136x Filetype PDF File size 0.32 MB Source: media.ruddcenter.uconn.edu
RESEARCH Original Research Trying to Lose Weight: Diet Strategies among Americans with Overweight or Obesity in 1996 and 2003 TATIANA ANDREYEVA, PhD; MICHAEL W. LONG, MPH; KATHRYN E. HENDERSON, PhD; GABRIELLE M. GRODE, MPH ABSTRACT any Americans struggle to lose weight and are Background Health professionals recommend that individ- Mlooking for guidance on appropriate weight loss uals with overweight and obesity lose weight by reducing strategies. Based on clear findings (1) that the energy intake while maintaining a healthful diet. This macronutrientcompositionofdietsisunrelatedtoweight study was designed to examine trends in weight loss loss, clinicians, as well as public health campaigns, have attempts and strategies for adults with overweight or focused recent weight loss education efforts on energy obesity among different sociodemographic groups. restriction while encouraging a healthful overall diet (2,3). Methods Data from the 1996 and 2003 Behavioral Risk Despite the increased focus on energy restriction in Factor Surveillance System were used to estimate expert weight loss recommendations, public interest in changes in weight loss attempts and strategies across diets limiting intake of particular macronutrients with- population groups. Data were analyzed in 2009. out emphasizing energy restriction persists (4). While Results Slightly more adults with overweight or obesity many different diet strategies have been popularized in attempted weight loss in 2003 compared to 1996. There the United States (3), this research focuses on low-fat weresubstantialchangesinthedietapproachesreported: diets. Low-fat diets were widely promoted in the 1990s rates of those using energy restriction to lose weight (2,4), receiving significant expert endorsement (eg, Amer- doubled between 1996 and 2003, whereas low-fat dieting ican Heart Association), drawing many adherents, and decreased by one third. Hispanic and less-educated prompting food manufacturers to flood the market with adults did not shift away from low-fat diets. Attempted low-fat products (5). Determining the extent and speed weight loss was associated with higher fruit and vegeta- withwhichdifferentpopulationgroupshaveshiftedaway ble consumption for most population groups. fromlow-fatdietstostrategiesfocusingonenergyrestric- Conclusions Increasingly more adults with overweight or tion could inform clinical and public health responses to obesity tried to lose weight through energy reduction, but ongoing misconceptions about weight loss (6-9). some at-risk groups did not follow this beneficial trend In addition to encouraging weight loss through energy between 1996 and 2003. Dietetics practitioners and pub- restriction, the National Heart, Lung, and Blood Insti- lic health campaigns should target such groups with con- tute’s clinical guidelines for the treatment of overweight crete recommendations to reduce energy intake while andobesity emphasize the importance of implementing a maintaining a healthful diet, including adequate con- weight maintenance program to prevent commonly expe- sumption of fruit and vegetables. rienced weight regain (10). As a part of this program, the J Am Diet Assoc. 2010;110:535-542. NationalHeart,Lung,andBloodInstituteguidelinesrec- ommendconsumptionofdietsrichinfruitandvegetables as good sources of vitamins, minerals, and fiber, which may aid in weight maintenance (10). In observational T. Andreyeva is director of economic initiatives, K. E. studies, higher fruit and vegetable consumption has been Henderson is director of school and community initia- linked to long-term weight maintenance (11,12). Limited tives, and G. M. Grode is a research associate, Rudd experimental evidence suggests that increasing fruit and Center for Food Policy and Obesity, Yale University, vegetable consumption may aid in weight loss when com- NewHaven, CT. M. W. Long is a doctoral student, De- bined with energy restriction or reduction in dietary fat partment of Society, Human Development, and Health, (13-15). Diets rich in fruit and vegetables also reduce Harvard University School of Public Health, Cam- chronic disease risk, which is particularly elevated in bridge, MA. overweight and obese populations (10,16). Given the low Address correspondence to: Tatiana Andreyeva, PhD, rates of adequate fruit and vegetable consumption in the Rudd Center for Food Policy & Obesity, Yale University, United States (17) and clinical and public health recom- 309 Edwards St, New Haven, CT 06520-8369. E-mail: mendationstoconsumelow-energydietsrichinfruitand tatiana.andreyeva@yale.edu vegetables to lose or maintain weight (2,18), it is impor- Manuscript accepted: November 2, 2009. tant to determine whether those trying to lose weight Copyright © 2010 by the American Dietetic consume adequate amounts of fruit and vegetables. Association. This study used data from the 1996 and 2003 Behav- 0002-8223/10/11004-0004$36.00/0 ioral Risk Factor Surveillance System (BRFSS) to exam- doi: 10.1016/j.jada.2009.12.029 ine nationwide trends in reported weight loss diet strat- ©2010bytheAmericanDietetic Association Journal of the AMERICAN DIETETIC ASSOCIATION 535 egies for Americans with overweight or obesity (19,20). gaining weight)?” and “Are you using physical activity or Specifically, the analysis assessed changes in the preva- exercise to lose weight (or keep from gaining weight)?” lence of reported weight loss attempts and key diet strat- Our analysis focused on self-reported attempts to lose egies among adults with overweight or obesity between weight (with a yes/no indicator for trying to lose weight) 1996and2003,identifiedsociodemographicdifferencesin and several key diet strategies evaluated in the BRFSS: these trends, and estimated the association between try- trying to eat fewer calories vs trying to eat less fat vs ing to lose weight and reported consumption of diets rich trying to eat both fewer calories and less fat. Note that in fruit and vegetables. these diet strategies were assessed among adults report- ing weight loss attempts. The high percentage of partic- METHODS ipants with overweight or obesity reporting using diet methods to lose weight (about 90%) and the lack of evi- Sample dence supporting exercise as a weight loss strategy (21) The study source is the BRFSS, which provides state- prompted us to concentrate on diet practices in this specific estimates of health conditions, preventive health analysis. practices, and risk behaviors among adults in the United Fruit and Vegetable Consumption. The BRFSS module on States. The BRFSS is a random-digit-dial telephone sur- fruit and vegetables poses the same six questions in all vey conducted by state health departments on indepen- survey years the module was administered. Probed by dent probability samples of state residents aged 18 years interviewerstoconsiderthefoodtheyusuallyeatordrink or older. It is the world’s largest ongoing telephone health at home and away from home, participants report the systemsurvey;ithasgrowntoincludemorethan350,000 frequencyofconsumingfruit(excludingfruitjuice),green adults annually. The survey data are weighted by race/ salad, potatoes (excluding french fries, fried potatoes, and ethnicity, age, and sex to represent the probability of potato chips), carrots, other vegetables, and fruit juice. selection and each state’s population. A detailed descrip- Interviewers give examples of frequency responses tion of the BRFSS data and methodology is available (“twice a week, three times a month,andsoforth”),butno elsewhere (19,20). definition of a serving size. Participants can report the The analysis draws data from the earliest (1996) and frequency of eating or drinking per day, week, month, latest years (2003) of the BRFSS in which fruit and veg- year, or never. These responses are used to create the etable and weight control modules were administered in daily number of servings of individual food items, which all states. There were 124,085 participants in 1996 and are summedtoconstructthemeasuresoftotaldailyfruit 264,684respondentsin2003,amuchlargersampleinthe consumption,totalvegetableconsumption,andtotalfruit later wave because the BRFSS increased participation and vegetable consumption (sum of the two measures). over time. Exclusion criteria included participants under The assessment of fruit and vegetable consumption age 18 years; missing age or older than age 85 years; was based on a summary fruit and vegetable servings pregnant women; and participants missing information index available in the survey. It categorizes the daily for questions about weight or height, reported fruit and intake of fruit and vegetable servings on a 4-point scale, vegetable consumption, reports of trying to lose weight including less than once per day or never, once but less and diet-related questions from the weight control mod- than three times per day, at least three but less than five ule, and sociodemographic characteristics (education, times per day, and five or more times per day. A variable sex, or race/ethnicity). Participants reporting attempts to was created for whether individuals reported eating the lose weight but not reporting using either diet or physical Centers for Disease Control and Prevention recom- activity weight loss strategy were also excluded. Another mended intake—at the time of the survey—of a mini- important restriction was limiting this sample to adults mumoffivedaily servings of fruits and vegetables (18). with overweight or obesity (body mass index [BMI] of 25 Weight status was assessed based on the BRFSS self- andabove). Normal weight and underweight people were reported body measurements with participants assigned excluded to address weight loss attempts and strategies into three weight groups: overweight (BMI25 to 29.99), amongthoseatelevatedriskforexcessiveweight-related obese Class I (BMI30 to 34.99), and obesity Class II/III diseases. After all exclusions, 57,464 individuals in 1996 (BMI35).Therewerenosignificantdifferencesineither and 136,756 participants in 2003 remained eligible for the prevalence of weight loss attempt, diet practices, or analysis. fruit and vegetable intake between adults with BMI of 35 This study was deemed exempt from institutional re- to 39.99 and BMI 40, so these groups were combined. view board approval under federal regulation 45 CFR §46.101(b). Statistical Analysis Measures Comparative analyses of sociodemographic differences in Weight Loss and Diet Strategies. The weight control module weight loss practices across population groups and their starts with a question to all participants: “Are you now changes over time focused on rates of trying to lose trying to lose weight?” If the response is negative, the weight, diet strategies used, and reported fruit and veg- module continues with, “Are you now trying to maintain etable consumption in 1996 and 2003. Differences in your weight, that is to keep from gaining weight?” Only means between groups were tested using t tests, and participants reporting affirmatively to either of the two associations between categorical variables were assessed using 2 questions are asked the questions “Are you trying to eat tests. Stratifying by sex, weight status, race/ fewer calories or less fat to lose weight (or keep from ethnicity, and education, the hypothesis that weight loss 536 April 2010 Volume 110 Number 4 Table 1. Trends in prevalence of trying to lose weight among adults with overweight or obesity (n194,220) between 1996 and 2003 1996 2003 Men Women Men Women Characteristic (n29,554) (n27,910) (n65,679) (n71,077) 4™™™™™™™™™™™™™™™™™%a™™™™™™™™™™™™™™™™™3 b c bc Overall 41.7 62.9 44.1 64.9 Body mass index 25-24.99 35.1 59.2 34.6 59.3 30-34.99 58.4b b b b 68.4 59.5 70.2 b b b bc 35 and above 65.5 70.9 69.9 75.2 Race/ethnicity c c White 42.2 63.3 44.7 65.8 b b bc b African American 35.4 58.5 40.5 60.7 Hispanic 41.5 65.5 43.1 65.1 Other 48.3 67.3 44.4 64.7 Education b b b Less than high school 34.2 51.2 37.3 54.0 High school 38.1 61.8 39.9 62.4 b b b b Some college/technical school 44.2 67.9 45.7 68.7 b b b b College graduate and above 46.9 69.7 48.9 70.2 Professional advice to lose weight c Not received 36.3 56.5 37.2 58.6 b b bc b Advice received 73.8 82.4 77.5 82.8 a The reported estimates are percentages weighted to be nationally representative. bRates of trying to lose weight different within groups (body mass index: compared to overweight; race/ethnicity: compared to white non-Hispanic adults; education: compared to people with high school diploma; professional advice: compared to not receiving advice to lose weight) significant at P0.01. cRates of trying to lose weight different between 1996 and 2003 significant at P0.01. practices in the group remainedunchangedbetween1996 ducted using Stata (version 9.0, 2008, Stata Corporation, and 2003 was tested using a t test. College Station, TX). To determine whether trying to lose weight was asso- ciated with fruit and vegetable consumption and if this relationship differed across demographic groups, a mul- RESULTS tivariate linear probability regression model for the di- Trying to Lose Weight chotomous measure of meeting the five per day fruit and The prevalence of adults with overweight or obesity at- vegetable servings* was estimated. The model was run tempting weight loss varied significantly across demo- separately for men and women from the 1996 survey, the graphic groups (Table 1). Women with overweight or obe- 2003 wave and the combined 1996 and 2003 sample. It sity were 1.5 times more likely to attempt weight loss included measures of trying to lose weight, weight group, than men, and they started doing so at much lower BMI age, race/ethnicity, education, household income, receiv- levels. The sex difference in trying to lose weight dimin- ing professional advice to lose weight, and a year dummy ished with increasing BMI, but remained statistically in the combined 1996 and 2003 sample. Interactions be- significant. Although the differences in attempted weight tween trying to lose weight and race/ethnicity, weight loss by sex and degree of overweight were most notable, group, education and income were considered, and those there were also significant gradients by education and with significant effects such as weight group and race/ race.Theleasteducatedparticipants(lessthanhighschool) ethnicity were retained. To account for the complex sam- were the least likely to try to lose weight even though they pling design and obtain nationally representative es- timates, individual sample weights were used in all had significantly higher rates of obesity, especially severe analyses. The Huber-White estimator of variance was forms(eg,amongtheleasteducated,7.4%and13.7%ofmen used to calculate standard errors. All analyses were con- and women, respectively, had BMI of 35 in 2003; for men and women with a graduate degree, these rates were 4.6% and 5.4%, respectively). African Americans were less likely *While one typically does not include truly dichoto- to try to lose weight than white non-Hispanic and Hispanic mous outcomes in linear regression models, the underly- individuals. It is notable that Hispanic adults with over- ing distribution of fruit and vegetable consumption can weightorobesityhadratesofattemptedweightlosssimilar be considered continuous with a true zero point and in- to white non-Hispanic participants. Receiving professional finite possible positive true values. advice to lose weight was associated with considerably April 2010 ● Journal of the AMERICAN DIETETIC ASSOCIATION 537 Table 2. Trends in prevalence of diet methods among adults with overweight or obesity who are dieting to lose weight, 1996 and 2003 (n92,596) 1996 2003 Fewer Less Fewer calories Fewer Less Fewer calories Characteristic calories fat and less fat calories fat and less fat a 4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™% ™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 b b Overall 11.3 41.6 47.1 24.9 29.1 46.0 Sex b b Male 13.2 40.3 46.5 26.2 28.9 44.8 c c bc b c Female 9.6 42.8 47.6 23.7 29.2 47.1 Body mass index b b 25-24.99 10.7 42.7 46.6 25.3 29.3 45.3 b b 30-34.99 12.3 40.7 46.9 24.2 29.0 46.7 c b b 35 and above 11.7 38.5 49.7 24.6 28.6 46.8 Race/ethnicity b b White 9.9 40.9 49.1 26.0 25.1 48.8 c c bc bc c African American 16.0 39.8 44.2 22.8 34.5 42.7 c c c bc c c Hispanic 13.6 46.9 39.5 21.8 43.8 34.3 c c c bc Other 17.4 43.6 38.9 21.8 31.1 47.1 Education b c c Less than high school 13.5 43.8 42.7 21.1 39.9 38.9 b b High school 12.2 41.6 46.2 23.9 32.5 43.6 b bc c Some college/technical school 10.7 43.3 45.9 24.6 28.8 46.5 c c bc bc c College graduate 9.6 38.6 51.8 27.3 22.1 50.5 Professional advice to lose weight b b Not received 11.7 42.9 45.4 25.8 29.6 44.6 c c c b c Advice received 10.3 38.6 51.1 22.9 28.1 48.9 a The reported estimates are percentages weighted to be nationally representative. bDiet methods different between 1996 and 2003 significant at P0.01. cDiet methods different within groups (sex: compared to males; body mass index: compared to overweight; race/ethnicity: compared to whites; education: compared to people with high school diploma; professional advice: compared to not receiving advice to lose weight) significant at P0.01. higher rates of attempted weight loss, for example, almost pants. At the same time, the low-fat diet without energy doubling rates for men. restriction dropped in prevalence by one third, with Agreater proportion of adults with overweight or obe- greater reductions among college graduates and white sity tried to lose weight in 2003 than in 1996, although non-Hispanic participants. In contrast, Hispanic and theincreasewasrelativelymodest(53%vs51%,P0.01). least-educated individuals continued to use low-fat diet- This increase was twice as high for men as for women, ing without energy restriction at virtually the same rate which reduced the sex difference in attempted weight in 2003 as in 1996. loss. The largest gains in rates of attempted weight loss between 1996 and 2003 emerged for African-American menwithoverweightorobesity, white non-Hispanic par- Fruit and Vegetable Consumption ticipants, and women with BMI 35. Although some Ashypothesized,adultswithoverweightorobesitytrying sociodemographic differences in attempted weight loss to lose weight were more likely to report consuming the narrowed during this time, significant gradients re- recommended servings of fruit and vegetables than their mained, especially by education (Table 1). peers not attempting weight loss, by about five percent- Weight Loss Strategies age points (Table 3). Even with this gain, only a quarter of them met the recommendation for fruit and vegetable Among adults with overweight or obesity trying to lose consumption, which highlights inadequate fruit/vegeta- weight, about 86% of men and 92% of women reported ble intake in this population. Attempting to lose weight either eating less energy (energy restriction), or eating was associated with higher rates of meeting the recom- less fat (low-fat dieting), or following both strategies. mended intake of fruit and vegetables across most popu- Between 1996 and 2003, there were substantial changes lation groups, with the largest difference among those in the distribution of these diet approaches among adults with BMI of 35 and above. However, such differences in withoverweightorobesitytryingtoloseweight(Table2). fruit and vegetable intake were small for minorities and The proportion of people who reported using energy re- men. Despite all the promotional messages in recent striction alone doubled during this period with even years to increase fruit and vegetable consumption, rates larger increases for women and highly educated partici- of meeting the recommended intake of five servings of 538 April 2010 Volume 110 Number 4
no reviews yet
Please Login to review.