Is Rural Residency a Risk Factor for Overweight and Obesity for U.S. Children?
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Department of Family and Community Medicine, University of Illinois-Chicago, College of Medicine at Rockford, 1601 Parkview Avenue, Rockford, IL 61107. E-mail: email@example.com
Objective: Despite studies suggesting that there is a higher prevalence of overweight or obese children in rural areas in the U.S., there are no national studies comparing the prevalence levels of overweight or obese rural to metropolitan children. The objective of this research was to examine the hypothesis that living in a rural area is a risk factor for children being overweight or obese.
Research Methods and Procedures: Using the National Survey of Children's Heath, the prevalence of overweight and/or obese rural children was compared with that of children in metropolitan settings. Multivariate analyses were performed on the data to detect if differences varied by health services use factors or demographic factors, such as household income, gender, and race.
Results: Multivariate analysis revealed that overweight or obese children ≥5 years of age were more likely to live in rural rather than metropolitan areas (odds ratio = 1.252; 95% confidence interval, 1.248, 1.256). Rural overweight U.S. children ≥5 years of age of age were more likely than their metropolitan counterparts to: be white, live in households ≤200% of the federal poverty level, have no health insurance, have not received preventive health care in the past 12 months, be female, use a computer for non-school work >3 hours a day, and watch television for >3 hours a day. In addition, they were more likely to have comorbidities.
Discussion: Living in rural areas is a risk factor for children being overweight or obese.
Obesity has reached epidemic proportions among all Americans (1). Perhaps even more disturbing is the dramatic increase in the prevalence of overweight children (2, 3, 4, 5). Over the past few decades, the prevalence of overweight children between 6 and 11 years of age has doubled and even more worrisome tripled among those 12 to 17 years of age (6, 7, 8). Currently, 16% of children 6 to 19 years of age are overweight (9, 10, 11).
Obesity places an enormous burden on the health and economy of the U.S. with annual overall healthcare costs of approximately $100 billion (12). Overweight children, especially adolescents, are at greater risk for becoming obese adults and obesity in adulthood poses serious health risks (13, 14, 15, 16, 17, 18). An estimated 300,000 Americans die from obesity-related diseases each year; and as today's overweight children become adults, this death toll is likely to rise (19, 20, 21, 22, 23). In addition to future risk, children being overweight causes health problems during childhood (16, 24, 25, 26). These problems are caused by the mechanical stress of excess weight and the metabolic consequences of adipose tissue and fat inhibiting normal organ function (27). Less apparent, but maybe even more significant, are the psychological consequences associated with children being overweight (24, 25).
Many risk factors have been associated with children being overweight or obese (28, 29). Regional and local data also suggest that residency is a risk factor and that overweight is more prevalent in rural children and adolescents than their metropolitan counterparts (30). For example, a study of school children in rural Georgia found an increased risk for being overweight that could not be explained by demographic factors alone (31). Another study examining primarily rural practices in Michigan identified a higher prevalence of overweight children than in national studies, suggesting that rural residence might be a specific risk factor for children being overweight (32). Despite research that supports the notion that overweight children are more prevalent in rural areas, no national studies comparing rural to metropolitan children are available.
The purpose of this study was to examine the hypothesis that rural residence is an independent risk factor for children being overweight or obese. Using the 2003 National Survey of Children's Health (NSCH),1 a national database, the prevalence of overweight and obese children in rural settings was compared with that of children living in metropolitan ones. The analysis explored whether the detected differences varied by demographic characteristics as well as a number of other covariates such as hours watching television, hours of non-school computer use, and amounts of physical activity.
Research Methods and Procedures
Our study is a secondary analysis of data from the NSCH. Results were used to determine whether living in a rural area is a risk factor for being overweight or obese among children 5 to 18 years of age. NSCH is a random digit dial survey conducted by the Centers for Disease Control and Prevention's National Center for Health Statistics and is a part of the State and Local Area Integrated Telephone Survey program. For data collection, one child from each identified household was randomly selected and data were generated through a phone interview with the parent or guardian in the household who was the most familiar with the targeted child. For our study, a single year of data for the year 2003 to 2004, (n = 46,396, 7972 of which were overweight or obese) were analyzed. In our analysis, we included only those children who had an unsuppressed Metropolitan or Micropolitan Statistical Area (MSA) designation. For analysis, these data were weighted to represent 42,159,461 school-aged U.S. children, 7429,448 of whom were overweight or obese. The weighting variable is calculated by the Centers for Disease Control and Prevention using the most recently available census data to provide a stratified representation of the nation's children (33).
The NSCH collects data in eight domains including: demographics, physical and mental health status, health insurance, healthcare use and access to health care, medical home, family functioning, parent's health, and neighborhood characteristics. A full description of the NSCH and the development of sampling frames have been extensively described in other publications (33).
The data used in this analysis were weighted to be representative of the U.S. child population while maintaining complete subject anonymity. Consent for participation in the study was obtained from interview respondents as soon as it was determined that their household contained an age-eligible child. At that time, respondents were informed about the voluntary nature of the survey, the authorizing legislation, and confidentiality of data collected (34).
A number of the variables were re-coded to test our hypothesis. The variables, original survey questions used for data generation and the re-coded categories used in the analysis, are displayed in Table 1. Gender- and age-specific BMI percentiles were used to measure overweight or obesity among children. For this measure, the BMI number is calculated the same way for children as adults (BMI); however, the criteria used to interpret the meaning of the BMI number for children are different. Once the BMI is calculated, its percentile must be determined using a gender- and age-specific BMI percentile chart and then interpreted as follows: underweight, <5th percentile; normal weight, ≥5th percentile to <85th percentile; overweight, ≥85th to <95th percentile; obese, ≥95th percentile.
Table 1. . Variables, original survey questions, and recoded factors used in analysis: 2003 to 2004 NCHS data
|Age||Please tell me the [age/ages] of the [child/children]||5–9 years|
| ||less than 18 years old living in this household.||10–13 years|
| || ||14–17 years|
|Household income as percentage of federal poverty level||Two variables were used to determine a household's poverty status: the number of people residing in a household and the household's income during the prior year.||≤100% FPL 100–200% FPL >200% FPL|
|Child's non-school computer use||On an average school day, about how many hours does [child] use a computer for purposes other than schoolwork?||<1 hour Between 1 and 2 hours 3 or more hours|
|Child's TV use||On an average school day, about how many hours does [child] usually watch TV, watch videos, or play video games?||<1 hour Between 1 and 2 hours 3 or more hours|
|Comorbidities||Has a doctor or health professional ever told you that [child] has any of the following conditions?||No comorbidities 1 or more comorbidities|
| ||Asthma/Attention Deficit Disorder or Attention Deficit Hyperactive Disorder (ADD or ADHD)/depression or anxiety problems/bone, joint, or muscle problems/diabetes/autism|| |
|Child physical activity||During the past week, on how many days did [child] exercise or participate in physical activity for at least 20 minutes that made [him/her] sweat and breathe hard, such as basketball, soccer, running, swimming laps, fast bicycling, fast dancing, or similar aerobic activities?||Yes No (<5 days coded as “no” for not getting minimum amounts of moderate physical activity, ≥5 days coded as “yes” for getting minimum amounts of moderate physical activity)|
|Sports team participant||During the past 12 months, was [child] on a sports team or did [he/she] take sports lessons after school or on weekends?||Yes No|
|Have health plan||Does [child] have any kind of health care coverage, including health insurance, prepaid plans such as HMOs, or government plans such as Medicaid?||Yes No|
|Received all preventive care in past 12 months||Preventive care visits include things like a well-child check-up, a routine physical exam, immunizations, or health screening tests. During the past 12 months/since [child]'s birth, did [he/she] visit [his/her] personal doctor or nurse for preventive care?||Yes No|
|Have PCP||Do you have one or more persons you think of as [child]'s personal doctor or nurse?||Yes No|
|Race||Is [child] white, black or African-American, American Indian, Alaska Native, Asian, or Native Hawaiian or other Pacific Islander?||White Non-white|
|Sex||Is [child] male or female?||Male Female|
The definitions for metropolitan and rural used in this analysis were based on whether the respondent lived in or outside of an MSA. Those living within an MSA were considered to be urban or metropolitan residents and those outside an MSA were considered to be rural residents. These designations were made by the Centers for Disease Control and Prevention (33) and not changed for this analysis. In some instances, to ensure the anonymity of individual respondents, the rural/metropolitan designation was suppressed in the dataset. This suppression occurred whenever the sum total population for all MSA areas in a given state was <500,000 persons or whenever the sum total population for all of the non-MSA areas in a given state was <500,000 persons (33). A more detailed description of the survey parameters and definitions is described elsewhere (33).
Statistical Package for Social Scientists (SPSS, Inc., Chicago, IL), version 14.0, was used to complete all statistical analyses performed for this study. Univariate descriptions of and bivariate comparisons between rural and metropolitan children were made using either a χ2 or an unadjusted odds ratio (OR) to test for between- and/or within-groups differences. Multivariate logistic regression was performed to ascertain adjusted ORs for 2 different models: the first examining predictor variables for the outcome variable overweight or obese children and the second examining predictor variables for the outcome variable rural overweight or obese children. For all analyses, α was set at ≤0.05.
Univariate, bivariate, and multivariate analyses were performed on the data to detect if differences varied by health services use factors or demographic factors, such as household income, gender, race, and ethnic group. Univariate descriptions of and bivariate comparisons between rural and metropolitan children are presented in Table 2. Using either a χ2 or an unadjusted OR to examine between- and/or within-groups differences, the within-group differences for both rural and metropolitan overweight or obese children for hours of non-school computer use and hours of television (TV) use were more remarkable than the between-group differences. For example, the comparison of the TV watching habits of overweight or obese rural to metropolitan children yielded no between-group differences (Table 2). In both groups, about 17% of overweight or obese rural and metropolitan children watched <1 hour of TV, 57% watched between 1 and 2 hours of television daily, and 25% of both groups of overweight or obese children watched 3 or more hours of TV (Table 2). The within-group differences for levels of TV watching, however, revealed significant differences (Table 2).
Table 2. . Bivariate predictor variable comparison of rural and metropolitan overweight or obese children ≥5 years of age (2003 to 2004 NCHS data; unweighted, n = 7972)*
|Age group|| || || |
| 5–9 years||24.6%||27.3%|| |
| 10–13 years||33.1%||31.8%||0.00|
| 14–17 years||42.3%||40.9%|| |
|Household income as percent of federal poverty level (FPL)|| || || |
| ≤100% FPL||35.2%||33.0%|| |
| 100–200% FPL||43.1%||35.0%||0.00|
| >200% FPL||21.7%||32.0%|| |
|Child's non-school computer use|| || || |
| <1 hour||55.4%||54.8%|| |
| Between 1 and 2 hours||36.3%||35.4%||0.00|
| 3 or more hours||8.3%||9.7%|| |
|Child's TV use|| || || |
| <1 hour||17.3%||17.7%|| |
| Between 1 and 2 hours||57.8%||57.3%||0.00|
| 3 or more hours||24.9%||25.0%|| |
| || ||95% CI|
| ||Unadjusted odds ratios||Lower||Upper|
|Comorbidities|| || || |
| No comorbidities||1.05 In favor of rural children||1.05||1.06|
|Child physical activity|| || || |
| Not meeting recommended levels of moderate PA||1.046 In favor of metro children||1.045||1.047|
|Sports team participant|| || || |
| No||1.013 In favor of metro children||1.012||1.014|
|Have health plan|| || || |
| No||1.016 In favor of metro children||1.015||1.017|
|Received all preventive care in past 12 months|| || || |
| No||1.371 In favor of rural children||1.367||1.375|
|Have PCP|| || || |
| No||1.035 In favor of metro children||1.034||1.036|
|Race|| || || |
| White||1.502 In favor of rural children||1.498||1.507|
For household income, the between-group differences were pronounced with only 21.7% of rural overweight or obese children living in households with incomes >200% of the federal poverty level (FPL) in comparison to 32.0% of metropolitan children. Additionally, in comparison to their metropolitan counterparts, rural children who were overweight or obese were more likely to have not received preventive care in the past 12 months [OR = 1.371; 95% confidence interval (CI), 1.367, 1.375]. Overweight or obese rural children were also more likely to be white than overweight or obese metropolitan children (OR = 1.502; 95% CI, 1.48, 1.509).
Two multivariate logistic regression models were tested. Overweight or obese U.S. children was the dependent variable in the first model (Table 3), which yielded that in 2003 to 2004 U.S. children ≥5 years of age who were overweight or obese were more likely to: live in households with <100% FPL (OR = 1.371; 95% CI, 1.366, 1.377) and 100% ≥200% FPL (OR = 1.384; 95% CI, 1.380, 1.388) than >200% FPL; not get recommended levels of moderate physical activity (OR = 1.467; 95% CI, 1.463, 1.472); not participate on a sports team (OR = 1.323; 95% CI, 1.320, 1.327); not have health insurance (OR = 1.317; 95% CI, 1.310, 1.324); and be male (OR = 1.314; 95% CI, 1.338, 1.344). Additionally, this analysis showed that overweight or obese school-aged children were more likely to live in rural rather than metropolitan areas (OR = 1.252; 95% CI, 1.248, 1.256). These same children were less likely to: use a computer for non-school work <1 hour (OR = 0.574; 95% CI, 0.572, 0.577) or between 1 and 2 hours (OR = 0.649; 95% CI, 0.646, 0.651) rather than ≥3 hours a day; watch television <1 hour (OR = 0.463; 95% CI, 0.461, 0.465) or between 1 and 2 hours (OR = 0.654; 95% CI, 0.652, 0.656) rather than ≥3 hours a day; have no comorbidities (OR = 0.670; 95% CI, 0.668, 0.671) rather than at least 1 comorbidity; and be white (OR = 0.662; 95% CI, 0.660, 0.665) rather than non-white. In this analysis, demographic variables explained 85% of the overall increased risk for children being overweight or obese.
Table 3. . Overweight or obese U.S. children ≥5 years of age: logistic regression model (2003 to 2004 NCHS data)
|Child physical activity (vs. meeting at least recommended levels of moderate physical activity)||Not meeting recommended levels of moderate physical activity||1.467||1.463||1.472|
|Household income as percentage of FPL (vs. >200% FPL)||≤100% FPL||1.371||1.366||1.377|
| ||100–200% FPL||1.384||1.380||1.388|
|Sex (vs. female)||Male||1.341||1.338||1.344|
|Sports team participant (vs. yes)||No||1.323||1.320||1.327|
|Health plan (vs. yes)||No||1.317||1.310||1.324|
|Metropolitan statistical area (vs. metropolitan)||Rural||1.252||1.248||1.256|
|Preventive care in past 12 months (vs. yes)||No||1.047||1.044||1.050|
|Comorbidities (vs. at least 1 comorbidity)||No comorbidities||0.670||0.668||0.671|
|Race (vs. non-white)||White||0.662||0.660||0.665|
|Child's non-school computer use (vs. 3 or more hours)||<1 hour||0.574||0.572||0.577|
| ||Between 1 and 2 hours||0.649||0.646||0.651|
|Child's TV use (vs. 3 or more hours)||<1 hour||0.463||0.461||0.465|
| ||Between 1 and 2 hours||0.654||0.652||0.656|
The dependent variable for the second model was rural U.S. children ≥5 years of age who were overweight or obese (Table 4). This analysis is summarized in Table 4. This analysis revealed that rural U.S. children ≥5 years of age who were overweight or obese were more likely than their metropolitan counterparts to have the following characteristics: be white rather than non-white (OR = 1.418; 95% CI, 1.411, 1.424); be more likely to live below or slightly above the poverty level; be uninsured; and have not received preventive health care in the past 12 months. In addition, compared with overweight or obese metropolitan children, rural overweight, or obese children were more likely to be female and to use a computer for non-school work >3 hours a day. They were also more likely to watch TV for >3 hours a day and less likely to have no comorbidities (OR = 0.801; 95% CI, 0.798, 0.805). In this logistics regression model, demographic variables explained 75% of the increased risk for rural children being overweight or obese.
Table 4. . Rural (vs. metropolitan) overweight or obese U.S. children ≥5 years of age: logistic regression model (2003 to 2004 NCHS data)
|Predictive variables and factors||Adjusted odds ratio||Lower||Upper|
|Race (vs. non-white)||White||1.418||1.411||1.424|
|Household income as percentage of federal poverty level (vs. >200% FPL)||≤100% FPL||1.433||1.426||1.441|
| ||100–200% FPL||1.372||1.366||1.379|
|Health plan (vs. yes)||No||1.218||1.208||1.228|
|Received preventive care in past 12 months (vs. yes)||No||1.230||1.224||1.235|
|Sex (vs. female)||Male||0.889||0.885||0.893|
|Comorbidities (vs. at least 1 comorbidity)||No comorbidities||0.801||0.798||0.805|
|Child's TV use (vs. 3 or more hours)||<1 hour||0.646||0.642||0.650|
| ||Between 1 and 2 hours||0.656||0.653||0.660|
|Child's non-school computer use (vs. 3 or more hours)||<1 hour||0.412||0.409||0.414|
| ||Between 1 and 2 hours||0.386||0.384||0.388|
|Child physical activity (vs. not getting recommended levels of moderate PA)||Getting at least recommended levels of moderate physical activity||0.648||0.645||0.651|
|Sports team participant (vs. no)||Yes||0.843||0.839||0.846|
The major finding of this study is that children living in rural areas in the U.S. are about 25% (OR = 1.252; CI, 1.248, 1.256) more likely to be overweight or obese than their metropolitan counterparts. In the dataset used, known risk factors for children being overweight such as physical inactivity, television watching, and computer use accounted for some but not all of the increased risk, suggesting that rural residency is an independent childhood risk factor for being overweight or obese. Although previous smaller regional studies (32, 35, 36, 37) found that children living in rural settings were more likely to be overweight, this is the first known study to use national data to explore the link between rural residence and childhood overweight or obesity.
It is established that health is unevenly distributed across factors such as socioeconomic status (38) and that lower income groups experience poorer health outcomes (39). Only more recently has geographic location been recognized as creating its own set of health issues and that there might be a health benefit to non-rural residence (40). A national study linked adult obesity to rural residency across virtually every racial and ethnic group (41). Combining those findings with our results makes the link between overweight children and rural residence as an independent risk factor even more compelling. This finding represents a change from the past when children from large metropolitan centers were at greater risk for being overweight than rural children (42, 43, 44, 45, 46, 47).
Our results suggest that not only is rural residency a risk factor for overweight in children, but that overweight rural children had the additional risk factors of poverty, no health insurance, no preventive care in the past year, and little physical activity. These demographic findings are consistent with previous research and are not unexpected. However, in the multivariate analyses these factors do not account for all of the increased risk of being at-risk-for- overweight or overweight, and the exact reason why rural children are more likely to be at-risk-for-overweight or overweight is uncertain. One factor may be that the rural “built environment” (48) creates an obesegenic rural ecology (49). The built environment is loosely defined as those aspects of the environment that are created or modified by humans such as homes, schools, workplaces, parks, industrial areas, farms, food sources, and roads and highways (48). Components of the built environment include transportation systems and other features that influence physical activity. There is evidence that the built environment can promote a sedentary lifestyle and that children being overweight is linked to physical inactivity and a sedentary lifestyle (48, 49, 50, 51, 52). Similar to past research, our study found that rural children are less likely to be physically active than their metropolitan counterparts and this may be related to features of the rural environment that create special challenges for rural children to be physically active (51). These include limited access to parks, exercise facilities (51), fewer sidewalks, a lack of public transportation, and limited physical education classes (53, 54). However, research about the built environment is still in its infancy (55), and more research is needed to determine the role of the built environment in promoting obesity.
The rural setting may also have additional elements that make it obesegenic. These additional elements might include less access to healthy food, increased cost of fruits and vegetables, fewer opportunities for education, and limited access to dietitians or other nutrition experts. Although there is substantial speculation about the causes for the epidemic of obesity, there is still little available evidence to disentangle the reasons (3).
Another interesting finding was that rural children were almost 50% less likely to have a preventive healthcare visit and also less likely to have health insurance. Primary care provider visits can be a source of information about healthy lifestyle, including diet and exercise. Because smaller rural schools also offer fewer nutrition services (56), the loss of educational opportunities resulting from fewer preventive care visits in rural locations may have a more profound effect than in a metropolitan area. Improving access for preventive care and highlighting the importance to rural practitioners of addressing healthy lifestyles to rural children and parents may be of even more benefit than in a metropolitan setting.
There are several limitations to this study. First, the data are self-reported and subject to error. However, the bias would most likely be to under-report overweight since there are negative social connotations to being labeled as overweight (24). This suggests that the prevalence of overweight might be even higher. However, there is no reason to suspect that errors in self-report vary by characteristics likely to impact rural/metropolitan differences. Although several confounding variables associated with children being at-risk-for-overweight or overweight were accounted for, it is possible other unidentified factors may account for the increased risk of overweight rather than the rural residence. For example, the composition of the rural adult community who tend to be less educated and more likely to be overweight than their metropolitan counterparts may account for the increase in prevalence of children being overweight. These or other compositional factors rather than the context of rural living may account for the excess incidence of being at-risk-for-overweight or overweight. However, even though some degree of childhood overweight might be attributed to the demographics of their parents, the fact that several small studies also identified an increased prevalence of rural overweight children, combined with the magnitude of the OR in our study, suggests that the impact of rural residency on a child's weight is real. In addition, studies demonstrating that the economic prosperity of a region interacts with socioeconomic status in determining risk for becoming overweight illustrate the influence of context (57).
The operational definitions of rural and metropolitan used in this study are a final limitation. These definitions represent gross geographic distinctions, not community specific ones. We fully recognize that rural communities, like metropolitan ones, vary greatly from one another and should be seen as heterogeneous rather than homogeneous. There is a considerable literature from rural sociology addressing the fallacy or myth of homogeneous rural communities (58). The definitions of rural and metropolitan used in this study, however, are ones imposed by the data and cannot be changed.
In summary, the U.S. now has the highest percentage of overweight youth in its history, and this study suggests that rural residency may be a contributing factor. The added weight places a burden on the health of our children now and also when they become adults (59). Identifying rural residency as a risk factor for children suggests that strategies to reduce the prevalence of overweight or obese children should consider interventions targeted to children living in rural settings. Physicians practicing in rural settings should also consider strategies they might adapt to their practices that address rural obesity. However, because many of these children have the additional risk factors of being poor and uninsured, physician offices may not be the best site for interventions targeted at overweight rural children. Public health interventions entailing partnerships between local schools and local family physicians or pediatricians might have a better chance of successfully reaching the targeted population. Future research to unravel the environmental factors contributing to rural obesity may aid in developing effective strategies geared to rural settings for preventing obesity.
There was no funding/outside support for this study.
Nonstandard abbreviations: NSCH, National Survey of Children's Health; MSA, metropolitan or micropolitan statistical area; OR, odds ratio; TV, television; FPL, federal poverty level; CI, confidence interval.