Trends in Weight Management Goals and Behaviors Among 9^th–12^th Grade Students — United States, 1999–2009

Abstract

Objective

To examine trends in weight management goals and behaviors among U.S. high school students during 1999–2009.

Methods

Data from 6 biennial cycles (1999–2009) of the national Youth Risk Behavior Survey were analyzed. Cross-sectional, nationally representative samples of 9^th–12^th grade students (approximately 14,000 students/cycle) completed self-administered questionnaires. Logistic regression models adjusted for grade, race/ethnicity, and obesity were used to test for trends in weight management goals and behaviors among subgroups of students.

Results

Combined prevalences and trends differed by sex and by race/ethnicity and weight status within sex. During 1999–2009, the prevalence of female students trying to gain weight decreased (7.6%–5.7%). Among female students trying to lose or stay the same weight, prevalences decreased for eating less (69.6%–63.2%); fasting (23.3%–17.6%); using diet pills/powders/liquids (13.7%–7.8%); and vomiting/laxatives (9.5%–6.6%) for weight control.

During 1999–2009, the prevalence of male students trying to lose weight increased (26.1%–30.5%). Among male students trying to lose or stay the same weight, the prevalence of exercising to control weight did not change during 1999–2003 and then increased (74.0%–79.1%) while the prevalence of taking diet pills/powders/liquids for weight control decreased (6.9%–5.1%) during 1999–2009.

Conclusion

Weight management goals and behaviors changed during 1999–2009 and differed by subgroup. To combat the use of unhealthy weight control behaviors, efforts may be needed to teach adolescents about recommended weight management strategies and avoiding the risks associated with unhealthy methods.

Introduction

The prevalence of obesity among adolescents has more than tripled since the 1970s [1] and in 2009–2010, nearly one-fifth (18.4%) of U.S. adolescents (12–19 years) were obese [2]. Obesity affects all racial/ethnic, sex, and socioeconomic subgroups [3]; however, racial/ethnic minorities and low income children and adolescents are more likely to be obese [2, 4]. Obesity in youth may have both short-term and long-term negative health effects, including high blood pressure, high cholesterol, type 2 diabetes, metabolic syndrome, sleep disturbances, orthopedic problems, and social and psychological problems [3, 5]. In addition, obese adolescents are more likely to become overweight or obese adults, resulting in increases in chronic diseases and premature death [3, 5]. In 2002–2005, obesity among youth (6–19 years) was associated with $2.9 billion in increased annual direct medical expenditures [6, 7]. As a result, weight management among adolescents is an important public health concern. The reduction of obesity prevalence among children and adolescents is a high priority issue for the United States and is a Healthy People 2020 objective [8]. Population-level changes in weight management can help achieve this national goal.

Weight management involves at least two dimensions: a weight goal and behaviors employed to reach the goal [9]. The weight goal for overweight and obese adolescents (12–18 years) is to achieve a body mass index (BMI) less than the 85^th percentile for age and sex [10]. The Expert Committee Recommendations Regarding the Prevention, Assessment, and Treatment of Child and Adolescent Overweight and Obesity recommend overweight adolescents (85^th percentile ≤ BMI < 95^th percentile) achieve a healthy weight by maintaining their current weight while stature increases; obese adolescents (BMI ≥95^th percentile) can pursue weight loss that is not to exceed an average of 2 pounds per week [10]. For obese and overweight adolescents, all weight management strategies should be supervised by a health care provider [10]. Recommended behaviors for weight management for overweight and obese youth include increasing intake of fruits and vegetables, participating in 1 or more hours of physical activity each day, decreasing television viewing to less than 2 hours per day, and reducing sugar-sweetened beverage intake [11]. In addition, adolescents should eat breakfast daily, eat meals prepared at home, and eat the majority of meals with their family. Family members should be encouraged to engage in healthy lifestyle changes along with their child [11].

In 2005–2006, a national study found that more than 80% of obese adolescent females and 60–80% of obese adolescent males, varying by race/ethnicity, wanted to weigh less and had tried to lose weight in the past year [12]. For both sexes, the percentage of adolescents who wanted to lose weight and who had tried to lose weight increased with body mass index [12].

While adolescents engage in both healthy and unhealthy weight management behaviors, unhealthy methods can have both negative psychosocial and physical health effects and may be counter-productive to weight loss [13–15]. Unhealthy weight management methods include skipping meals; purging; fasting; excessive exercise; and use of diet pills, laxatives, and supplements [10]. Community-based studies have found that, in the adolescent population, obese adolescents are more likely than normal weight adolescents to use healthy and unhealthy weight control behaviors [16–18].

One longitudinal community-based cohort study of adolescents in Minnesota has provided evidence that unhealthy weight control behaviors may persist over time, indicating that behaviors utilized in adolescence may track into adulthood [19]. The study found that approximately 45% of female students and 17% of male students persistently used (at both at baseline and follow-up) unhealthy weight control behaviors from 1998–1999 (when in junior or senior high school) to 2003–2004 (when in high school or post-high school). Only 10% of female students and 15% of male students persistently used healthy weight control behaviors over this same time period. Another study out of Minnesota observed decreasing secular trends in unhealthy weight control behaviors among middle and high school students [20].

No study to date has examined trends in these unhealthy behaviors over time among cross-sectional samples of 9^th–12^th grade school students nationally. Our study examined the trends of weight management goals and behaviors among students in grades 9–12 in the United States during 1999–2009 by sex, race/ethnicity, and weight status. This trend data may be useful for public health and school professionals conducting needs assessments and planning and implementing programs related to obesity and weight management, and also to clinicians and families who may also be addressing these issues among adolescents. The data may be used to identify changing at-risk populations, aid in targeting intervention efforts, and help assess the progress the nation is making in addressing this public health issue.

Methods

Study design and participants

Data were obtained from six biennial cycles (1999, 2001, 2003, 2005, 2007, 2009) of the national Youth Risk Behavior Survey (YRBS), a cross-sectional, school-based survey that has monitored six categories of priority health risk behaviors, obesity, and asthma since 1991. The sampling frame included all regular public and private schools that enrolled students in at least one of grades 9–12 in the 50 states and the District of Columbia. A three-stage cluster sample design was used to select nationally representative samples of students (ranging in size from 13,601 in 2001 to 16,410 in 2009). Overall response rate ranged from a low of 63% in 2001 to 71% in 2009. Additional detail about the study design is described elsewhere [21, 22]. Participation by schools and students was voluntary. Before survey administration, parental permission was obtained following local procedures. Questionnaires were completed anonymously and study protocols were designed to protect student’s privacy. Students completed the self-administered questionnaire during a regular class period and recorded their responses on a computer-scannable booklet or answer sheet. Psychometric properties of the questionnaire items have been published elsewhere [23]. The YRBS study protocol was reviewed and approved by the Institutional Review Board of the Centers for Disease Control and Prevention.

Measures

Students’ weight management goal was measured with the question “Which of the following are you trying to do about your weight?” The response options included “lose weight,” “gain weight,” “stay the same weight,” and “I am not trying to do anything about my weight.” Five weight control behaviors were assessed during 1999–2009. In separate questions, students were asked if during the past 30 days they had engaged in each of the following behaviors to lose weight or keep from gaining weight: 1) exercising; 2) eating less food, fewer calories, or foods low in fat; 3) going without eating for 24 hours or more (i.e., fasting); 4) taking diet pills, powders, or liquids without a doctor’s advice; and 5) vomiting or taking laxatives. Response options for each question were “yes” and “no.” Two of these behaviors are considered healthy when engaged in appropriately (exercising; eating less food, fewer calories, or foods low in fat) and three are considered unhealthy (fasting; taking diet pills, powders, or liquids without a doctor’s advice; and vomiting or taking laxatives). Other variables used in the analysis include sex (female, male), race/ethnicity (non-Hispanic black or African-American, Hispanic, non-Hispanic white, and other or multiple race), grade in school (9^th, 10^th, 11^th, or 12^th), and weight status (obese, non-obese). Weight status was determined by BMI, calculated from self-reported height and weight. Students were asked “How tall are you without your shoes on?” and “How much do you weight without your shoes on?” Students were considered obese if their BMI was ≥95^th percentile based on age- and sex-specific growth charts from the CDC [24]. All other students (overweight, normal weight, and underweight) were categorized as non-obese.

Statistical methods

The year 1999 was chosen as the first year for trend analysis as height and weight assessment was first conducted in 1999. The year 2009 was chosen as the end year as this was the last year that healthy weight control behaviors were included on the YRBS questionnaire. Trends in the prevalence of weight management goals were analyzed among all students, while weight management behavior trends were analyzed among those students trying to lose weight or stay the same weight. To analyze trends over time, time was modeled as a continuous variable with both linear and non-linear (quadratic) components. The time variables were created using orthogonal coefficients that reflected the biennial spacing of the surveys. Linear and quadratic time effects were assessed simultaneously in logistic regression models that tested for trends separately for male and female students in sex-combined models (using interaction terms) adjusted for race/ethnicity, grade, and weight status. Additional sex-stratified analyses were conducted to assess trends in weight management goals and behaviors among racial/ethnic and obesity subgroups. Each model was examined separately including all records with complete data on the variables included in that specific model.

Trends were reported by sex due to evidence of sex differences in weight management goals and behaviors among adolescents [12, 20, 25]. Trends for students from other or multiple-race subgroups are not presented because the numbers for any given year were too small for meaningful analysis. Models were adjusted for weight status because weight management goals and behaviors may change over time resulting from changes in weight status. For example, the goal of trying to lose weight may increase if there is an increase in obesity prevalence (which national studies show has occurred in recent years [2, 21]).

Data were weighted to calculate national estimates [22], and all analyses were conducted using SUDAAN, a statistical software package accounting for the complex sampling design. Biennial prevalence estimates were unadjusted. Adjusted odds ratios (ORs) for the main effects of sex, race/ethnicity, and weight status on weight management goals and behaviors were calculated for 2009 and were considered statistically significant if the 95% confidence intervals (CI) did not include 1. The linear and quadratic time effects were considered statistically significant when the regression coefficient (β) had a p-value of <0.05. A significant linear β in absence of a significant quadratic β indicated there was an overall linear increase or decrease in the variable during 1999–2009. A significant quadratic β in absence of a significant linear β indicated there was a non-linear change (i.e., an acceleration, leveling off, or change in direction) during 1999–2009. Significant linear and quadratic βs indicated a non-linear change in addition to an overall increase or decrease during 1999–2009 and are described by the quadratic trend. Non-significant linear and quadratic βs indicated no significant variation in the prevalence over time.

Results

Main effects, 2009

In 2009, female students had higher odds of trying to lose weight and lower odds of trying to gain weight, stay the same weight, or do nothing about their weight as compared to male students (Table 1). Among both sexes, black students had lower odds of trying to lose weight and higher odds of trying to gain weight as compared to white students. However, the association for black male students trying to lose weight did not reach statistical significance. Hispanic students of both sexes had lower odds of doing nothing about their weight. Among both female and male students, obese students had higher odds than non-obese students of trying to lose weight and lower odds of trying to gain weight, stay the same weight, or do nothing about their weight.

Table 1.

Trends in weight management goals^a among 9^th – 12^th grade students—United States, 1999–2009

	2009 Main effect	Prevalence by year (%)						Trend βb
Subgroup	OR(95% CI)	1999	2001	2003	2005	2007	2009	Linear	Quadratic
Lose weight
All females	4.51 (4.00, 5.07)	59.4	62.3	59.3	61.7	60.3	59.3	−0.009	−0.001
All males	referent	26.1	28.8	29.1	29.9	30.4	30.5	0.018c	−0.002
Black females	0.50 (0.40, 0.62)	48.3	49.4	46.7	52.7	49.5	47.3	−0.004	−0.001
Hispanic females	1.03 (0.90, 1.18)	63.6	63.4	61.7	64.1	62.1	62.4	−0.014	0.003
White females	referent	61.4	65.4	62.6	63.5	62.3	61.3	−0.009	−0.003
Black males	0.79 (0.62, 1.00)	23.6	23.6	22.7	24.4	24.9	26.3	0.013	0.006
Hispanic males	1.78 (1.53, 2.08)	37.3	39.1	37.4	38.6	38.5	41.8	0.015	0.007
White males	referent	24.9	27.9	27.9	28.8	29.0	28.4	0.016	−0.004
Obese females	5.84 (3.99, 8.54)	85.3	88.7	86.8	91.8	89.5	88.0	0.025	−0.012
Non-obese females	referent	59.3	60.5	58.6	58.7	58.0	57.6	−0.009	−0.000
Obese males	9.17 (7.36, 11.43)	67.6	67.7	67.8	74.4	74.7	73.2	0.037d	−0.001
Non-obese males	referent	19.1	22.6	22.0	21.5	22.3	23.0	0.013d	−0.002
Gain weight
All females	0.13 (0.11, 0.16)	7.6	5.9	10.3	6.3	6.0	5.7	−0.025d	−0.011
All males	referent	29.5	26.3	30.1	27.1	26.9	26.2	−0.013	−0.001
Black females	6.32 (4.64, 8.62)	18.6	15.2	17.7	13.0	14.7	14.1	−0.028	0.001
Hispanic females	2.37 (1.77, 3.17)	9.7	8.3	9.7	7.8	7.7	7.1	−0.029	−0.007
White females	referent	5.2	3.4	8.3	4.1	3.3	2.8	−0.041d	−0.022
Black males	1.67 (1.31, 2.12)	34.7	35.0	41.8	38.7	38.5	34.0	0.002	−0.011c
Hispanic males	0.91 (0.72, 1.15)	25.9	22.4	26.6	22.3	22.9	22.2	−0.018	−0.003
White males	referent	28.5	25.4	28.6	25.3	25.3	25.0	−0.016	0.001
Obese females	0.13 (0.05, 0.33)	2.3	1.4	4.1	0.7	0.6	0.9	−0.152	−0.021
Non-obese females	referent	6.6	5.8	9.0	6.7	5.8	5.5	−0.027d	−0.011
Obese males	0.13 (0.09, 0.18)	6.0	5.3	10.1	4.7	4.7	5.3	−0.035	−0.011
Non-obese males	referent	33.4	29.9	32.7	31.2	31.1	29.8	−0.011	−0.000
Stay same weight
All females	0.86 (0.77, 0.97)	18.0	16.0	17.2	17.7	18.4	18.5	0.012	0.002
All males	referent	20.1	21.5	19.4	21.1	21.8	19.8	0.003	−0.003
Black females	1.54 (1.27, 1.87)	19.9	20.0	22.2	21.2	19.8	23.7	0.020d	−0.000
Hispanic females	0.97 (0.80, 1.17)	14.2	13.6	18.3	15.9	18.1	16.8	0.039c	−0.008
White females	referent	18.8	15.6	15.5	17.4	18.3	18.3	0.008	0.005
Black males	1.11 (0.87, 1.42)	24.8	22.2	21.2	22.9	20.1	21.3	−0.015	0.003
Hispanic males	0.89 (0.75, 1.04)	18.9	20.8	17.5	22.9	23.2	17.9	0.007	−0.007
White males	referent	20.1	21.2	20.0	20.6	21.9	20.5	0.008	−0.002
Obese females	0.13 (0.07, 0.23)	5.9	2.7	2.3	1.9	2.9	3.2	−0.049	0.031c
Non-obese females	referent	19.0	16.9	18.7	19.3	20.6	19.9	0.014d	0.001
Obese males	0.40 (0.30, 0.54)	12.3	11.8	11.3	10.3	10.6	10.0	−0.025	0.001
Non-obese males	referent	21.4	22.7	21.7	23.3	23.8	21.6	0.005	−0.003
Do nothing
All females	0.61 (0.53, 0.70)	14.9	15.8	13.2	14.3	15.3	16.5	0.012	0.004
All males	referent	24.3	23.5	21.5	22.0	20.8	23.6	−0.005	0.005d
Black females	0.86 (0.64, 1.15)	13.2	15.4	13.4	13.1	15.9	14.9	0.010	0.000
Hispanic females	0.76 (0.61, 0.93)	12.6	14.6	10.3	12.1	12.1	13.8	−0.000	0.006
White females	referent	14.6	15.6	13.5	15.0	16.1	17.6	0.016	0.005
Black males	0.63 (0.50, 0.80)	16.9	19.2	14.4	14.0	16.6	18.5	0.002	0.007
Hispanic males	0.63 (0.52, 0.76)	18.0	17.7	18.5	16.2	15.4	18.0	−0.007	0.003
White males	referent	26.5	25.5	23.5	25.3	23.7	26.1	−0.005	0.005
Obese females	0.44 (0.29, 0.66)	6.5	7.2	6.8	5.6	7.0	7.9	0.019	0.006
Non-obese females	referent	15.1	16.8	13.7	15.3	15.5	17.0	0.011	0.004
Obese males	0.40 (0.30, 0.54)	14.1	15.1	10.8	10.6	9.9	11.6	−0.027	0.007
Non-obese males	referent	26.1	24.7	23.6	24.1	22.8	25.7	−0.003	0.005d

^aIn response to the question, “Which of the following are you trying to do about your weight?”

^bβ indicates logistic regression coefficient. Sex-stratified models are adjusted for race/ethnicity, grade, and weight status (body mass index ≥ 95^th percentile is considered obese). The initial sex comparison models include sex as a covariate.

^cp <0.01.

^dp <0.05.

Among those trying to lose or stay the same weight, female students had higher odds as compared to male students of engaging in all weight control behaviors except exercising to control weight (Table 2). Among females, as compared to white students, black and Hispanic students had lower odds of both healthy behaviors. In addition, black students had lower odds of taking diet pills, powders, or liquids without a doctor’s advice. Black male students had lower odds of healthy weight control behaviors as compared to white male students. Hispanic male students had higher odds exercising to control weight as compared to white male students. Among both sexes, obese students had higher odds of eating less, fewer calories, or foods low in fat. Exercising to control weight only differed by weight status among male students. Fasting to control weight and taking diet pills, powders, or liquids without a doctor’s advice only differed by weight status among female students.

Table 2.

Trends in weight management behaviors among 9^th-12^th grade students trying to lose or stay the same weight—United States, 1999–2009

	2009 Main effect	Prevalence by year (%)						Trend βa
Subgroup	OR(95% CI)	1999	2001	2003	2005	2007	2009	Linear	Quadratic
Exercised b
All females	1.06 (0.92, 1.23)	80.1	81.4	78.3	79.0	78.4	79.0	−0.011	0.003
All males	referent	76.0	75.8	74.0	76.1	76.7	79.1	0.017c	0.006c
Black females	0.42 (0.32, 0.56)	76.4	70.1	63.9	70.2	66.4	67.1	−0.033	0.010
Hispanic females	0.70 (0.56, 0.88)	78.9	78.8	75.1	79.4	77.4	77.1	−0.008	0.002
White females	referent	81.1	83.8	81.5	80.9	81.6	82.7	0.000	0.002
Black males	0.66 (0.51, 0.85)	66.7	67.2	67.8	71.4	74.2	72.3	0.024	0.003
Hispanic males	1.30 (1.00, 1.68)	75.8	74.2	75.0	83.1	76.3	83.8	0.048d	0.005
White males	referent	77.9	78.1	74.9	75.4	76.8	79.2	0.002	0.007
Obese females	1.09 (0.81, 1.47)	84.5	84.5	80.6	83.5	80.1	79.1	−0.032	−0.001
Non-obese females	referent	80.1	81.2	78.7	78.5	78.4	79.6	−0.007	0.004
Obese males	1.88 (1.42, 2.48)	87.3	83.7	81.3	86.3	86.8	85.9	0.015	0.010
Non-obese males	referent	72.9	74.5	73.0	72.5	74.3	77.4	0.016	0.005
Ate less, fewer calories, or foods low in fat b
All females	2.09 (1.88, 2.33)	69.6	71.7	67.9	66.2	65.0	63.2	−0.037e	−0.001
All males	referent	46.2	48.4	48.5	45.8	46.8	48.2	0.001	0.001
Black females	0.37 (0.29, 0.47)	59.8	54.7	52.7	50.7	48.0	45.8	−0.053e	0.003
Hispanic females	0.63 (0.54, 0.73)	63.0	69.9	63.6	64.0	62.3	58.4	−0.035c	−0.005
White females	referent	73.2	75.2	72.3	69.8	69.9	68.3	−0.035d	0.000
Black males	0.61 (0.45, 0.82)	41.5	41.9	35.1	37.5	35.0	39.1	−0.013	0.008
Hispanic males	0.87 (0.71, 1.07)	46.0	49.0	48.9	46.3	45.3	48.0	−0.007	0.001
White males	referent	49.7	48.9	51.3	46.6	49.2	50.0	−0.003	0.001
Obese females	2.07 (1.66, 2.59)	82.9	80.0	75.2	78.5	73.5	74.5	−0.049c	0.004
Non-obese females	referent	69.1	71.4	67.3	64.6	64.4	62.1	−0.035e	−0.001
Obese males	2.06 (1.73, 2.46)	62.7	63.6	65.6	58.8	66.7	61.1	−0.004	−0.001
Non-obese males	referent	40.8	44.4	43.2	40.8	40.9	43.9	0.002	0.001
Fasted b f
All females	1.99 (1.71, 2.32)	23.3	23.1	21.6	20.9	19.4	17.6	−0.036e	−0.003
All males	referent	11.2	12.2	11.8	11.8	10.9	10.4	−0.013	−0.003
Black females	0.92 (0.72, 1.19)	23.6	20.0	19.1	17.7	17.1	16.7	−0.042d	0.006
Hispanic females	1.02 (0.84, 1.24)	21.5	28.0	20.7	21.1	19.7	17.9	−0.037d	−0.002
White females	referent	22.9	23.4	21.5	21.5	19.9	17.8	−0.035e	−0.004
Black males	1.27 (0.91, 1.77)	14.5	16.1	12.0	12.5	11.0	11.9	−0.039	0.005
Hispanic males	1.31 (0.92, 1.87)	9.7	11.8	10.3	10.3	14.6	11.6	0.030	0.000
White males	referent	10.4	11.1	10.9	12.4	8.9	9.6	−0.018	−0.007
Obese females	1.64 (1.30, 2.06)	29.8	28.4	24.1	22.6	28.2	24.1	−0.019	0.007
Non-obese females	referent	22.6	22.8	21.3	20.5	18.2	16.5	−0.038e	−0.004
Obese males	1.10 (0.80, 1.52)	13.6	14.9	12.2	14.6	14.8	10.5	−0.012	−0.007
Non-obese males	referent	10.3	11.0	11.0	10.4	9.4	9.7	−0.013	−0.002
Took diet pills, powders, or liquids without a doctor’s advice b
All females	1.68 (1.29, 2.20)	13.7	15.6	13.0	9.6	8.7	7.8	−0.086e	−0.005
All males	referent	6.9	8.6	8.2	5.9	6.0	5.1	−0.051e	−0.007
Black females	0.48 (0.35, 0.68)	9.2	10.5	5.3	6.1	5.1	4.4	−0.094e	0.005
Hispanic females	1.00 (0.80, 1.25)	14.1	17.0	12.5	9.0	8.6	8.5	−0.090e	0.001
White females	referent	14.3	16.3	15.0	10.7	9.6	8.6	−0.081e	−0.008c
Black males	0.87 (0.48, 1.59)	7.5	7.0	6.0	5.4	4.9	5.3	−0.064	0.006
Hispanic males	0.96 (0.59, 1.57)	9.2	10.1	7.1	7.4	6.6	5.2	−0.062c	−0.000
White males	referent	7.0	8.1	8.8	5.5	5.3	5.4	−0.048c	−0.006
Obese females	1.85 (1.43, 2.40)	24.0	25.1	17.0	14.7	13.4	11.8	−0.098e	0.001
Non-obese females	referent	13.1	14.8	12.4	8.8	8.0	7.0	−0.084e	−0.006c
Obese males	1.20 (0.86, 1.67)	11.2	12.9	10.9	8.4	9.8	5.7	−0.070e	−0.009
Non-obese males	referent	5.4	6.9	6.8	4.8	4.2	4.7	−0.040c	−0.006
Vomited or took laxatives b
All females	2.38 (1.80, 3.13)	9.5	9.4	9.0	7.5	7.5	6.6	−0.044e	−0.003
All males	referent	3.2	4.2	3.5	3.7	2.9	3.1	−0.025	−0.004
Black females	0.69 (0.40, 1.19)	9.3	5.2	6.3	4.6	4.9	4.5	−0.050c	0.004
Hispanic females	1.31 (0.97, 1.78)	7.9	13.3	9.3	7.8	7.5	8.3	−0.035c	−0.003
White females	referent	8.6	9.8	9.4	8.1	8.2	6.3	−0.038d	−0.008
Black males	2.40 (1.38, 4.17)	4.8	5.1	5.2	3.5	3.1	5.4	−0.016	0.014
Hispanic males	2.09 (1.18, 3.70)	5.1	4.9	4.2	4.7	4.5	4.4	−0.019	0.007
White males	referent	2.6	3.4	3.1	3.0	2.1	2.2	−0.042	−0.008
Obese females	1.10 (0.69, 1.75)	10.9	11.7	8.5	7.9	9.3	6.8	−0.047	−0.000
Non-obese females	referent	9.2	9.1	8.8	7.4	7.1	6.4	−0.043e	−0.003
Obese males	1.17 (0.68, 2.02)	3.0	3.0	2.5	3.8	3.3	3.2	0.016	−0.000
Non-obese males	referent	3.2	4.3	3.4	3.4	2.4	2.7	−0.041	−0.006

^aβ indicates logistic regression coefficient. Sex-stratified models are adjusted for race/ethnicity, grade, and weight status (body mass index ≥ 95^th percentile is considered obese). The initial sex comparison models include sex as a covariate.

^bTo lose weight or keep from gaining weight, during the 30 days before the survey.

^cp<0.05.

^dp<0.01.

^ep<0.001.

^fGoing without eating for 24 hours or more.

Trends over time, weight management goals

Significant trends in weight management goals were observed (Table 1). The prevalence of trying to gain weight decreased during 1999–2009 among all female students (7.6%–5.7%), non-obese female students (6.6%–5.5%), and white female students (5.2%–2.8%). The prevalence of trying to stay the same weight increased during 1999–2009 for both black and Hispanic female students. Among non-obese female students, this prevalence increased during 1999–2009 (19.0%–19.9%), and among obese female students, the prevalence decreased during 1999–2005 (5.9%–1.9%) and then did not change during 2005–2009 (1.9%–3.2%).

During 1999–2009, among male students, increases in the prevalence of trying to lose weight were found among all students (26.1%–30.5%), obese students (67.6%–73.2%), and non-obese students (19.1%–23.0%). Among black male students, the prevalence of trying to gain weight increased during 1999–2003 (34.7%–41.8%) and then decreased during 2003–2009 (41.8%–34.0%). The percent of male students doing nothing about their weight did not change during 1999–2007 (24.3%−20.8%), and then increased during 2007–2009 (20.8%–23.6%). The same pattern was observed for non-obese male students.

Trends over time, healthy weight management behaviors

Significant trends in healthy weight management behaviors occurred during 1999–2009 (Table 2). Eating less food, fewer calories, or foods low in fat to control weight decreased during 1999–2009 among all groups of female students. Among male students overall, the prevalence of exercising to control weight did not change during 1999–2003 and then increased during 2003–2009 (74.0%–79.1%). Among Hispanic male students, this behavior increased during 1999–2009 (75.8%–83.8%).

Trends over time, unhealthy weight control behaviors

Significant trends in unhealthy weight management behaviors occurred during 1999–2009 (Table 2). Fasting to control weight decreased during 1999–2009 among all groups of female students except for obese female students. Taking diet pills, powders, or liquids without a doctor’s advice to control weight decreased during 1999–2009 among all groups of female students. However, among white female students and non-obese female students, the prevalence decreased only during 2003–2009 and 2001–2009, respectively. Vomiting and taking laxatives to control weight decreased during 1999–2009 among all groups of female students except obese students. During 1999–2009, using diet pills, powders, or liquids without a doctor’s advice to control weight decreased among all groups of male students except black students.

Discussion

During 1999–2009, prevalences and trends in weight management goals and behaviors differed between female and male students and between race/ethnicity and obesity subgroups among the sexes. One positive finding regarding weight management goals is that obese students of both sexes consistently had goals in line with weight control: higher odds of trying to lose weight, lower odds of trying to gain weight, lower odds of trying to stay the same weight, and lower odds of doing nothing about one’s weight as compared to non-obese students of the same sex. This is positive because it is recommended that obese students pursue weight loss in order to achieve a healthy weight. Race/ethnic differences in weight management goals may result not only from potential differences in obesity status, but from cultural differences in body image and social norms surrounding preferred body size [26–28]. These social influences may also influence weight control behaviors [29]. In line with previous literature [17, 18], obese students also had higher odds of using both healthy and unhealthy weight control behaviors, though the associations were not consistently significant. We highlight important trend findings for these behaviors below.

We found that students in grades 9–12 use both healthy and unhealthy behaviors to control weight. Physical activity promotes a healthy body weight among all youth and can reduce adiposity in obese youth [30, 31]. In 2009, more than half of both female students (79.0%) and male students (79.1%) trying to lose or stay the same weight used exercise as a method to manage their weight. However, during 1999–2009, the prevalence of exercising to control weight only changed among overall and Hispanic male students. Given that a small proportion of high school students, on a whole, meet physical activity recommendations [32], more research is needed to identify barriers to increasing exercise, specifically among obese students, and develop appropriate interventions to address these barriers.

Reducing calories, reducing food intake, and choosing a low-fat diet are healthy dietary changes that help with weight management [33]. However, the prevalence of eating less food, fewer calories, or foods low in fat to control weight decreased among all groups of female students and did not change among male students during 1999–2009. Students of racial/ethnic minority groups were less likely to eat less food, fewer calories, or foods low in fat as compared to white students. Literature shows that culture influences food choices and that food choices differ by race/ethnicity [34–36]. These cultural influences may play a role in an adolescent’s choice of using dietary methods for weight control. Previous research has also shown that the prevalence of weight control behaviors, including dietary methods, differs by race/ethnicity among adolescents [25]. It is important that these cultural differences be understood in order to develop culturally-relevant interventions and programs.

Unhealthy weight control behaviors, such as fasting; diet pills, powders, or liquids; and vomiting/taking laxatives to control their weight, should be avoided by all students due to their negative psychosocial and physical health effects [13–15]. In addition to the negative consequences of these unhealthy methods, they are counter-productive to weight control by increasing one’s risk of being obese through a variety of mechanisms [14, 37]. Fortunate news is that the data presented here indicate that these behaviors are decreasing, especially among female students. The reasons for these trends are speculative. The changes could be influenced by efforts from health providers and school or community interventions on healthy ways for students of this age to manage their weight and changing social norms on weight management.

There are at least four limitations to this study. First, data are only collected among adolescents who attend school and, therefore, are not representative of all persons in this age group. Nationwide, in 2008, of persons aged 16–17 years, approximately 4% were not enrolled in a high school program and had not completed high school [38]. Second, YRBS data are self-reported and students may under- or over-report these behaviors. However, YRBS dietary behavior questions, which includes the weight management behaviors, generally have moderate test-retest reliability [23]. Third, BMI based on self-reported height and weight data tends to underestimate the prevalence of obesity [39]. Lastly, this study was descriptive and was not designed to explain the reasons for the observed trends.

It is important to note a few other issues. First, the healthiness of weight management goals depends on individual factors such as age and obese status. Also, the weight control behaviors classified as healthy may not be healthy among all students. It may be unhealthy for students who are underweight to engage in these behaviors for the purpose of losing or maintaining weight. Also, if these behaviors are taken to extremes (e.g., skipping meals, reducing food intake severely, or exercising excessively), it is no longer a healthy behavior. Further, these weight control behaviors may not be used in isolation. Students combine the use of healthy and unhealthy weight control behaviors, which can also be unhealthy.

In conclusion, some improvements in weight management goals and behaviors were seen during 1999–2009. Differences in weight management goals and behaviors by sex, race/ethnicity, and weight status were observed. Healthy methods of weight control were more prevalent among both sexes than unhealthy methods. Obese students were more likely than non-obese students to engage in weight control behaviors. A significant reduction in all unhealthy weight control behaviors occurred during 1999–2009 among females. Though declines have occurred, still too many adolescents, especially females, are using unhealthy weight control behaviors. More research is needed to inform intervention strategies, including factors that influence adolescent weight control behaviors and the level of knowledge regarding the negative consequences of using unhealthy weight control behaviors. These strategies involve the development of messages that counter negative advice that students receive regarding weight management and unhealthy social norms and should be responsive to cultural factors related to weight management goals and behaviors. A variety of healthcare, community, school, and home-based strategies may be needed to ensure adolescents receive appropriate evidence-based information about weight management and understand the risks of unhealthy weight control behaviors. Unhealthy weight control behaviors may be indicators of current or future eating disorders [15]; therefore, both medical and psychiatric clinicians may inquire regarding adolescent weight control behaviors and use the opportunity to educate about their negative effects, encourage healthy behaviors such as healthy eating and regular physical activity, deemphasize body dissatisfaction, and address any weight stigmatization/discrimination [40, 41]. Clinicians and families can work together to develop a home environment supportive of healthy choices [40].