hg, P<0.05; and β=1.68, P<0.01, respectively) within residential buffers. In GPS neighbourhoods, no associations were documented. Examining how food environments are associated with CVD risk and how differences in relationships vary by buffer types have the potential to shed light on determinants of CVD risk. Further research is needed to investigate these relationships, including refined measures of spatial accessibility/exposure, considering individual's mobility.

BACKGROUND:Local food environments include food stores (eg, supermarkets, grocery stores, bakeries) and restaurants. However, the extent to which other storefront businesses offer food/drink is not well described, nor is the extent to which food/drink availability through a full range of storefront businesses might change over time. OBJECTIVES/OBJECTIVE:This study aimed to assess food/drink availability from a full range of storefront businesses and the change over time and to consider implications for food-environment research. DESIGN/METHODS:Investigators compared direct observations from 2010 and 2015. PARTICIPANTS/SETTING/METHODS:Included were all storefront businesses offering foods/drinks on 153 street segments in the Bronx, NY. MAIN OUTCOME MEASURES/METHODS:The main outcome was change between 2010 and 2015 as determined by matches between businesses. Matches could be strict (businesses with the same name on the same street segment in both years) or lenient (similar businesses on the same street segment in both years). Investigators categorized businesses as general grocers, specialty food stores, restaurants, or other storefront businesses (eg, barber shops/beauty salons, clothing outlets, hardware stores, laundromats, and newsstands). STATISTICAL ANALYSES PERFORMED/METHODS:Investigators quantified change, specifically calculating how often businesses in 2015 were present in 2010 and vice versa. RESULTS:Strict matches for businesses in 2015 present in 2010 ranged from 29% to 52%, depending on business category; lenient matches ranged from 43% to 72%. Strict matches for businesses in 2010 present in 2015 ranged from 34% to 63%; lenient matches ranged from 72% to 83%. In 2015 compared with 2010, on 22% more of the sampled street segments, 30% more businesses were offering food/drink: 66 vs 46 general grocers, 22 vs 19 specialty food stores, 99 vs 99 restaurants, 98 vs 56 other storefront businesses. CONCLUSIONS:Over 5 years, an urban food environment changed substantially, even by lenient standards, particularly among "other storefront businesses" and in the direction of markedly greater food availability (more businesses offering food on more streets). Failure to consider a full range of food/drink sources and change in food/drink sources could result in erroneous food-environment conclusions.

BACKGROUND:Excess calorie consumption and poor diet are major contributors to the obesity epidemic. Food retailers, in particular at supermarkets, are key shapers of the food environment which influences consumers' diets. This study seeks to understand the decision-making processes of supermarket retailers-including motivators for and barriers to promoting more healthy products-and to catalogue elements of the complex relationships between customers, suppliers, and, supermarket retailers. METHODS:We recruited 20 supermarket retailers from a convenience sample of full service supermarkets and national supermarket chain headquarters serving low- and high-income consumers in urban and non-urban areas of New York. Individuals responsible for making in-store decisions about retail practices engaged in online surveys and semi-structured interviews. We employed thematic analysis to analyze the transcripts. RESULTS:Supermarket retailers, mostly representing independent stores, perceived customer demand and suppliers' product availability and deals as key factors influencing their in-store practices around product selection, placement, pricing, and promotion. Unexpectedly, retailers expressed a high level of autonomy when making decisions about food retail strategies. Overall, retailers described a willingness to engage in healthy food retail and a desire for greater support from healthy food retail initiatives. CONCLUSIONS:Understanding retailers' in-store decision making will allow development of targeted healthy food retail policy approaches and interventions, and provide important insights into how to improve the food environment.