Advertisement
Research Article

TV Viewing and BMI by Race/Ethnicity and Socio-Economic Status

  • Kerem Shuval mail,

    Kerem.Shuval@uth.tmc.edu

    Affiliations: Division of Epidemiology, Human Genetics and Environmental Sciences, University of Texas Health Science Center at Houston, Dallas Regional Campus, Dallas, Texas, United States of America, Harold C Simmons Cancer Center, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, United States of America

    X
  • Kelley Pettee Gabriel,

    Affiliation: Division of Epidemiology, Human Genetics and Environmental Sciences, University of Texas Health Science Center at Houston, Austin Regional Campus, Austin, Texas, United States of America

    X
  • Tammy Leonard

    Affiliation: The Department of Economics, University of Texas at Dallas, Dallas, Texas, United States of America

    X
  • Published: May 15, 2013
  • DOI: 10.1371/journal.pone.0063579

Abstract

Objective

To assess the association between TV viewing and obesity by race/ethnicity and socio-economic status.

Design

Cross-sectional analysis of 5,087 respondents to the Health Information National Trends Survey (HINTS), a nationally representative sample of US adults. Multivariate regression models were computed to assess the association between quartiles of TV viewing and BMI, stratified by race/ethnicity, educational attainment, employment and health insurance status.

Results

Findings indicate that increased TV viewing was associated with higher odds for being overweight/obese in the entire sample, while adjusting for physical activity and other confounders. After stratification by race/ethnicity, increased odds for overweight/obesity in the 3rd and 4th quartiles of TV viewing (e.g., 3rd quartile- cumulative OR = 1.43, 95%CI 1.07–1.92) was observed in non-Hispanic whites, with statistical significance. In non-Hispanic blacks and Hispanics, the odds were similar to whites, but did not reach statistical significance. Significant relations between greater TV viewing and increased BMI were observed in college graduates and non-graduates, those with health insurance and the employed.

Conclusions

This study extends previous research by examining potential inconsistencies in this association between various racial/ethnic groups and some socio-economic variables, which primarily were not found.

Introduction

Technological advancements over the past several decades have led to a constant decrease in individual- and population- level energy expenditure, resulting in increasing rates of obesity and other related chronic diseases [1]. The research focus for years has been on the health benefits of moderate- to vigorous- intensity physical activity as a means to prevent or delay the onset of morbidity and mortality [2]. In the past decade, however, numerous studies have emerged demonstrating the effects of prolonged sedentary time (e.g., TV viewing, computer United States of Americage) on increased risk for obesity, diabetes, metabolic syndrome, cancer and mortality while adjusting for physical activity [3][6]. For example, Hu et al. (2003) found in the Nurses’ Health Study that a 2-hour increase in TV viewing per day increased the risk of obesity by 23% and diabetes by 14% during a 6 year follow-up period [7]. These studies, however, have primarily focused on ethnic majority populations (e.g., predominately white females in the Nurses’ Health Study) rather than ethnic minorities, who experience a disproportionately higher prevalence of obesity and other chronic diseases (e.g., diabetes) [4][9]. Therefore, the present study examines the association between TV viewing and obesity among adults responding to the Health Information National Trends Survey (HINTS) [10], while assessing how this association might differ based on race/ethnicity and socio-economic status.

Methods

The HINTS survey, described elaborately elsewhere [10], [11], is a nationally representative survey of US adults aged 18 years and older conducted by the National Cancer Institute with the aim of gleaning information pertaining to health communications, cancer knowledge, and behaviors related to cancer prevention and control. The HINTS used a list assisted random-digit-dial (RDD) sampling plan of the general adult population (i.e., all phone exchanges in the US) [12]; where one adult (aged ≥18 years) from each household was selected for an extended interview via a household screener [13]. Both non-Hispanic blacks and Hispanics were oversampled and data were weighted in order to be nationally representative; i.e., areas estimated to have >15% of non-Hispanic blacks and Hispanics were oversampled at a higher rate in an attempt to increase response rates [12], [14]. Weights are provided by the HINTS study for data analysis [10], [15].

In the current study, we examined the association between TV viewing and obesity among respondents to the HINTS 2005 survey (n = 5,586). The HINTS 2005 response rate was 34.0% for the screener and 61.2% for the extended interview [12], [13]. A total of 5,087 participants with complete information on TV viewing, height and weight, and covariates (e.g., physical activity, race/ethnicity, employment), were included in the analytic sample. Due to the large number of participants with missing information on income (n = 1,389), this covariate was not included; adjusting for this variable did not change results materially. The study received exempt status from the University of Texas Health Science Center at Houston Institutional Review Board.

Measures

Dependent variable.

BMI was calculated using the standard formula (kg/m2) based on participants’ self-reported height and weight. Participants were categorized into three groups: 1) BMI <25 kg/m2: neither overweight nor obese; 2) BMI 25–29.9 kg/m2: overweight; and 3) BMI ≥30 kg/m2: obese.

Primary independent variable.

TV viewing was gleaned from respondents’ answer to the number of hours spent watching TV on a typical weekday and during a typical weekend (i.e., both Saturday and Sunday). TV viewing hours per day (h/d) was calculated using the following formula: (typical weekday TV viewing hours multiplied by 5 days)+(TV viewing hours on the weekend)/7 days. For analysis, quartiles of TV viewing were determined based on the full population sample of the HINTS 2005 survey with available data for TV viewing; computed quartiles are: 0.00–1.70; 1.71–2.60; 2.61–3.70; >3.70 h/d.

Race/Ethnicity and Socio-Economic Variables.

Based on responses related to race and ethnicity, participants were classified as non-Hispanic whites, non-Hispanic blacks, and Hispanics (of any race). Other racial/ethnic groups were categorized as ‘others’. Participants’ education levels were dichotomized according to whether or not they completed a college degree as was employment status (i.e., employed or unemployed), and health insurance coverage (yes/no).

Other Covariates: Covariates included age, gender, marital status (married/not married), children <18 years (yes/no), smoking (current, former, never), self-reported health status (excellent, very good, good, fair or poor), and physical activity. Physical activity was dichotomized into meeting guidelines (i.e., ≥150 minute of moderate intensity physical activity) or not (i.e., <150 minutes a week) based on reported frequency and duration of moderate-intensity physical activity [2].

Statistical Analysis

A multivariate ordered logistic regression model was used to examine the relationship between TV viewing (primary independent variable) and BMI (dependent variable). First the association was examined in the entire sample, while controlling for all covariates (age, gender, marital status, children <18 years, race/ethnicity, education, employment, health insurance, health status, smoking, and physical activity). Then, similar models were estimated while stratifying for education, employment, health insurance, and race/ethnicity. Given the insufficient sample size in the “other” race/ethnicity group, the association between TV viewing and BMI, stratified by race/ethnicity, was examined in the Hispanic, non-Hispanic black, and non-Hispanic white sub-groups only. In multivariate analysis, ordered logistic regression was utilized due to the natural order of the dependent variables, i.e., higher categories suggest increased odds for overweight or obesity [16], and jackknife weights (provided by the HINTS study) were used for calculating standard errors [10]. Results are reported as adjusted cumulative odds ratios (OR) and 95% confidence intervals (CI) for transition to higher BMI categories versus remaining in the same category. STATA 12 (STATA, College Station, TX) was utilized for statistical analysis.

Results

The analytic sample’s mean age was 52 (SD = 18) years, 66% were women, and 38% had a college degree (Table 1). More than half (59%) were married, 88% had health insurance, and 43% perceived their health status to be either very good or excellent. More than two-thirds (68%) met public health recommendations for physical activity, whereas 46% watched TV >2.6 h/d, and 62% were either overweight or obese. In addition, non-Hispanic blacks spent 1.5 and 1.6 time more hours per day watching TV than non-Hispanic whites and Hispanics, respectively (i.e., non-Hispanic blacks-4.3 h/d, non-Hispanic whites- 2.9 h/d, and Hispanics- 2.7- h/d; P<0.001, P<0.001 respectively). Similarly, non-Hispanic blacks had a significantly higher BMI than non-Hispanic whites (P<0.001) and Hispanics (P<0.001).

thumbnail

Table 1. Descriptive Characteristics of 2005 HINTS Analytic Sample (n = 5,087).

doi:10.1371/journal.pone.0063579.t001

In the full analytic sample, the odds for being overweight/obese increased in the 3rd (2.61–3.70 h/d) and 4th (>3.70 h/d) quartiles of TV viewing in comparison to the 1st quartile (cumulative OR = 1.35, 95%CI 1.04–1.75; cumulative OR = 1.67, 95%CI 1.29–2.16; respectively) (Table 2). When stratifying by race/ethnicity, in non-Hispanic whites the 3rd and 4th quartiles of TV viewing significantly increased the odds for being overweight/obese (cumulative OR = 1.43, 95%CI 1.07–1.92; cumulative OR = 1.79; 95%CI = 1.39–2.31; respectively). In non-Hispanic blacks, increased odds were observed for the 3rd and 4th quartile of TV viewing (e.g., 4th quartile- cumulative OR = 2.14; 95%CI 0.90–5.13), yet without statistical significance. In Hispanics, TV viewing was associated with higher odds for being overweight/obese in the 2nd, 3rd and 4th quartiles; however without statistical significance (e.g., 4th quartile- cumulative OR = 1.27; 95%CI 0.52–3.11).

thumbnail

Table 2. Multivariate Ordered Logistic regressiona for overweight/obesity according to TV viewing- Full sample and stratified by Race/Ethnicity.

doi:10.1371/journal.pone.0063579.t002

Stratification by education revealed that both college graduates and non-graduates were at increased odds for overweight/obesity in the 4th quartile of TV viewing (Table 3). Moreover, those employed were at increased risk for being overweight/obese in the 3rd and 4th quartile of TV viewing (e.g., 4th quartile- cumulative OR = 1.64; 95%CI 1.26–2.14); while the unemployed were not (e.g., 4th quartile- cumulative OR = 1.01; 95%CI 0.20–5.00). When stratifying by health insurance, participants had increased odds for being overweight/obese when viewing more TV per day regardless of insurance status; however this relation was only statistically significant among those with insurance (Table 3).

thumbnail

Table 3. Multivariate Ordered Logistic regressiona for overweight/obesity according to TV viewing Stratified by Education, Employment, and Health Insurance.

doi:10.1371/journal.pone.0063579.t003

Discussion

To our knowledge, few studies have explored potential variations in the association between time spent watching TV and BMI by race/ethnicity and other socio-economic variables within a nationally representative sample in the US. Our findings indicate that the observed positive association between TV viewing and risk for being overweight/obese varies only somewhat by race/ethnicity and socio-economic variables. TV viewing in excess of 2.6 hours per day increased the odds for a higher BMI in non-Hispanic whites, non-Hispanic blacks, and Hispanics; however, findings were only statistically significant in non-Hispanic whites. Additionally, the likelihood for being overweight/obese with increased TV viewing was similar for both college graduates and non-graduates and for those with health insurance and without. In comparison, differences were found when stratifying by employment status: the employed were at increased risk for a higher BMI, whereas unemployed were not.

Most of the literature to date has focused on eliciting social-demographic correlates of either TV viewing or obesity as outcome measures, rather than variations in the TV viewing-obesity relation [17][20]. However, a study by Richmond et al. (2010) specifically examined this association stratified by race/ethnicity (but not socio-economic status) in a sample of young adult women [21]. They found that TV viewing of >14 hours per week increased the risk of a higher BMI in white women, but not in non-Hispanic black and Hispanic women. Richmond et al. hypothesized that in ethnic minority groups, TV viewing might not be necessarily indicative of sitting time, i.e., the TV might be on in the background without individuals actually sitting and watching programs [21], [22]. Another explanation provided was that racial/ethnic minorities, particularly low income, are exposed to many other factors (beyond TV viewing) that affect overweight/obesity, e.g., obesogenic environment with little opportunity for physical activity and insufficient access or means to consume a healthful diet [21], [23].

In comparison to the study by Richmond, the current study findings indicate that non-Hispanic blacks and whites as well as Hispanics are 1.3 to 2.1 times more likely to be in the overweight/obese category if viewing TV more than 3.7 hours per day, but the relationship was only statistically significant in non-Hispanic whites. The lack of statistical significance among the racial/ethnic minority sub-populations is most probably due to smaller sample sizes (i.e., non-Hispanic blacks- n = 434, Hispanics- n = 473, and non-Hispanic whites- n = 3,960). These smaller sample sizes most likely affected power and the ability to detect statistically significant associations, particularly since the strength of the associations were similar between subgroups. Consistently, the odds ratios were similar in those with and without health insurance; however the lower sample size in the uninsured (no health insurance- n = 599; health insurance- n = 4,488) might have led to the inability to detect a statistically significant relation in this group. These suppositions, however, need to be substantiated in further studies where the sample sizes are larger in the various strata. In contrast, our findings pertaining to employment status are more ‘clear cut’: more daily TV viewing was not linked to increased risk for a higher BMI in the unemployed. While this finding warrants additional exploration in future studies, potential explanations could range from the TV being on in the background without actually sitting (i.e., multi-tasking) to significant heterogeneity in lifestyles among the unemployed.

The current study has several limitations that should be taken into account when interpreting the findings. First, the study design is cross-sectional, therefore a temporal relationship between TV viewing (independent variable) and BMI (dependent variable) cannot be determined. Second, the proportion of non-Hispanic blacks and Hispanics is less in the analytic sample than the weighted sample (in spite of oversampling of these groups), therefore it is likely that response rates were lower among these subpopulations. Additionally, the distribution of the analytic sample by gender and age differed from the weighted ones. Third, a large number of participants did not report their income, which is an important socio-economical variable; therefore income was not included in the analysis. To compensate for this we adjusted for income in multivariate analysis (in a subsample) finding consistent results with the presented findings. Fourth, both TV viewing and BMI are self-reported, which might result in differential misclassification of these variables among the various sub-populations. Fifth, even if TV viewing was monitored directly, this measure is a proxy of sedentary behavior, which was not measured objectively. Sixth, though we controlled for physical activity in multivariate analysis, this measure only includes moderate intensity physical activity. Time spent in light and vigorous intensity physical activity was not queried. However, these intensity categories contribute significantly to overall physical activity volume which is directly related to obesity status [1], [24]. Finally, TV viewing has been linked to increased energy intake as a result of food consumption during TV viewing and potentially due to exposure to advertisement of energy density food products [25]. Unfortunately, the HINTS survey does not include sufficient nutritional information to adjust for energy intake in the current analysis.

Nonetheless, scant evidence exists specifically examining the effects of TV viewing on BMI stratified by race/ethnicity and socio-economic variables. The present study contributes to the literature by examining these associations among a nationally representative large sample of US adults. Study findings indicate that TV viewing of >2.6 hours per days increases the odds for a higher BMI. This finding was mostly consistent among all racial/ethnic and socio-economic strata (with the exception of employment status), yet did not always achieve statistical significance. Future research should continue to explore potential variations in the association between TV viewing and obesity by race/ethnicity and socio-economic variables among larger samples of the various subpopulations to confirm or refute current findings. This further exploration is of importance to inform program planners when designing intervention studies aimed at decreasing TV viewing as a means to reduce obesity among these sub-populations.

Author Contributions

Conceived and designed the experiments: KS KPG TL. Performed the experiments: KS KPG TL. Analyzed the data: TL. Wrote the paper: KS KPG TL.

References

  1. 1. Church TS, Thomas DM, Tudor-Locke C, Katzmarzyk PT, Earnest CP, et al. (2011) Rodarte RQ, Martin CK, Blair SN, Bouchard C. Trends over 5 decades in U.S. occupation-related physical activity and their associations with obesity. PLoS One 6: e19657. doi: 10.1371/journal.pone.0019657
  2. 2. US Department of Health and Human Services. 2008 Physical Activity Guidelines for Americans. Available: http://www.health.gov/paguidelines/defau​lt.aspx. Accessed: 2012 Nov 2.
  3. 3. van der Ploeg HP, Chey T, Korda RJ, Banks E, Bauman A (2012) Sitting time and all-cause mortality risk in 222 497 Australian adults. Arch Intern Med 172: 494–500. doi: 10.1001/archinternmed.2011.2174
  4. 4. Dunstan DW, Salmon J, Owen N, Armstrong T, Zimmet PZ, et al. (2004) Physical activity and television viewing in relation to risk of undiagnosed abnormal glucose metabolism in adults. Diabetes Care 27: 2603–2609. doi: 10.2337/diacare.27.11.2603
  5. 5. Patel AV, Rodriguez C, Pavluck AL, Thun MJ, Calle EE (2006) Recreational physical activity and sedentary behavior in relation to ovarian cancer risk in a large cohort of US women. Am J Epidemiol 163: 709–716. doi: 10.1093/aje/kwj098
  6. 6. Wijndaele K, Healy GN, Dunstan DW, Barnett AG, Salmon J, et al. (2010) Increased cardiometabolic risk is associated with increased TV viewing time. Med Sci Sports Exerc 42: 1511–1518. doi: 10.1249/mss.0b013e3181d322ac
  7. 7. Hu FB, Li TY, Colditz GA, Willett WC, Manson JE (2003) Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA 289: 1785–1791. doi: 10.1001/jama.289.14.1785
  8. 8. Faridi Z, Shuval K, Njike VY, Katz JA, Jennings G, et al. (2010) Partners reducing effects of diabetes (PREDICT): a diabetes prevention physical activity and dietary intervention through African-American churches. Health Educ Res 25: 306–315. doi: 10.1093/her/cyp005
  9. 9. Shuval K, DeVahl J, Tong L, Gimpel N, Lee JJ, et al. (2011) Anthropometric measures, presence of metabolic syndrome, and adherence to physical activity guidelines among African American church members, Dallas, Texas, 2008. Prev Chronic Dis 8: A18.
  10. 10. National Cancer Institute. HINTS: Health Information National Trends Survey. http://hints.cancer.gov/. Accessed 2012 July 2.
  11. 11. Nelson DE, Kreps GL, Hesse BW, Croyle RT, Willis G, et al. (2004) The Health Information National Trends Survey (HINTS): development, design, and dissemination. J Health Commun 9: 443–460. doi: 10.1080/10810730490504233
  12. 12. National Cancer Institute. (2005) HINTS: Health Information National Trends Survey (HINTS 2005). Final Report. http://hints.cancer.gov/docs/HINTS_2005_​Final_Report.pdf. Accessed 2013 Jan 16.
  13. 13. Sullivan HW, Rutten LJ, Hesse BW, Moser RP, et al. (2010) Lay representations of cancer prevention and early detection: associations with prevention behaviors. Prev Chronic Dis 7: A14.
  14. 14. Lathan CS, Okechukwu C, Drake BF, Bennett GG (2010) Racial differences in the perception of lung cancer: the 2005 Health Information National Trends Survey. Cancer 116: 1981–1986. doi: 10.1002/cncr.24923
  15. 15. Nelson DE, Kreps GL, Hesse BW, Croyle RT, et al. (2004) The Health Information National Trends Survey (HINTS): development, design, and dissemination. J Health Commun 9: 443–460. doi: 10.1080/10810730490504233
  16. 16. Williams R (2006) Generalized ordered logit/partial proportional odds models for ordinal dependent variables. The Stata Journal 6: 58–82.
  17. 17. Flegal KM, Carroll MD, Ogden CL, Curtin LR (2010) Prevalence and trends in obesity among US adults, 1999–2008. JAMA 303: 235–241. doi: 10.1001/jama.2009.2014
  18. 18. Clark BK, Sugiyama T, Healy GN, Salmon J, Dunstan DW, et al. (2010) Socio-demographic correlates of prolonged television viewing time in Australian men and women: the AusDiab study. J Phys Act Health 7: 595–601.
  19. 19. King AC, Goldberg JH, Salmon J, Owen N, Dunstan D, et al. (2010) Identifying subgroups of U.S. adults at risk for prolonged television viewing to inform program development. Am J Prev Med 38: 17–26. doi: 10.1016/j.amepre.2009.08.032
  20. 20. Stamatakis E, Hillsdon M, Mishra G, Hamer M, Marmot M (2009) Television viewing time and other screened-based entertainment in relation to multiple socioeconomic status indicators and area deprivation: the Scotish Health Survey 2003. J Epidemiol Community Health 63: 734–740. doi: 10.1136/jech.2008.085902
  21. 21. Richmond TK, Walls CE, Gooding HC, Field AE (2010) Television viewing is not predictive of BMI in Black and Hispanic young adult females. Obesity 18: 1015–1020. doi: 10.1038/oby.2009.391
  22. 22. Henderson VR (2007) Longitudinal associations between television viewing and body mass index among black and white girls. J Adolesc Health 41: 544–550. doi: 10.1016/j.jadohealth.2007.04.018
  23. 23. Tucker LA, Tucker JM (2011) Television viewing and obesity in 300 women: evaluation of the pathways of energy intake and physical activity. Obesity 19: 1950–1956. doi: 10.1038/oby.2011.184
  24. 24. Britton KA, Lee IM, Wang L, Gaziano JM, Manson JE, et al. (2012) Physical Activity and the Risk of Becoming Overweight or Obese in Middle-Aged and Older Women. Obesity 20: 1096–1103. doi: 10.1038/oby.2011.359
  25. 25. Swinburn B, Shelly A (2008) Effects of TV time and other sedentary pursuits. Int J Obes (Lond) 32: S132–136. doi: 10.1038/ijo.2008.249