<%BANNER%>

Unhealthy Weight Control Behaviors among High School Students

Permanent Link: http://ufdc.ufl.edu/UFE0043677/00001

Material Information

Title: Unhealthy Weight Control Behaviors among High School Students
Physical Description: 1 online resource (98 p.)
Language: english
Creator: Weng, Chung-Bang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: adolescent -- behavior -- control -- health -- risk -- weight
Health Education and Behavior -- Dissertations, Academic -- UF
Genre: Health and Human Performance thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Weight control has become a major health concern nationwide. Adolescents often practice weight control behaviors using both healthy and unhealthy approaches. The present work utilized Youth Risk Behavior Survey (YRBS) data to estimate the trends of Healthy Weight Control Behaviors (HWCB) and Unhealthy Weight Control Behaviors (UWCB), identify the predictive factors of UWCB, and explore the relationship between UWCB and selected health risk behaviors. Results showed: 1) The prevalence rates of UWCB decreased and the prevalence rates of HWCB increased in a significant linear trend from 1999 to 2009. The subgroup analyses (by sex, grade, race/ethnicity, Body Image Distortion (BID) also demonstrated similar trends in both UWCB and HWCB. 2) UWCB is significantly associated with the following risk factors: currently trying to lose weight (vs. not trying); being American Indian/Alaska Native (vs. White) or multiple race/ethnicity (vs. White); nonoverweight but having body image distortion (BID) (vs. nonoverweight and no BID); eating the recommended amount of vegetables (vs. not eating the recommended amount); being female (vs. male); drinking sugar-sweetened beverage daily (vs. not drinking sugar-sweetened beverage daily); and being not vigorous in physical activity (vs. vigorous physical activity). 3) UWCB was significantly associated with substance use, suicidal behavior, and been bullied among female adolescents while male adolescents demonstrated significant associations in alcohol and tobacco use, multiple sexual partners, suicidal behaviors, physical and sexual violence, and been bullied. Findings suggested a moderate portion of high school students engaged in UWCB, and several demographic, psychological, lifestyle factors, and several health risk behaviors. Moreover, male adolescents with UWCB tended to engage in more health risk behaviors than females. Design of intervention programs accounting for the aforementioned risk factors is suggested to help reduce adolescent's risk of UWCB.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Chung-Bang Weng.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Chen, Wei W.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043677:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043677/00001

Material Information

Title: Unhealthy Weight Control Behaviors among High School Students
Physical Description: 1 online resource (98 p.)
Language: english
Creator: Weng, Chung-Bang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: adolescent -- behavior -- control -- health -- risk -- weight
Health Education and Behavior -- Dissertations, Academic -- UF
Genre: Health and Human Performance thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Weight control has become a major health concern nationwide. Adolescents often practice weight control behaviors using both healthy and unhealthy approaches. The present work utilized Youth Risk Behavior Survey (YRBS) data to estimate the trends of Healthy Weight Control Behaviors (HWCB) and Unhealthy Weight Control Behaviors (UWCB), identify the predictive factors of UWCB, and explore the relationship between UWCB and selected health risk behaviors. Results showed: 1) The prevalence rates of UWCB decreased and the prevalence rates of HWCB increased in a significant linear trend from 1999 to 2009. The subgroup analyses (by sex, grade, race/ethnicity, Body Image Distortion (BID) also demonstrated similar trends in both UWCB and HWCB. 2) UWCB is significantly associated with the following risk factors: currently trying to lose weight (vs. not trying); being American Indian/Alaska Native (vs. White) or multiple race/ethnicity (vs. White); nonoverweight but having body image distortion (BID) (vs. nonoverweight and no BID); eating the recommended amount of vegetables (vs. not eating the recommended amount); being female (vs. male); drinking sugar-sweetened beverage daily (vs. not drinking sugar-sweetened beverage daily); and being not vigorous in physical activity (vs. vigorous physical activity). 3) UWCB was significantly associated with substance use, suicidal behavior, and been bullied among female adolescents while male adolescents demonstrated significant associations in alcohol and tobacco use, multiple sexual partners, suicidal behaviors, physical and sexual violence, and been bullied. Findings suggested a moderate portion of high school students engaged in UWCB, and several demographic, psychological, lifestyle factors, and several health risk behaviors. Moreover, male adolescents with UWCB tended to engage in more health risk behaviors than females. Design of intervention programs accounting for the aforementioned risk factors is suggested to help reduce adolescent's risk of UWCB.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Chung-Bang Weng.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Chen, Wei W.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043677:00001


This item has the following downloads:


Full Text

PAGE 1

1 UNHEALTHY WEIGHT CONTROL BEHAVIORS AMONG HIGH SCHOOL STUDENTS By CHUNGBANG WENG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 ChungBang Weng

PAGE 3

3 This dissertation is dedicated, with deepest love and appreciation, t o m y parents without their support, this accomplishment would not have been possible

PAGE 4

4 ACKNOWLEDGMENTS I thank my committee members, Drs. William W. Chen, Jiunn Jye Sheu, R. Morgan Pigg, Jr., Delores C.S. James, and I Chan Huang, for their patience, encouragement, and unwavering support. I appreciate each person for his/her unique gifts and special talents. I am grateful for the attention shown to me, time devoted to me, and sage advice shared with me.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................. 7 ABSTRACT ..................................................................................................................... 8 CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW ...................................................... 10 General Introduction .................................................................................................. 10 Overweight and Obesity among Adolescents in the U nited S tates ................................... 13 Healthy Weight Control Behaviors (HWCB) ................................................................ 14 Unhealthy Weight Control Behaviors (UWCB) ............................................................ 16 Health Consequences of UWCB ................................................................................. 18 Predictive Factors of UWCB ...................................................................................... 20 Gender, Race/Ethnicity, and Overweight Status ..................................................... 20 Weight Perception, Body Image Distortion, and Weight Loss Intention .................... 22 Sedentary Behaviors Physical Activity and Dietary Behaviors ............................... 24 Associations between UWCB and Other Risk Behaviors ............................................... 25 Alcohol, Tobacco, or Drug Use ............................................................................ 26 Suicidal Ideation and Suic idal Attempt .................................................................. 27 Sexual Risk Behaviors ......................................................................................... 28 Description of the Youth Risk Behavior Survey (YRBS) ............................................... 29 Population and Sampling ............................................................................................ 30 Three stage cluster sampling design ...................................................................... 30 Weighting .......................................................................................................... 31 Instrumentation ......................................................................................................... 31 Validity and Reliability of YRBS ................................................................................ 31 Quality Control Procedures ........................................................................................ 33 Advantages and Disadvantages of Secondary Data Analysis .......................................... 34 Delimitations ............................................................................................................ 35 Statement of Purpose ................................................................................................. 36 Significance .............................................................................................................. 37 Research Aims .......................................................................................................... 38 2 TRENDS OF WEIGHT CONTROL BEHAVIORS AMONG HIGH SCHOOL STUDENTS ............................................................................................................. 40 B ackground .............................................................................................................. 40 M ethods ................................................................................................................... 42 R esults ..................................................................................................................... 45 D iscussion ................................................................................................................ 48

PAGE 6

6 Limitations ......................................................................................................... 49 Future R esearch and Implications ......................................................................... 49 3 RISK FACTORS OF UNHEALTHY WEIGHT CONTROL BEHAVIORS AMONG HIGH SCHOOL STUDENTS .................................................................................... 54 Background .............................................................................................................. 54 M ethods ................................................................................................................... 55 Sample Demographics ......................................................................................... 56 Measures ............................................................................................................ 56 Data Analysis ..................................................................................................... 60 R esults ..................................................................................................................... 61 D iscussion ................................................................................................................ 63 4 THE RELATIONSHIP BETWEEN UNHEALTHY WEIGHT CONTROL BEHAVIORS AND SELECTED HEALTH RISK BEHAVIORS AMONG HIGH SCHOOL STUDENTS .............................................................................................. 68 Background .............................................................................................................. 68 M ethods ................................................................................................................... 71 Study Design ...................................................................................................... 71 Measures ............................................................................................................ 71 Data Analysis ..................................................................................................... 72 R esults ..................................................................................................................... 75 D iscussion ................................................................................................................ 75 5 SUMMARY, CONCLUSIONS, RECOMMENDATIONS, AND IMPLICATIONS ......... 84 Summary .................................................................................................................. 84 Conclusions .............................................................................................................. 85 Recommendations for Future Research ........................................................................ 86 Implications .............................................................................................................. 87 LIST OF REFERENCES .................................................................................................. 89 BIOGRAPHICAL SKETCH ............................................................................................. 98

PAGE 7

7 LIST OF TABLES Table page 21 Descriptive statistics of participants demographics, BID, weight loss tendency, UWCB, and HWCB by year ................................................................................... 51 22 UWCB prevalence rates by year and participants demographics and BID and their bivariate odds ratios .............................................................................................. 52 23 HWCB prevalence rates by year and participants demographics and BID and their bivariate odds ratios .............................................................................................. 53 31 UWCB distribut ion by the candidate predictors ........................................................ 66 32 Results of stepwise multiple logistic regression ........................................................ 67 41 UWCB Distribution by s ex and s elected h ealth r isk b ehaviors ................................... 82 42 Relat ionships between UWCB and s el ected h ealth r isk behaviors by s ex .................... 83

PAGE 8

8 ABSTRACT OF DISSERTA TION PRESENTED TO TH E GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLME NT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNHEALTHY WEIGHT CONTROL BEHAVIORS AMONG HIGH SCHOOL STUDENTS By Chung Bang Weng December 2011 Chair: William W. Chen Major: Health and Human Performance W eight control has become a major health concern nationwide. Adolescents often practice weight control behaviors using both healthy and unhealthy approaches. The present work u tilized Youth Risk Behavior Survey (YRBS) data to estimate the trends of Healthy Weight Control Behaviors (HWCB) and Unhealthy Weight Control Behaviors (UWCB) identify the predictive factors of UWCB, and explore the relationship between UWCB and selected health risk behaviors. Results showed: 1) T he prevalence rates of UWCB decreased and the prevalence rates of HWCB increased in a significant linear trend from 1999 to 2009. The s ubgroup analys e s ( by sex, grade, race/ethnic ity, Body Image Distortion (BID ) also demonstrated similar trends in both UWCB and HWCB. 2) UWCB is significantly associated with the following risk factors: currently trying to lose weight (vs. not trying ); being American Indian/Alaska Native (vs. White ) or multiple race/ethnicity (vs. White ) ; nonover weight but having body image distortion (BID) (vs. nonoverweight and no BID ); eating the recommended amount of vegetables (vs. not eating the recommended amount); being female (vs. male); drinking sugar sweetened beverage daily (vs. not drinking s ugar sweetened beverage daily ); and being not vigorous in physical activ ity

PAGE 9

9 (vs. vigorous physical activity) 3) UWCB was significantly associated with substance use, suicidal behavior, and been bullied among female adolescents while male adolescents demon strated significant associations in alcohol and tobacco use, multiple sexual partners, suicidal behaviors, physical and sexual violence and been bullied Findings suggested a moderate portion of high school students engaged in UWCB and several demographic, psychological, lifestyle factors and several health risk behaviors Moreover, m ale adolescents with UWCB tend ed to engage in more health risk behaviors than females Design of intervention programs accounting for the aforementioned risk facto rs is suggested to help reduce adolescents risk of UWCB.

PAGE 10

10 CHAPTER 1 INTRODUCTION AND LIT ERATURE REVIEW General Introduction The prevalence of obesity has been more than doubled in the past 3 decades (Y. Wang & Beydoun, 2007) Recent data from 2007 2008 National Health and Nutrition Examination Survey (NHANES) indicated that 33.8% of American adults were obese and 68% were overweight or obese (Flegal, Carroll, Ogden, & Curtin, 2010) Among adolescents aged 1219 years old, 18% ha d age specific BMI greater than 95th percentile, which is considered obese, and 34% ha d age specific Body Mass Index (BMI) greater than 85th percentile, which is considered overweight (Ogden & Flegal, 2010; Ogden, Carroll, Curtin, Lamb & Flegal, 2010) Such trends, if persist, can cause enormous life lost and economic costs. To prevent weight gain, people tend to engage in different weight control practices. Although many adolescents use healthy strategies, unhealthy or extreme wei ght control strategies, such as fasting, using diet pills/powders/liquids, vomiting, or taking laxatives are also common among adolescents (Eaton et al., 2010) Unhealthy Weight Control Behaviors (UWCB), such as the a forementioned ones (Eisenberg, NeumarkSztainer, Story, & Perry, 2005; Haines, Neumark Sztainer, Perry, Hannan, & Levine, 2006; D. Neumark Sztainer, Wall, Story, & Pe rry, 2003) predict the onset of more severe disordered eating behaviors and eating disorders, including anorexia nervosa, bulimia nervosa, and binge eating disorder, as well as depression and other medical and psychosocial morbidity (Fonseca, Matos, Guerra, & Pedro, 2009) According to the 2009 Youth Risk Behavioral Surveillance System (YRBSS), 59% of female and 31% of male students in grades 912 reported trying to lose weight. By looking into each unhealthy beha viors, i n 2009, 11% of high school students did not eat for 24 or more hours, 5% had taken diet pills, powders, or liquids without a doctors advice, and 4% had vomited or

PAGE 11

11 taken laxatives to lose weight or to keep from gaining weight during the 30 days bef ore the survey (Eaton et al., 2010) Contrary to the projected weight loss, particularly during adolescence, dieting has been shown to predict weight gain over time (Fonseca et al., 2009) (D. Neumark Sztainer, Wall, Story, & Sherwood, 2009) In addition, adolescents who practiced UWCB have approximately 3 times greater risk for being overweight and 6 times higher risk for uncontrolla ble binge eating 5 years later (D. Neumark Sztainer et al., 2006) (D. Neumark Sztainer et al., 2006) Gender difference may also play a role in the weight control related problem s Many studies ha d identified the gender difference in body image, weight concern, and weight control behaviors. Girls engaged in more bulimic symptoms than boys; particularly among overweight girls (McCabe & Ricciardelli, 2009) About 40% of overweight girls and 20% of overweight boys engaged in at least one of the disordered eating behaviors (binge eating and / or extreme weight control) (D. R. Neuma rk Sztainer et al., 2007) However, past research has not evaluated the racial/ ethnical differences of weight control behaviors among adolescents. Studies further identified that dissatisfaction with body size is correlated with eating disorders, ineff ective dieting, low self esteem, depression, and high risk behaviors among adolescent girls (Liechty, 2010) In addition, dissatisfaction with body size contributes to Body Image Distortion (BID), the perceived discrepancy between their current and ideal body among both men and women (Liechty, 2010) Common strategies utilized as a result of dissatisfaction with body size include dieting and exercise (Heywood & McCabe, 2006) Additionally, weight body concerns were found to have a strong correlat ion with UWCB (D. Neumark Sztainer et al., 2003; D. R. Neumark Sztainer et al., 2007) Regardless of body weight, dieting practices were prevalent among female students (Malinauskas, Raedeke, Aeby,

PAGE 12

12 Smith, & Dallas, 2006) Researchers indicated that t his may be caused by the fact that girls reported relatively more weight and shape problems and concerns than boys (Smolak & Levine, 2001) Moreover, studies have suggested that female students with an inflated body weight percepti on were significantly more likely to engage in UWCB and report depressive symptoms than were female students with an accurate body weight perception (Harring, Montgomery, & Hardin, 2010; Rahman & Be renson, 2010) Despite the physical and psychological effects on adolescents, preliminary studies showed that UWCB is associated with other risk behaviors such as illicit drug use, tobacco use, binge drinking, (Haley Hedberg, & Leman, 2010) physical and sexual dating violence against adolescent girls (Silverman, Raj, Mucci, & Hathaway, 2001) and suicidal ideation (H aley et al., 2010; Kim, Cho, Cho, & Lim, 2009) However, s uch associations were mostly based on small samples and have not been well established from a national representative data. To monitor health risk behaviors among youth and young adults, the C enters for D isease C ontrol and Prevention (CDC) established the nations biennial epidemiologic surveillance system in 1991. The Youth Risk Behavior Surveillance System (YRBSS) monitors the prevalence of obesity and asthma as well as the six categories of priority high risk behaviors, including behaviors that contribute to unintentional injuries and violence, tobacco use, alcohol and other drug use, sexual behaviors, unhealthy dietary behaviors, and physical inactivity. The YRBSS surveys students in grades 912 from nationally sampled schools across the United States, thus providing the most representative data for youth health practices. The Youth Risk Behavior Survey (YRBS) has been given biennially since 1999. Each official brief descriptive statistics we re reported in the year following the survey. The data collected by the YRBSS is an

PAGE 13

13 important source of information in the area of unhealthy weight control behaviors and were used to analyz e trends, associat e risk behaviors, and identify predictors in this study. Overweight and Obesity among Adolescents in the U nited S tates The most frequently cited data on obesity prevalence rates over time in the U.S. came from the results of the National Health and Nutrition Examination Survey (NHANES). The National Cent er for Health Statistics (NCHS) periodically collects measured heights and weights in representative samples of adults, adolescents, and children in the U.S. NHANES I in 1971 74 indicated nearly 6% of adolescents were overweight (Catenacci, Hill, & Wyatt, 2009) The trend of overweight among adolescents has been increasing since NHANES I. Although no statistically significant linear trends were found over the time periods, the most recent NHANES data uncover the fact that the rate of overweight adolescents had reached 34% in the 20072008 survey (Ogden et al., 2010) These results demonstrated a six fold increase over the past 30 years. In addition, the NHANES data also suggest ed the prevalence of obesity increased from 5% to 18% during the same period among adolescents (Ogden & Carroll, 2010) Being overweight or obese has life long health impacts including heart disease, stroke, hyperte nsion, diabetes, and various forms of cancer Research estimated that up to 80% of overweight or obese children are likely to remain obese as adults (Guo & Chumlea, 1999) Moreover, obese adolescents were 16 times more likely to develop severe obesity, which can lead to severe health consequences, in their young adulthood than normal weight or overweight adolescents (The, Suchindran, North, Popkin, & GordonLarsen, 2010) While the rates of overweight and obese increased over the years, overweight and obese adolescents were found to engage themselves more in weight control behaviors, both healthy and unhealthy than their nonoverweight peers (Fonseca et al., 2009)

PAGE 14

14 Healthy Weight Control Behaviors (HWCB) Weight loss practice was found prevalent among adolescents. According to 2009 YRBS data, 44.4% of adolescents were trying to lose weight. Overall, the prevalence of weight control practice was higher among female (59.3%) than male (30.5%) students (Eaton et al., 2010) The majority of the weight control behaviors which are typically viewed as health promoting acts include d decreased fat intake and increased physical activity (D. Neumark Sztainer, 1999) Nationally, the American Dietetic Association supported it since late 1980s and stated that the achievement and maintenance of optimal weight through sound e ating practices and adequate exercise are an essential health promotion strategy (Position of the American Dietetic Association : Optimal weight as a health promotion strategy.1989) In addition, the traditional medical models promote the use of healthy weight loss methods, such as increased physical activity and moderate modifications in the types and amounts of food consumed (D. Neumark Sztainer, 1999) Engaging in healt hy weight control behaviors to lose body weight can help to gain significant health benefits. Research reported that m oderate weight control behaviors such as increased fruit and vegetable intake, a lowered fat intake, and increased physical activity can b enefit adolescents in maintaining idea l weight, reducing the risk of becoming obese, as well as improving fitness (Adam, Westenhofer, Rudolphi, & Kraaibeek, 2009; Daee et al., 2002) Healthy weight control behaviors are lifestyle interventions, which include physical activity, nutrition education, and behavior therapy, and result in extended effectiveness of weight control For example, a lifestyle intervention, which included nutrition education, physi cal and exercise intervention, psychological intervention, parents education, and medical lessons, was conducted in 129 centers specialized in pediatric obesity care in Germany. This study recruited 21,784 overweight children and adolescents aged from 2 to 20. Results demonstrated a longterm

PAGE 15

15 effectiveness in weight control among overweight children and adolescents after 6 months intervention and at a 2 year follow up (Reinehr et al., 2009) Another randomized treatment stud y with 10 year follow up was conducted to identify the variables associated with long term weight control among overweight children aged 6 12 years old. The basic treatment for all groups included weekly meetings for 812 weeks and monthly meetings last ing for 6 12 months from the start of the program. Results showed significant treatment effects in all groups. Particularly, the aerobic exercise and lifestyle intervention group s demonstrated the most significant reduction in overweight proportion among obese children after 10 years of follow up. Moreover, the lifestyle physical activity intervention combined with diet showed the superior decrease in overweight porportion when compared to structured aerobic exercise or calis thenics alone (Epstein, Valoski, Wing, & McCurley, 1994) Physical activity contributes to many positive benefits in weight control, especially in long term effectiveness. Fulton and colleagues reviewed the weight loss treatment and weight gain prevention studies conducted among adolescents and suggested that treatment of uncomplicated childhood obesity should reinforce healthy eating and physical activity behaviors to achieve weight loss without adverse effects on growth. In addition, lifestyle exercise programs, which incorporate unstructured physical activity or activity that is deliberately to increase the physical activity in an individuals daily regimen had been shown to produce the best and consistent long t erm results (Fulton, McGuire, Caspersen, & Dietz, 2001) On the other hand, decreases in sedentary behavior had also been shown to facilitate weight loss among children and adolescents because they have substituted a ctive behavior and thereby lowered the food intake (K. N. Boutelle, Libbey, Neumark Sztainer, & Story, 2009; Epstein et al., 1995; Epstein, Roemmich, Paluch, & Raynor, 200 5)

PAGE 16

16 The other core component of standard behavior weight loss interventions involves dietary prescription to produce a healthy, caloriereduced diet. In order to determine whether adolescent eating behavior s such as intake of fruit, vegetable, snack foods, and reduced calorie snack foods, are associated with BMI change over the course of weight loss program, researchers recruited 72 adolescents aged 13 16 years old to participate in a 16week weight loss trial. Results identified male adolescents who had higher initial frequency of vegetable intake, increased frequency of fruit intake, and reduced calorie snack foods accounted for 43% of the variance in BMI reduction. Therefore, researchers suggested early changes in eating habits, including increased frequency of vegetable and fruit intake can promote greater BMI reduction among adolescents (C. N. Hart et al., 2010) In summar y the recommended weight control methods among adolescents should include lifestyle mod ification of modest eating and physical activity as well as decreases in sedentary behaviors and early initiation of increased frequency of vegetable and fruit intake to maintain long term weight control outcomes. Unhealthy Weight Control Behaviors (UWCB) The concerns of overweight and obesity among adolescents had been shown to promote the use of healthy weight control behaviors such as decreasing fat intake, increasing fruit and vegetable consumption, and increasing physical activity (D. Neumark Sztainer, Story, & French, 1996) But, i n the meantime, the Unhealthy Weight Control Behaviors (UWCB), such as fasting, taking diet pills or laxatives, or inducing vomiting, (Centers for Disease Control and Prevention (CDC), 2010b) (Centers for Disease Control and Prevention (CDC), 2010b) remained very common among adolescents (Centers for Disease Control and Prevention (CDC), 2010b) The first UWCB study examin ed CDCs 1999 YRBS data and indicated that among adolescents trying to lose weight or stay the same weight, 62% of females and 41% of males combined

PAGE 17

17 exercise wit h a reduced fat and calorie diet, while 32% of females and 17% of males also practiced UWCB (Lowry, Galuska, Fulton, Wechsler, & Kann, 2002) With a closer look at the YRBS results concerning the UWCB among adolescents in the U.S., the most recent YRBS report indicated that 10.6% of adolescents fasted, i.e. did not eat for 24 or more hours to lose weight or to keep from gaining weight during the 30 days before the survey. The prevalence of fasting was higher among femal e (14.5%) than male (6.9%) (Eaton et al., 2010) The Morbidity and Mortality Weekly Report (MMWR) stated that the prevalence rates of fasting increased during 19992001 (12.6% to 13.5%) and then decreased during 20012009 (13.5% to 10.6%). In addition, 5% of students had taken diet pills, powders, or liquids without a doctors advice to lose weight or to keep from gaining weight during the 30 days before the survey. The prevalence was higher among female (6.3%) than ma le (3.8%) students, especially among W hite females (7%) and Hispanic females (6.9%). The prevalence of taking diet pills exhibited similar trends as fasting increased during 19992001 (7.6% to 9.2%) and then decreased during 20012009 (9.2% to 5.0%). The other unhealthy weight control practice is vomiting or taking laxatives. Four percent s of high school students in the U.S. had vomited or taken laxatives to lose weight or to keep from gaining weight during the 30 days before the survey. The prevalence rate was higher among Hispanic females (6.9%) when compared with other ethnic groups. Trends of vomited or taken laxatives increased during 19952003 (4.8% to 6.0%) and then decreased during 20032009 (6.0% to 4.0%) (Eaton e t al., 2010) Although MMWR has indicated the overall trends of decrease for these three UWCB in recent years, the detailed trends of UWCB and healthy weight control behaviors grouped by gender, ethnicity, or weight perception remain unavailable.

PAGE 18

18 Health Consequences of UWCB In order to lose weight, adolescents engaged in all kinds of weight control behaviors. As a result, w eight control behaviors have become increasingly widespread among adolescents, especially females. They do so not only to achie ve ideal thinness but also due to the bombardment of messages and images of so called beauty from the media. When adolescents take the shortcut by engaging themselves in UWCB, potential dire outcomes can exacerbate the health conditions both physiologicall y and psychologically. One of the major consequences of UWCB for adolescen t s is long term weight gain and increased risk for obesity later in life (Stice, Cameron, Killen, Hayward, & Taylor, 1999) Research showed that adolescents who reported dieting and other weight loss efforts, either healthy or unhealthy, were more likely to gain weight than those who did not report these efforts. The same study also revealed that UWCB is also predictive of greater weight gain over 3 years (Daee et al., 2002) Similarly, a 5 year longitudinal study examining overweight adolescents identified UWCB as a major risk factor for disordered eating and suggested weight gain over time (D. Neumark Sztainer et al., 2006; D. Neumark Sztainer et al., 2009) A study examined the biomedical consequences of severely obese outpatients on a very low calorie diet and suggested dieting resul t ed in significant short term improvements in health, such as body weight, systolic and diastolic blood pressure, triglycerides, and cholesterol. However, only a small percentage (less than 25%) of individuals actually maintain ed their lowered weight over substantial time periods (K. E. Hart & Warriner, 2005) Moreover, dieting without exercise was found to be associated with an increased risk of eating disorders later among youth. ( Patton, Selzer, Coffey, Carlin, & Wolfe, 1999) In addition, UWCB can impact adolescents psychological health, particularly during a time of rapid physical, psychological, and social development (Daee et al., 2002) A dolescents who

PAGE 19

19 were engaged in UWCB exhibited the most negative patterns of psychosocial and health behaviors. Neumark Sztainers studyshowed the associations between UWCB and other problem behaviors, such as alcohol or marijuana use, school delinquency, u nprotected sex, and suicide attempts (D. Neumark Sztainer et al., 1996) Additionally, it was reported that the more serious danger about UWCB is the subsequent development of eating disorders, including anor exia nervosa, bulimia nervosa, and binge eating (Daee et al., 2002) A large sample study also suggested that adolescents who reported binge / purge to lose weight were less likely to report sense of purpose, self esteem personal power, and positive view of personal future (French et al., 2001) Similarly, a study suggested dieting tends to cooccur with measures of diminished emotional well being, such as low self esteem, body dis satisfaction, and depressive symptoms, regardless of weight status (Crow, Eisenberg, Story, & Neumark Sztainer, 2006) Furthermore, research es showed that an individuals cognitive functioning may be negatively affected by UWCB (Mathias & Kent, 1998) (D. Neumark Sztainer et al., 1996; D. Neumark Sztainer, Story, French, & Resnick, 1997) (D. Neumark Sztainer et al., 1996; D. Neumark Sztainer, Story, French, & Resnick, 1997) In regards to the major physiologic consequences of UWCB among adolescents, diet pills take the center stage. A clinical study showed diet pills can only produce limited e ffectiveness to palliation of overweight and obesity {{593 Bray,G.A. 2009}}, but the side effect s may jeopardize adolescents health. The acute side effects of taking diet pills can cause problems in the cardiovascular system (arrhythmias, hypertension, an d myocardial ischemia) and the central nervous system (tremulousness, insomnia, psychosis, and seizures). Researchers identified cardiac valvular lesions and pulmonary hypertension after long term use of diet pills, which

PAGE 20

2 0 contain fenfluramine and dexfenfluramine (Daee et al., 2002) Due to health complications, these diet pills had been withdrawn from clinical use since 1997 (Bray, 2009) (Bray, 2009) Other health consequences of UWCB in clude d fatigue, dizziness, lack of energy, sore throat, abdominal pain, constipation, loss of dental enamel, and pancreatitis (Pritts & Susman, 2003) These medical complications may be irreversible and may simultaneously cause growth retardation, structural brain changes, and pubertal delay or arrest (Golden et al., 2003) To summarize, UWCB may help to control body weight in short term. However, adolescents who engaged in UWCB would eventually gain weight overtime. In addition, UWCB can jeopardize adolescents health both physiologically and psychologically. Further studies are warranted to identify pred ictive factors to help developing prevention and intervention programs. Predictive Factors of UWCB Research efforts have contributed to identify the predictive factors to explain why adolescents engage in UWCB. Selected factors related to UWCB have be en shown to includ e gender, race/ ethnicity, overweight status, weight concern, body image distortion, sedentary behaviors, dietary behaviors, and physical activities. These predictive factors from the literature are summarized below. Gender, Race/ Ethnicity and Overweight Status Recent studies highlighted the importance of gender differences in understanding and predicting adolescent weight control behaviors. In fact, most of the studies concerning weight related issues examined the gender differences. This indicates researchers support on genders role in weight loss. Among the studies, female adolescents reported more frequent engagement in both healthy and unhealthy weight control behaviors (Fonseca et al., 2009) I n addition, studies examined the nature of UWCB among overweight and normal weight adolescents indicated that girls engaged in more bulimic symptoms than boys; particularly among overweight

PAGE 21

21 girls (K. Boutelle, Neumark Sztainer, Story, & Resnick, 2002; Crow et al., 2006; McCabe & Ricciardelli, 2009) Study indicated that although female adolescents tend to have lower overweight prevalence than male adolescents, they are more likely to engage in UWCB (Lowry et al., 2002) Because females were more likely to engage in weight control beha viors, some studies purposely recruit ed females only. Although statistically significant gender differences were observed, with females significantly more likely than men to report fasting and vomiting, but effect sizes (Number Needed to Treat) were foun d to be small at 4% (Striegel Moore et al., 2009) Genders influences over UWCB are notably worthwhile for further investigat tion After years of intensive studies and efforts in weight control interventi ons, the prevalence of overweight, obese, and UWCB among high school youth did not decrease significantly. Ethnic differences were considered as a confounding variable by many health professionals. A regional study in Oregon using 2005 YRBS questionnaire f ound that American Indian and Hispanic girls had higher prevalence rates of UWCB when compared with non Hispanics Caucasians (Haley et al., 2010) Another regional study in Massachusetts revealed the lowest prevalence of UWCB among Caucasian youth and Asian youth, while Hawaiian/Pacific Islander and American Indian/Alaskan Native youth exhibited the highest prevalence (Silverman et al., 2001) A national survey targeting Ameri can Indian Alaska Native adolescents also revealed UWCB were common among American Indian adolescent with 27% report ed self induced vomiting at some time to try to lose weight and 11% reported using diet pills (Story et a l., 1994) Additionally, the odds of UWCB reached 10 times higher among female students who belong to these two aforementioned high prevalence groups when compared to W hite females (Austin et al., 2011) Further s tudies using the national representative data to explore the ethnic differences are suggested.

PAGE 22

22 Many studies also identified overweight status as a major factor related to UWCB among adolescents. Studies have suggested that overweight youth are more likely to engage in weight control behaviors, both healthy and unhealthy than their nonoverweight peers (Austin et al., 2011; Fonseca et al., 2009; Foti & Lowry, 2010) In addition, a cross sectional study consisting of a statewide representative sample of 7th 11th grade public school students in Connecticut suggested overweight adolescents use more unhealthy weight management strategies and avoid the healthier strategies, such as increased physical activity or healthier eating (K. Boutelle et al., 2002) Another cross s ectional study, which aimed to explore the perception of body size among overweight adolescents and to determine whether they consider it necessary to adjust weight and adopt healthier life habits, revealed overweight adolescents are eager to lose weight, and more likely to choose unhealthy methods (Chen, Fan, Jane, & Wu, 2009) Weight Perception, Body Image Distortion, and Weight Loss Intention Weight perception represents an important predictor of diet and weight man agement behaviors especially for those who perceived overweight (Cheung, Ip, Lam, & Bibby, 2007) As shown in the literature individuals who perceived their body as fat tend ed to engage in weight control behaviors in both health and unhealthy methods (Fonseca et al., 2009) Progressively, the UWCB prevalence increased from those who perceived themselves as thin, to those who perceived themselves as being the right size and those who perceived themselves as fat (Fonseca et al., 2009) Furthermore, weight body concern has been shown to significantly correlate to UWCB (D. Neumark Sztainer et al., 2003) especially those who see themselves as overweight or underweight (Cook, MacPherson, & Langille, 2007) Based on the YRBS reports, although the prevalence of overweight increased from 1999 to 2007, the prevalence of perceived overweight did not change among US high school students. In fact, the prevalence of perceived overweight decreased among non overweight adolescents (Foti

PAGE 23

23 & Lowry, 2010) While these results could s uggest positive implications for reducing UWCB, the relationship between weight misperception and UWCB remains unclear. For example, a study examining 2003 YRBS data to explore the prevalence of misperceptions and weight control behavior among normal weigh t adolescents revealed that a significant portion of normal weight adolescents who misperceived themselves as overweight engaged in UWCB (Talamayan, Springer, Kelder, Gorospe, & Joye, 2006) In addition, results f rom the National Longitudinal Study of Adolescent Health indicated girls with Body Image Distortion (BID), which refers to a discrepancy between perceived and actual weight status, exhibited 4.3 times greater odds of extreme weight loss behavior s (e.g., vomiting, laxatives, diet pills) and 2.3 times higher odds of dieting to control weight 1 year later (Liechty, 2010) A study aimed to explore the relationships between weight misperception and weight contr ol behaviors, especially UWCB among women, suggested normal weight females with misperception of their body weight were more likely to report healthy and unhealthy weight reduction behaviors compared with normal weight females who perceived themselves norm al weight (Rahman & Berenson, 2010) In addition, studies explored the gender difference on weight misperception and UWCB among adolescents indic ated that female students with an inflated body weight perception were si gnificantly more likely to engage in UWCB and to report depressive symptoms than were female students with an accurate body weight perception (Cheung et al., 2007; Harring et al., 2010) The in fluences of body weight distortion over UWCB we re reported statistically significant. However, how the direction of misconception, by the actual weight status and perceived weight status, is associated with UWCB prevalence remains unclear. Whether a doser esponse relationship exists is of interest to investigate.

PAGE 24

24 Weight loss intention can be an important factor determining the selection of weight control behavior whether healthy or unhealthy. A large sample study suggested UWCB was significantly higher amon g those not dieting but thinking they should (Fonseca et al., 2009) Sedentary Behaviors Physical Activity and Dietary Behaviors Physical activity creates many positive benefits for children and adolescents, and pr ovides an opportunity for play activity and pleasure. Physical activity not only facilitates the development of motor skills and increases energy expenditure (Batch, 2005) but also provides more effective and long te rm weight control results Physical activity also has been shown to affect body composition favorably during weight loss by preserving or increasing lean mass while promoting fat loss (Fulton et al., 2001) Numerous studies indicated that the combination of proper nutrition and regular physical activity is the most effective intervention for weight loss and maintenance of weight loss (Rippe & Hess, 1998) These studies focused mainly on the predictors of UWCB such as, gender, overweight status, and BID as well as the benefits of physical activity on weight control practices. V ery limited research investigated the relationship among UWCB, sedentary behavior, and physical activity While the limited research on this topic persists, a recent longitudinal study suggested that moderateto vigorous physical activity wa s a protective factor against UWCB among adolescents while sedentary behaviors such as, watching TV, using computer, and reading or doing homework did not show significant association (D. R. Neumark Sztainer et al., 2007) In addition, study has shown overweight adolescents were less likely to engage in vigorous physical activity and more likely to engage in UWCB than nonoverweight youths (K. Boutelle et al., 2002) In r egard to dietary behaviors, more healthy food choices can usually help to prevent UWCB. However, the 1999 YRBS data showed that, among male adolescents, the consumption

PAGE 25

25 of greater than 5 servings/day of fruits and vegetables was associated with engaging in fasting for weight control, while female adolescents who consumed the same amount of fruits and vegetables were more likely to use exercise, diet, fasting, vomiting, and laxatives for weight control (Lowry et al., 2002) A school based, multi component intervention, which included one hour of physical activity and the provision of a low fat, nutrient dense snack at individual level, designed to reduce the engagement in UWCB among elementary schoolchildren suggested sig nificant improvement of confidence to impact weigh teasing norms. However, the prevalence of UWCB remained the same in the intervention group while the control group showed decreased prevalence (Haines et al., 2006) H ow dietary behavior intervention interacts with other psychosocial factors and how it relates to UWCB remain s unclear and demands further investigation. In sum, the relationships between UWCB and sedentary behavior, physical activity, and dietary behavior still remain tentative especially without a national representative data; thus warrant further studies. Associations between UWCB and Other Risk Behaviors Health risk behaviors usually link with each other. As Dr. DiClemente suggested, practitioners and r esearchers should attempt to address multiple behaviors, not merely one health habit at a time, and consider contextual and environmental factors necessary to understand the big picture and more effectively influence health behaviors (Werch, Ames, Moore, Thombs, & Hart, 2009) While the health consequences of UWCB may not cause immediate life threatening dangers, studies have suggested that UWCB can lead to more complicated ailments, such as eating disorders, anorexia nervosa, and bulimia nervosa. In addition, UWCB has been reported to associate with a cluster of risk behaviors, including alcohol, tobacco, and illicit drug use, sexual risk behaviors, and suicide ideation/attempts. Researchers suggested that adolescents who engage

PAGE 26

26 in UWCB are more likely to engage in a range of other health compromising behaviors, including suicide attempts, delinquency, tobacco, alcohol, and marijuana use, unprotected sexual intercourse, and multiple sexual partners (D. Neumark Sztainer et al., 1996) The related literature of the selected risk behaviors under investigation are reviewed as following. Alcohol, Tobacco, or Drug Use Alcohol, tobacco, and drug use are frequently associated with each other. This review of literature found it difficult to separate their associations with UWCB by individual substances. UWCB among adolescents was found to correlate with cigarette smoking, appearance concerns, (Cance, Ashley, & Penne, 2005; Croll, NeumarkSztainer, Story, & Ireland, 2002; French et al., 2001) (Cance, Ashley, & Penne, 2005; Croll et al., 2002; French et al., 2001) and alcohol use (Croll et al., 2002) Researchers examined UWCB and found them associated with alcohol consumption among adolescents (Cance et al., 2005; Croll et al., 2002; French et al., 2001; Haley et al., 2010) Underage alcohol drinking, especially heavy drinking, wa s associated with numerous negative consequences, such as drunk driving, acute impairment, violence, poor school performance, and suicide (National Research Council (US) and Institute of Medicine (US) Committee on Developing a Strategy to Red uce and Prevent Underage Drinking, 2004) The m ajority of the inter connected web of risk behaviors were reported by researchers with smaller sample size studies Hence more in depth investigations are warranted. In addition to alcohol, tobacco use is also associated with UWCB among adolescents. Researchers pointed out the fact that some students smoked cigarettes as a method of weight control (Johnson, Eaton, Pederson, & Lowry, 2009) especially among female ad olescents (Lowry et al., 2002) The literature also showed the increased likelihood of UWCB among smokers. In terms of risk adolescent smokers were significantly more likely to engage in UWCB (Strauss & Mir, 2001) with odds ratio approximately 2 times higher when compared with those

PAGE 27

27 who did not smoke (Johnson et al., 2009) More specifically, adolescents who smoked currently or in the pas t were more likely than nonsmokers to report purging (vomiting, laxatives, diuretics) (Breitkopf & Berenson, 2004) In regard to the relationship between UWCB and drug use among adolescents, those who engaged in UW CB reported more health compromising behaviors such as substance use in the past year (McGuire et al., 2002) The YR B S 2003 report suggested juvenile delinquents, which include illicit drug use, were more likely tha n nondeliquents to engage in UWCB to lose weight or to keep from gaining weight (Ho, Kingree, & Thompson, 2006) In addition, adolescent females who had used laxatives or vomited to lose weight in the past year were mor e likely than those who had not used substances during the past year, including ecstasy, inhalants, nonmedical psychotherapeutics, marijuana, cigarettes, and alcohol (Cance et al., 2005) Girls who reported the use of illicit drugs or tobacco were four times more likely to report UWCB than non users. Among both sexes, UWCB was found significantly associated with all substance use behaviors including illicit drug use, tobacco use, and binge drinking (Haley et al., 2010) On the other hand from the prevention standpoint, adolescents who reported abstinence from alcohol, drug, or sex were less likely to report binge eating or purging (French et al ., 2001) Suicidal Ideation and Suicidal Attempt Suicide wa s the third leading cause of death among youth ages 15 19 years in 2010. A prior suicide attempt wa s one of the most significant risk factors for a fatal adolescent suicide attempt (Centers for Disease Control and Prevention (CDC), 2010b) Adolescents who engaged in UWCB are more likely to report a history of physical and sexual abuse, and suicide attempts in the past year (McGuire et al., 2002) In addition, the adolescents engaging in multiple UWCB showed greater risk of having suicidal thoughts (Kim et al., 2009) In terms of their ris k, adolescents with suicidal ideation exhibited 3 times higher odds engaging in UWCB when

PAGE 28

28 compared with their counterparts (Haley et al., 2010) As discussed previously, female adolescents were more likely to engage U WCB. Those female adolescents who engaged in UWCB exhibit 4.2 times higher odds of suicidal thoughts, 2.9 times higher odds of suicide planning, and 3.4 times higher odds of suicide attempts than their counterparts (Cook et al., 2007) Sexual Risk Behaviors Early initiation of sexual intercourse was associated with having a greater number of lifetime sexual partners. In addition, adolescents who initiate sexual intercourse early were less likely to use contraception an d were at higher risk of unexpected pregnancy (Centers for Disease Control and Prevention (CDC), 2010b) Recent estimates suggested that while representing 25% of the ever sexuall y active population, young people aged 1524 acquire d nearly half of all new STDs (Weinstock, Berman, & Cates, 2004) There are significant relationships between UWCB and sex related behaviors. For example, research ers suggested that female adolescents who engaged in UWCB were more likely to report being sexually active (McGuire et al., 2002) Data from 1997 and 1999 Massachusetts Youth Risk Behavior Surveys indicated that phy sical and sexual dating violence against adolescent girls is associated with increased risk of UWCB (Silverman et al., 2001) Quantitatively speaking, the adolescent girls who engaged in UWCB exhibited 1.6 times higher odds of ever having had vaginal intercourse compared with the girls who did not engaged in UWCB (Cook et al., 2007) When examined the relationships between UWCB and sexual behaviors from the prevention perspecti ves, researchers suggested that values related to abstinence from alcohol and sex appear to be protective against unhealthy eating behaviors and may reflect a general resilience that buffers against a broad range of health risk behaviors (French et al., 2001)

PAGE 29

29 The above reviewed literature based on researcher initiated studies with relatively smaller sample sizes show ed preliminary evidences of the association between UWCB and other selected risk behaviors. These findings support the aggregation of risk behaviors among adolescents. However, the strength of association varies from study to study. While some studies examin ed the complete definition of UWCB, some others examined UWCB sub behaviors separately. The use of the nationally representative YRBS data in this study can provide greater statistical power to examine the detailed association between UWCB and other select ed risk behaviors. Description of the Youth Risk Behavior Survey (YRBS) The Division of Adolescent and School Health of the Centers for Disease Control and Prevention established this biennial epidemiologic Youth Risk Behavior Surveillance System (YRBSS) i n 1991 to monitor health risk behaviors among youth and young adults at national, state, and local levels. The YRBSS surveys students in grades 9 12 from scientifically sampled schools across the United States, thus provides the most representative data fo r youth health practices. CDC designed YRBSS to determine the prevalence of health risk behaviors among high school students, to assess whether these behaviors increase, decrease, or stay the same over time, and to examine the co occurrence of health risk behaviors. YRBSS focuses almost exclusively on healthrisk behaviors, rather than the determinants of these behaviors (e.g., knowledge, attitudes, beliefs and so on ). In addition, YRBSS is designed not only to provide comparable national, state, and local data as well as comparable data among subpopulations of youth (e.g., racial/ethnic groups), but also to monitor progress toward achieving the national health objectives for 2000 and 2010 along with other program indicators (Brener et al., 2004) The Youth Risk Behavior Survey (YRBS) is the major component of the YRBSS. It contains six categories of priority high risk behaviors that contribute substantially to the leading causes of death, disability, and social probl ems among youth and adults in the United States.

PAGE 30

30 These six categories include: 1) behaviors that contribute to unintentional injuries and violence; 2) tobacco use; 3) alcohol and other drug use; 4) sexual behaviors that contribute to unintended pregnancy a nd sexually transmitted diseases (STDs), including human immunodeficiency virus (HIV) infection; 5) unhealthy dietary behaviors; and 6) physical inactivity, plus overweight and asthma (Centers for Disease Control and Prevention (CDC), 2010a) Population and Sampling The national school based YRBS was conducted by CDC since 1991 biennially using a similar three stage cluster sampling design to obtain a nationally representative s ample of high school students in the United States. The target population consisted of all public and private school students in grades 9 to 12 in the 50 states and District of Columbia. This study retrieve d only the data since 1999 because YRBS from 1991 to 1997 did not include UWCB related questions. In the years 1999, 2001, 2003, 2005, 2007, and 2009, school response rate were 77%, 75%, 81%, 78%, 81%, and 81%, respectively; student response rate were 86%, 83%, 83%, 86%, 84%, and 88%, respectively; overall response rate were 66%, 63%, 67%, 67%, 68%, and 71%, respectively; and sample size were 15,349, 13,601, 15,214, 13,917, 14,041, and 16,410, respectively (Eaton et al., 2006; Eaton et al., 2008; Eaton et al., 2010; Grunbaum et al., 2002; Grunbaum et al., 2004; Kann et al., 2000) Threestage cluster sampling desi gn The national YRBS use d a threestage cluster sampling design to produce a representative sample of 9th through 12th grade students in the United States. In the first stage, the Primary Sampling Units (PSUs), consisting of large counties or groups of sm aller adjacent counties selected from 16 strata formed according to the degree of urbanization and the relative percentages of Black and Hispanic students in the PSU. The PSUs were selected with probability proportional to the total school enrollment in the PSU. In the second stage of sampling, schools

PAGE 31

31 were selected with probability proportional to school enrollment size. Schools with substantial numbers of Black and Hispanic students were sampled at relatively higher rates than were other schools. The fina l stage of sampling comprised of random selection from each chosen school and grade level 1 or 2 intact classes of either a required subject (e.g., English or social studies) or a required period (e.g., homeroom or second period). All students in sampled c lasses were eligible to participate. Schools, classes, and students that refused to participate were not replaced. Weighting A weighting factor based on student gender, race/ethnicity, and grade was applied to each record to adjust for student non respo nse and for oversampling of Black and Hispanic students. Final overall weights were scaled so that the weighted count of students equaled the total sample size and the weighted proportions of students in each grade matched national population projections. Therefore, weighted estimates are representative of all students in grades 9 12 attending public and private schools in the United States. Instrumentation A questionnaire containing approximately 88 items (the exact number varied by survey year) was admi nistered in the classroom by trained data collectors each survey year Before survey administration, local parental permission procedures were followed. Students completed the self administered questionnaire during one class period and recorded their respo nses directly on a computer scannable booklet or answer sheet. An institution review board at the CDC reviewed and approved the YRBS. Validity and Reliability of YRBS The available literature indicated CDC had conducted a test retest reliability study o ut of the 1999 version of national YRBS in 2000. The questionnaire was administered to a convenience sample of 4,619 high school students. The questionnaire was administered on two

PAGE 32

32 occasions, approximately 2 weeks apart. Approximately one of five items (22 %) had significantly different prevalence estimates for the first and second times that the questionnaire was administered. Ten items (14%) had both kappas less than 61% and significantly different time 1 and time 2 prevalence estimates, indicating that th e reliability of these items may be questionable. Certain items (e.g., an item related to injury during physical activity) have been revised or deleted from later versions of the questionnaire (Brener et al., 2004) In 2000, CDC also conducted a study to assess the validity of the two YRBS questions regarding self reported height and weight. In that study, 2,965 high school students completed the 1999 version of the YRBS questionnaire on two occasions approximately 2 weeks apart. After completing the questionnaire, the students were weighed and had their height measured. Self reported height, weight, and BMI calculated from these values were substantially reliable, but on average, students in the study overreported th eir height by 2.7 inches and underreported their weight by 3.5 pounds. These results indicate YRBSS may potentially underestimate the prevalence of overweight in adolescent populations (Brener et al., 2004) The CD C had conducted a review of existing empirical literature to assess cognitive and situational factors that might affect the validity of adolescent self reporting of behaviors measured by the YRBS questionnaire in 2003. Results suggested although self repor ts of risk behaviors are affected by cognitive and situational factors, these factors do not threaten the validity of self reports of each type of behavior equally (Brener et al., 2004) Although no existing literature suggested high validity of current YRBS questionnaire, a logical consistency edits mechanism was used to improve the validity. The majority of the edit criteria compare two questions at a time to ensure logical consistency. If responses from two questions conflict logically, both questions are set to blank. For example, If students who

PAGE 33

33 responded they have never tried cigarette smoking, even one or two puffs and then responded a certain age to the other question asking How ol d were you when you smoked a whole cigarette for the first time?, both questions are set to missing data. However, data from demographic questions are not set to missing in this case (Centers for Disease Control and Prevention (CDC), 2010a) Another mechanism to ensure the data quality was subverting records. If a student answers 20 or less questions or answers with b, c, d, e, f, g, or h 15 or more times in a row, all values except the demographic variables are set to missing (Centers for Disease Control and Prevention (CDC), 2010a) Quality Control Procedures All hypothesis testing s w ere performed on weighted data using IBM SPSS Statistics 19. The 19992009 YRBS data sets, instructional menus, YRBS biennial reports, and questionnaires were downloaded from the Centers for Disease Control and Prevention, Division of Adolescent and School Hea lth, Youth Risk Behavior Surveillance Systems website. To avoid misunderstanding and errors during data processing, descriptive statistical procedures were performed and the results were compared with Youth Online, the web based inquiry system provided by YRBSS (http://apps.nccd.cdc.gov/youthonline/App/Default.aspx). Youth Online allows users to filter and sort data on the basis of race/ethnicity, sex, grade, or survey collection site, create customized tables and graphs, and perform statistical tests by s ite and health topic. The results inquired from Youth Online were compared with the descriptive statistics produced from the downloaded data sets. Inconsistent results, if any, were reviewed. In such incident, the data sets and descriptive statistical proc edures were examined and corrected to ensure the quality of data and future analysis for this study.

PAGE 34

34 Advantages and Disadvantages of Secondary Data Analysis Secondary data analysis is any further analysis of an existing data set which presents interpretat ions, conclusions or knowledge additional to, or different from, those presented in the first report on an inquiry and its main results. Given the huge human and financial investment in data collection by funders, investigators, and subjects, it seems reasonable if not imperative, ethically speaking, to make the maximum use possible of research data (Clarke & Cossette, 2000) The choice of primary or secondary data need not be an either/or question. Research methodologist suggested scientists to select data that are appropriate to the research question being studied and the resources available to the researchers. Several advant ages were detailed in the literature (Boslaugh, 2007) The first advantage of working with secondary data is economy in terms of money and time. Because someone else already collected the data, the researchers do not have to devote resources to this phase of research. In addition, the data are already collected, cleaned and stored in electronic format, thus the researchers can spend the bulk of their time analyzing the data. The second major advantage of using secondary data is the breadth of data available. Few researchers would have the resources to collect data from a representative sample of adolescents in every state in the U.S., let alone repeat this data collection process in every other year. However, the f ederal government conducts numerous surveys on that scale. The third major advantage in using secondary data is that often the data collection process is informed by expertise and professionalism that may not available to smaller research projects. For in stance, many of the federal health surveys use a complex sampling design and system of weighting that allows the researchers to compute population based estimates of health conditions and behaviors; thus obtains greater generalizability. To take another ex ample, data collection for

PAGE 35

35 many federal data sets is often performed by staff members who specialize in that task and who may have years of experience working on a particular survey. In Addition, secondary data analysis affords the researchers the opportu nity to explore variables of interest and study sub samples such as gender, race/ethnicity, grade, of the original group using a variety of statistical techniques. Lastly, when longitudinal data or data collected with same questions repeated annually or at regular intervals are available, researchers may track trends and patterns of health status and health behaviors in the population across time (Clarke & Cossette, 2000) One major disadvantage when using secondary data is inherent in its nature. Because the data were not collected to answer the researchers specific research questions, particular information that the researcher would like to have may not have been collected (B oslaugh, 2007) The second major disadvantage of using secondary data is that because the researcher did not participate in the planning and execution of the data collection process, he or she does not know exactly how it was done. More to the point, the researcher does not know how well it was done and therefore how seriously the data are affected by problems such as low response rate or respondent misunderstanding of specific survey questions (Boslaugh, 2007) However, this disadvantage could be overcome by closely examining the data available in the original document or literature. Delimitations The study employ ed secondary data analysis, which is inherent in the nature of the original survey. Because of this restriction, the research questions were confined within the original scope. Some possible factors reported in the literature as the outcomes of researcher initiated studies, such as body image, body dissatisfaction, peer influence (Hutchinson & Rapee, 2007) and frequent reading of magazines articles about dieting/weight loss and weight control

PAGE 36

36 behavior (van den Berg, Neumark Sztainer, Hannan, & Haines, 2007) cannot be included because these factors were not covered by YRBS (van den Berg et al., 2007) (van den Berg et al., 2007). In order to adjust data to become nationally representative, CDCs YRBSS administrative team had dec ided to give a weight based on student sex, race/ethnicity, and grade level to each record to adjust for school and student nonresponse and oversampling of African American and Hispanic students. The overall weights were scaled so that the weighted count o f students equals the total sample size, and the weighted proportions of students in each grade match the national population proportions. Thus the data are deemed representative of students in grades 9 12 in public and private schools in the United States (Centers for Disease Control and Prevention (CDC), 2010a) YRBS data sets have multiple limitations. First, all YRBS data are self reported, and the extent of under reporting or over reporting of behaviors may not be determined, although that the data retain acceptable quality (Brener et al., 2004) Second, data apply only to adolescents who attend school and therefore, are not representative of all persons in this age group. Third, YRBSS is designed to produce information to help assess the effect of broad national, state, and local policies and pr ograms. It was not designed to evaluate the effectiveness of specific interventions. Fourth, YRBSS only addresses behaviors that contribute to the leading causes of morbidity and mortality among youth and adults. Statement of Purpose The Centers for Disease Control and Preventions Morbidity and Mortality Weekly Report ( MMWR ) had briefly illustrated some trends of UWCB. For example, the most recent YRBS data indicated that 10.6% of adolescents fasted, i.e. did not eat for 24 or more hours, to lose weight o r to keep from gaining weight during the 30 days before the survey. In addition, the prevalence of fasting was higher among female (14.5%) than male (6.9%) (Eaton et al., 2010) However, no

PAGE 37

37 further analysis except des criptive statistics was done to date. More sophisticated analyses, including bivariate and multivariate analyses, to describe detailed trends of UWCB over the past 10 years, its risk factors, and its association with other risk behaviors we re lacking. This type of information would be very helpful when designing prevention and control program specifically related to the health problems associated with UWCB. The findings from these analyses can reflect more accurate and comprehensive trends to raise the awar eness of UWCB among the general public, school teachers and counselors, and the health professionals. While several studies using partial YRBSS data or researcher initiated surveys had identified some preventive and risk factors for UWCB, analyzing a more thorough, comprehensive, and nationally representative data brings much greater benefits and allows for more accurate scientific discovery with national generalizability. Significance Unhealthy Weight Control Behaviors (UWCB) could predict the onset of mo re severe disordered eating behaviors and eating disorders, including anorexia nervosa, bulimia nervosa, and binge eating disorder, as well as depression and other medical and psychosocial morbidity (Fonseca et al., 2009) Eating disorders are widely recognized sources of morbidity among adolescents. Although not meeting the Diagnostics and Statistical Manual of Mental Disorders fourth edition (DSM IV) diagnostic criteria for eating disorders, many adolescents engaged i n extreme weight loss practice(Croll et al., 2002) Such practice include d prolonged fasting, vomiting, and use of unprescribed weight loss medications. Study indicated that t hese unhealthy behaviors may not be immedia tely life threatening, but they can lead to electrolyte imbalance and other physical manifestations, may progress to frank anorexia or bulimia, and often are associated with other health risk behaviors (Haley et al., 2010) A longitudinal study suggested female adolescents who engaged in moderate dieting were

PAGE 38

38 five times more likely to develop an eating disorder than nondieters, and female adolescents who engaged in severe dieting were 18 times more likely to develop an eating disorder 2 years later (Patton et al., 1999) Early identification of eating disorders has been recommended as it leads to more successful treatment outcomes (Bulik, Berkma n, Brownley, Sedway, & Lohr, 2007) YRBSS was designed to focus the nation on adolescent behaviors contributing to the leading causes of mortality and morbidity and to assess how risk behaviors change over time (Eaton et al., 2010) Analysis of such data not only explore d the UWCB trends, predictive factors, and relationships with other risk behaviors, but also establishe d greater generalizability, founded on the large sampling size and weighting methods. Although risk factors such as peer influences, social norms about weight loss behaviors, media influences were not included in YRBS, variables such as gender, weight, height, perceived weight status, physical activity, dietary behaviors were collected and can help to design more effective intervention curriculum against UWCB, reduce risk, and improve health outcomes among adolescents. Research has shown that a curricular program that uses cognitive social learning, developmental, and contextual principals may help prevent development of negative body image and unhealthy weight management techniques in girls and boys as they move from late childhood to early adolescence (Smolak & Levine, 2001) The findings of this study could be used as the foundation to understand adolescents risk behaviors and to develop tailored UWCB prevention curriculum and interventions, and optimally reduce the morbidity and mortality caused by UWCB and promote adolescents health. Research Aims This study was designed to assess the trends of UWCB among high school students, to investigate the possible risk and protective factors of UWCB, and explore the associations

PAGE 39

39 between UWCB and selected risk behaviors by analyzing the YRBS data. The specific aims are as following: This study aimed to analyze YRBS data to Aim 1: describe the 19992009 prevalence trends of healthy and unhealthy weight control behaviors as defined by CDCs YRBSS. Aim 2: e xamine the differences of UWCB prevalence in 2009 by gender, ethnicity/race, overweight status, present weight control tendency and Body Image Distortion (BID). Aim 3: identify the predictive factors of UWCB. Aim 4: explore the relationships between UWCB and selected risk behaviors, including tobacco use, alcohol use, drug use, multiple sexual partners, suicidal ideation and attempt physical/sexual violence, and been bullied

PAGE 40

40 CHAPTER 2 TRENDS OF WEIGHT CONTROL BEHAVIORS AMONG HIGH SCHOOL STUDENTS B ackground The prevalence of obesity in the United States has been more than doubled over the past 3 decades (Y. Wang & Beydoun, 2007) The latest National Health and Nutrition Examination Survey (NHANES) data indicated t hat over one third of American adults were obese and over two thirds were overweight or obese (Flegal et al., 2010) Among adolescents aged 1219 years old, 18% have age specific BMI greater than 95th percentile, whi ch is considered obese, and 34% have age specific BMI greater than 85th percentile, which is considered overweight (Ogden & Flegal, 2010; Ogden et al., 2010) Such tendency from adolescence to adu l t hood, if persist, can cause enormous health problem s and result in significant economic lost. To prevent weight gain, people voluntarily engage in various weight control practices. Although many adolescents use healthy strategies, unhealthy or extreme w eight control strategies, such as fasting, using diet pills/powders/liquids, vomiting, or taking laxatives are also common among adolescents (Eaton et al., 2010) Unhealthy weight control behaviors, such as these (Eisenberg et al., 2005; Haines et al., 2006; D. NeumarkSztainer et al., 2003) could predict the onset of more severe disordered eating behaviors, including anorexia nervosa, bulimia nervosa, and binge eating disorder, as well as depression and other medical and psychosocial morbidity (Fonseca et al., 2009) Contrary to the commonly projected weight loss, dieting (particularly dur ing adolescence) had been shown to predict weight gain over time (Fonseca et al., 2009; D. NeumarkSztainer et al., 2009) In addition, adolescents who had unhealthy weight control behavi ors had been shown to have approximately 3 times greater risk for being overweight and 6 times risk for uncontrollable binge eating 5 years later (D. Neumark Sztainer et al., 2006)

PAGE 41

41 According to the 2009 Youth Risk Behavioral Surveillance System (YRBSS), 59% of female and 31% of male students in grades 912 reported trying to lose weight. In 2009, for the purpose of losing weight or keeping from gaining weight, 10.6% of high school st udents did not eat for 24 or more hours, 5.0% had taken diet pills, powders, or liquids without a doctors advice, and 4.0% had vomited or taken laxatives during the 30 days before the survey (Eaton et al., 2010) How ever, t he longitudinal trends by participants demographics have not been well investigated. It is important to examine such trends in order to identify the high risk populations for better intervention effectiveness. Researchers ha d found the distinction between sexes in related to unhealthy weight control behaviors among adolescents. G ender difference was found in body image, weight concern, and weight control behaviors. Literature showed girls engaged in more bulimic symptoms than boys; particularly amon g overweight girls (McCabe & Ricciardelli, 2009) In addition, about 40% of overweight girls and 20% of overweight boys engaged in at least one of the disordered eating behaviors (binge eating and / or extreme weight control) (D. R. Neumark Sztainer et al., 2007) Other demographic characteristics, such as race/ethnicity, have not been clearly identified. Although several studies ha d presen ted the ethnical differences in unhealthy weight control behaviors (Austin et al., 2011; Haley et al., 2010; Silverman et al., 2001; Story et al., 1994) the overall trends remain inconsistent In addition, adolescents body weight status makes a difference in weight control behaviors. Researchers found overweight youth were more likely to engage in weight control behaviors, both healthy and unhealthy, than their nonoverweight pee rs (Austin et al., 2011; Fonseca et al., 2009; Foti & Lowry, 2010) Furthermore, the interaction of gender and body image was found to be a significant contributor of unhealthy weight control behaviors. Girls with

PAGE 42

42 Body Image Distortion (BID), which refers to a discrepancy between perceived and actual weight status, exhibited 4.3 times greater odds of extreme weight loss behaviors (e.g., vomiting, laxatives, diet pills) and 2.3 times higher likelihood of dieting to control weight one year later (Liechty, 2010) Although the findings of the aforementioned studies are quite inspiring, they are either lack of national sampling representation, s mall in sample size, cross sectional design in timing, or inconclusive in findings. To advance scientific understanding of adolescents weight control behaviors, th is study examine d the longitudinal data from YRBSS nationally representative samples. This study was designed to describe the longitudinal trends of healthy and unhealthy weight control behaviors as defined by the Centers for Disease Control and Prevention (CDC) YRBSS among high school stud ents and to examine the prevalence differences by gender ethnicity/race, and BID. Results are expected to help identify ing the high risk populations and provide critical information in designing tailored weight control programs specifically targeting adolescents. M ethods T he Division of Adolescent and School Health at the CDC established the nations biennial epidemiologic surveillance system in 1991 t o monitor health risk behaviors among youth and young adults. The Youth Risk Behavior Surveillance System (YRBSS) monitors the prevalence of obesity and asthma as well as the six categories of priority high risk behaviors, including behaviors that contribute to unintentional injuries and violence, tobacco use, alcohol and other drug use, sexual behaviors, unhealthy dietary behaviors, and physical inactivity. The Y RBSS surveys students in grades 9 12 from nationally sampled schools across the United States and applies a weighting system, thus providing the most representative data for youth health practices. The Youth Risk Behavior Survey (YRBS) has been given bienn ially since 1999.

PAGE 43

43 O fficial brief descriptive statistics were reported in the Morbidity and Mortality Weekly Report in the year following the survey and YRBSS website. The data collected by the YRBSS is an important source of information in the research of weight control behaviors and were used as a basis for analyzing trends in this study. We analyzed weighted data from 1999 to 2009 because only these year s surveys asked about weight control behaviors CDCs YRBS groups fasting, taking diet pills or laxa tives, and inducing vomiting during the past 30 days to lose weight or to keep from gaining weight into Unhealthy Weight Control Behaviors (UWCB) (Centers for Disease Control and Prevention (CDC), 2010b) In this study, UWCB was operationally defined as: When a student reported yes in one or more of the following three YRBS questions, he/she was classified as having UWCB: During the past 30 days, did you go without eating for 24 h ours or more (also called fasting) to lose weight or to keep from gaining weight? During the past 30 days, did you take any diet pills, powders, or liquids without a doctors advice to lose weight or to keep from gaining weight? (Do not include meal repla cement products such as Slim Fast.) During the past 30 days, did you vomit or take laxatives to lose weight or to keep from gaining weight? YRBS also groups exercise and eat less food, fewer calories, or foods low in fat to lose weight or to keep from ga ining weight during the past 30 days into Healthy Weight Control Behaviors (HWCB). When a student reported yes in one or both of the following t wo YRBS questions, he/she was classified as having HWCB: During the past 30 days, did you exercise to lose weight or to keep from gaining weight?

PAGE 44

44 During the past 30 days, did you eat less food, fewer calories, or foods low in fat to lose weight or to keep from gaining weight? Body Image Distortion (BID) is the variable generated by comparing the perceived body weight and actual weight of each participant Overweight was defined as a Body Mass Index ( BMI ) being equal to or larger than the 85th percentile of the age (Kuczmarski et al., 2002) The participant who had thei r BMI below their respective 85th percentiles w as classified as non overweight. I n YRBS, p articipants were asked How do you describe your weight? The response option s include Very underweight Slightly underweight About the right weight Slightly overweight Very overweight The responses in the first three options were pool ed as perceived nonoverweight and the last two were considered perceived overweight. The body weight perception was cross tabulated with the respondents overweight status based on their actual BMI percentiles The participants whose BMI was equal to or larger than the 85th percentile and perceived themselves as overweight and the ones whose BMI was less than the 85th percentile and perceived themselves as nonoverweight were defined as no BID. The participants whose BMI was equal to or larger than the 85th percentile and perceived themselves as nonoverweight and the ones whose BMI was less than the 85th percentile and perceived themselves as overweight were defined as having B ID. Consequently, four weight X BID categories were formed in this study : nonoverweight without BID, nonoverweight with BID, overweight without BID, and overweight with BID. Analyses were performed o ver the weighted YRBS data using SPSS 19. All statistics were conducted using the YRBS 19992009 data downloaded from YRBSS website. To assure the quality of analysis, our descriptive statistics were fully compared with the official YRBSS statistics on the Youth Online. To examine longitudinal trends, linear a nd quadratic time (year)

PAGE 45

45 variables were entered into the curve estimation that tested for UWCB s and HWCB s secular trends The presence of a significant linear time coefficients (beta) indicates an overall increase or decrease over the years. A significan t linear time beta accompanied by a nonsignificant quadratic time beta indicates the presence of linear change. A significant quadratic time beta accompanied by a significant linear time beta indicates the presence of nonlinear change in addition to an ove rall increase or decrease over time. A significant quadratic time beta accompanied by a nonsignificant linear time beta indicates the presence of nonlinear change with no overall increase or decrease over time. A nonsigificant quadratic time beta accompani ed by a nonsignificant linear time beta indicate s the absence of any significant variation in the data. Bivariate logistic regression was conducted to examine the associations between weight control behaviors (UWCB and HWCB) and sex, grade, race/ethnicit y, and BID. Odd ratios and their corresponding 95% confidence intervals (95% CI) were generated over the pooled data from 19992009. R esults The Table 21 contained the demographic characteristics of respondents as well as the prevalence of BID, trying to lose weight, HWCB, and UWCB during 1999 to 2009. Overall, 88,547 participants completed YRBS across 10 years. Male and female adolescents were evenly distributed as well as the grade with slightly more (28.8%) 9t h graders and less (21.4%) 12th graders. Most students were Caucasian (60.8%). About 44.6% of students were trying to lose weight; 65.4% engaged in HWCB and 17.2% engaged in UWCB. The majority of students were nonoverweight without BID (61.3%) while 20.2% of students had BID. To describe the trends, quadratic and linear models were estimated in model fit. All quadratic models showed near ly complete collinearity among the model terms. Thus, only linear models were performed. The curve estimation revealed th at UWCB exhibited a linear time beta

PAGE 46

46 ( = 0.05, p<.001) showing a significant descending trend. Additional close in observation found the presence of a linear as c ending trend over 1999 (18.2%) to 2001 (19.4%) and an de scending trend from 2003 (19.4%) to 2009 (15.8%). The descending trend was found in the majority of the years. Analysis over HWCB prevalence rates exhibited a significant linear increase ( =0.03, p<.001) from 1999 (63.2%) to 2009 (66.5%). Additional close in observation found the trend had no apparent fluctuation over the 10 years. The Table 22 exhibits the UWCB prevalence rates by demographic subgroups and BID categories A mong these four c haracteristics, three subgroups consistently exhibited higher UWCB prevalence rate over their counterparts First, female adolescents consistently exhibited about twice as high prevalence rates of UWCB over males. Moreover, female adolescents exhibited about 2.5 times higher risk of engaging in UWCB than male adolescents (OR=2.51, p<.001). Second, American Indian/Alaska Native adolescents show e d the highest UWCB prevalence rate among all races across the year s Compared to Caucasian Asian (OR=0.73, p<.001) and African American (OR=0.91, p<.001) adolescents tended not to engage in UWCB, while adolescents of Hispanic (OR=1.08, p<.05), m ultiple race with Hispanic origin (OR=1.20, p<.001), and m ultiple race without Hispan i c origin (OR=1.30, p<.001) tended to engage in UWCB. Lastly, nonoverweight adolescents with BID showed the highest UWCB p revalence rate in all surveys Compared with nonoverweight without BID, nonoverweight with BID (OR=3.93, p<.001), overweight without BID (OR=2.81, p<.001), and overweight with BID (OR=1.62, p<.001) had higher risk engaging in UWCB. In addition, t he curve estimation over each category revealed similar results as the pooled UWCB and HWCB data where quadratic models could not be established A lthough data showed some fluctuation over time the linear time beta showed significant negative values in all

PAGE 47

47 subgroups from 1999 to 2009 except American Indians /Alaska Natives and m ultiple race with Hispanic origin. Such trends indicat ed the UWCB prevalence rates decreased across sex, grade, race/ethnicity over the years. However, the UWCB prevalence rates among American Indians /Alaska Native adolescents and adolescents with m ultiple race and Hispanic origin remain ed relatively high overall. HWCB trends by the demographic subgroups and BID categories are presented in the Table 23. Three subgroups consistently exh ibited higher or lower prevalence rates of HWCB from their counterparts across the year s First, more female adolescents engaged in HWCB than male students Females adolescents had great er likelihood to practice HWCB than male adolescents (OR=2.38, p<.001) In addition, African American (OR=0.69, p<.001) and Asian (OR=0.92, p<.05) adolescents practiced less HWCB, while adolescents in m ultiple race with Hispanic origin (OR=1.20, p<.001) and Hispanic (OR=1.11, p<.001) engaged in more HWCB than Caucasian s Las tly, a bout 90% of nonoverweight adolescents with BID engaged in HWCB which also represented the highest HWCB prevalence rate among other BID categories. In fact, nonoverweight adolescents with BID (OR=7.47, p< .001), overweight adolescents without BID (OR=6.09, p<.001), and overweight adolescents with BID (OR=1.90, p<.001) tended to practice HWCB more than nonoverweight adolescents without BID. When we examined the longitudinal trends, s everal subgroups including male, 10th grade, 11th grade, 12th grade, American Indian/Alaska Native, Caucasian, Hispanic, nonoverweight without BID, and overweight with BID exhibited a significant linear ascending trend of HWCB across years. Besides, m ore than half of the adolescents engaged in HWCB in all subgroups and the prevalence rates were found gradually increased over the years.

PAGE 48

48 D iscussion This investigation represents the first effort to examine the trends of weight control behaviors by year among adolescents. Our findi ngs were consistent with the high prevalence rates of UWCB (ranged from 9.4% to 21.2 %) in a p revious study using l ongitudinal surveys in Minnesota (D. Neumark Sztainer, Wall, Eisenberg, Story, & Hannan, 2006) Over time, our f indings suggest ed the prevalence trend s of UWCB decreased and trend s of HWCB increased across years. Our findings of ascending HWCB trends echoed the success of long term (2 to 10 years) healthy weight loss interventions (Batch, 2005; Epstein et al., 1994; Reinehr et al., 2009) and benefits (K. E. Hart & Warriner, 2005) among adolescents and implied that adolescents m igh t have been more aware of the benefit s of HWCB over UWCB. S ubgroup analyses also show ed the presence of similar trends of HWCB and UWCB among sex, ethnic groups, and BID status. Female adolescents and nonoverweight adolescents with BID exhibited higher prevalence rate of weight control behavior s in both healthy and unhealthy approaches across years when compared to their counterparts. P revious research results also support ed our findings that female adolescents exhibited higher percentage of engaging in w eight control behavior than male adolescents (43.7% v.s. 22.3%) (Serdula et al., 1999) Also female exhibited significant higher rate of fasting than male (6.3% v.s. 4.0%) (Striegel Moore et al., 2009) Only one p revious study investigated the influence of BID on adolescent in which the researchers suggested female adolescents with BID exhibited 4.3 times greater odds of onset of UWCB (Liechty, 2010) W ith the national representative data, this study confirm ed th is difference and further indicat ed the risk of UWCB for adolescents by the combination of overweight status and BID These comparisons may provide more scientific evidences for the design of intervention and the identification of risk population.

PAGE 49

49 Limitations Several important limitations were identified in this study. Although YRBS has been well recognized as a sound instrument for youth risk behaviors, the use of 3 and 2 items to assess UWCB and HWCB respectively may not have satisfactory reliability and validity. This same scenario may be applicable to the assessment of BID They may be considered as less reliable than a more sophisticated instrument with known psychometrics. In addition, the findings from this sample of public high school attendees may not be generalizable to other groups of adolescents, such as private high school students and individuals who did not attend or dropped out of high school. Future R esearch and Implications Based on these national ly representative surveys, the results of this study clearly demonstrate d the trends of weight control behaviors and possible link age between weight control behavior and gender race/ethnicity, and body image distortion among high school students We suggest the continuation of such large scale longitudinal studies to monitor the weight control behaviors Additional studies can be done to investigate the risk/preventive factors of UWCB and its associated risk behaviors. F urther studies to investigat e the underl ying psychosocial mechanisms of choosing UWCB over HWCB are recommend ed to provide further insight s for proper weight control interventions Despite the need s for additional research, findings of this s tudy have significant implications for prevention programming in a number of areas. First, UWCB is prevalent. As a result, early identification and prevention efforts in this area are needed and the development and implementation of prevention programs spe cific to adolescent weight control are recommended Second, UWCB is prevalent in some demographic subgroups Thus, those high risk groups including females, American Indians/Alaska Natives, and overweight and normal weight

PAGE 50

50 adolescents who perceived themse lves as overweight need to be targeted Last ly, practitioners working in obesity prevention are suggested to foster proper body image and develop programs using healthy weight control methods.

PAGE 51

51 Table 21 Descriptive statistics of participants demographics, BID, weight loss tendency, UWCB, and HWCB by year 1999 2001 2003 2005 2007 2009 Overall Characteristics n (%) n (%) n (%) n (%) n (%) n (%) n (%) Sex Female 7608 (49.6) 6953 (51.3) 7361 (48.6) 6858 (49.5) 6942 (49.4) 7816 (47.8) 43538 (49.3) Male 7727 (50.3) 6609 (48.7) 7787 (51.4) 7006 (50.5) 7084 (50.5) 8537 (52.2) 44750 (50.7) Grade 9 th 4424 (28.9) 4030 (29.7) 4448 (29.4) 4021 (29.0) 4052 (29.0) 4570 (28.0) 25546 (28.8) 10 th 3984 (26.0) 3509 (25.9) 3947 (26.1) 3591 (25.9) 3663 (26.2) 4273 (26.2) 22967 (25.9) 11 th 3607 (23.6) 3126 (23.1) 3532 (23.3) 3230 (23.3) 3265 (23.4) 3843 (23.5) 20602 (23.3) 12 th 3282 (21.4) 2873 (21.2) 3181 (21.0) 2997 (21.6) 2976 (21.3) 3628 (22.2) 18938 (21.4) Race/Ethnicity AI/AN 131 (0.9) 94 (0.7) 158 (1.0) 143 (1.0) 135 (1.0) 101 (0.6) 762 (0.9) Asian 456 (3.0) 454 (3.4) 507 (3.4) 460 (3.4) 480 (3.5) 546 (3.4) 2903 (3.3) Black 2141 (14.1) 1747 (13) 2095 (13.9) 2004 (14.6) 2076 (15.1) 2320 (14.4) 12382 (14.2) NH/PI 86 (0.6) 100 (0.7) 149 (1.0) 114 (0.8) 108 (0.8) 130 (0.8) 687 (0.8) White 9219 (60.8) 9093 (67.5) 9271 (61.4) 8482 (61.9) 8321 (60.3) 9452 (58.7) 53838 (60.8) Hispanic 1242 (8.2) 1388 (10.3) 2139 (14.2) 1310 (9.6) 1184 (8.6) 1766 (11.0) 9029 (10.2) Multiple H 328 (2.2) 221 (1.6) 364 (2.4) 759 (5.5) 1142 (8.3) 1223 (7.6) 4037 (4.6) Multiple NH 1555 (10.3) 367 (2.7) 423 (2.8) 419 (3.1) 345 (2.5) 572 (3.6) 3682 (4.2) BID 1 .N O W NBID 8681 (60.0) 7791 (62.1) 8133 (61.5) 7863 (60.2) 7861 (61.3) 9489 (62.9) 49819 (61.3) 2 .N O W BID 1797 (12.4) 1527 (12.2) 1536 (11.6) 1446 (11.1) 1285 (10.0) 1412 (9.4) 9002 (11.1) 3 .OW BID 1414 (9.8) 1106 (8.8) 1108 (8.4) 1099 (8.4) 1208 (9.4) 1431 (9.5) 7366 (9.1) 4 .OW NBID 2573 (17.8) 2126 (16.9) 2440 (18.5) 2652 (20.3) 2479 (19.3) 2748 (18.2) 15018 (18.5) Trying to l ose w eight Yes 6527 (42.7) 6219 (46.0) 6554 (43.8) 6264 (45.6) 6209 (45.2) 7157 (44.4) 38930 (44.6) Weight control behavior HWCB 9548 (63.2) 8674 (65.2) 9461 (64.8) 9003 (65.8) 9095 (66.8) 10662 (66.5) 56442 (65.4) UWCB 2753 (18.2) 2570 (19.4) 2807 (19.4) 2216 (16.6) 2081 (15.8) 2277 (14.2) 14704 (17.2) BID 1: Nonoverweight students without BID. 2: Nonoverweight students with BID. 3: Overweight students with BID. 4: Overweight students without BID.

PAGE 52

52 Table 22 UWCB prevalence rates by year and participants demographics and BID and their bivariate odds ratios Characteristics 1999 % 2001 % 2003 % 2005 % 2007 % 2009 % Total sample 1999 2009 % OR 95% CI Sex Female 26.1 26.4 25.9 22.8 21.3 19.1 .008** 9953 23.6 2.51 2.42, 2.61 Male 10.4 12.0 13.1 10.5 10.3 9.6 .002** 4704 10.9 Reference Grade 9 th 17.8 20.6 18.9 16.1 14.1 13.5 .006** 4099 16.8 Reference 10 th 18.9 19.9 18.3 16.1 16.8 13.7 .005** 3820 17.2 1.03 0.98, 1.08 11 th 17.2 17.5 20.5 17.5 15.8 15.4 .002** 3456 17.3 1.03 0.98, 1.09 12 th 19.0 18.8 19.7 16.8 16.5 14.3 .005** 3208 17.4 1.04 0.99, 1.10 Race/Ethnicity AI/AN 20.6 26.6 29.6 16.9 22.7 26.7 .001 173 23.6 1.51 1.27, 1.79 Asian 13.6 12.0 22.1 11.3 13.4 6.3 .007** 350 13.0 0.73 0.65, 0.82 AA 18.4 17.4 16.5 15.0 13.5 13.8 .005** 1833 15.7 0.91 0.86, 0.96 NH/PI 17.9 17.4 22.8 20.9 14.6 15.9 .003 117 18.2 1.09 0.89, 1.33 White 18.0 19.3 18.8 17.1 15.2 13.8 .005** 8941 17.0 Reference Hispanic 18.3 22.8 21.4 15.5 15.8 13.8 .008** 1564 18.1 1.08 1.02, 1.14 Multiple H 15.0 26.3 18.6 19.0 21.5 19.4 .002 762 19.8 1.20 1.11, 1.31 Multiple NH 21.2 21.8 27.2 14.4 22.8 19.5 .002 755 21.1 1.30 1.20, 1.41 BID 1.NW NBID 12.0 12.8 12.1 10.9 10.4 8.9 .003** 5431 11.1 Reference 2.NW BID 35.8 35.6 36.1 30.7 27.9 29.7 .008** 2901 33.0 3.93 3.73, 4.14 3.OW BID 15.8 20.1 19.9 15.2 16.5 14.7 .003* 1205 16.8 1.62 1.51, 1.73 4.OW NBID 28.0 30.3 27.4 25.1 25.2 21.4 .007** 3842 26.0 2.81 2.69, 2.95 Note: *: p<.05; **: p<.01

PAGE 53

53 Table 23 H WCB prevalence rates by year and participants demographics and BID and their bivariate odds ratios Characteristics 1999 % 2001 % 2003 % 2005 % 2007 % 2009 % Total sample 1999 2009 % OR 95% CI Sex Female 73.8 75.8 75.7 75.5 75.2 74.9 .001 32004 75.1 2. 38 2. 31 2. 45 Male 52.7 54.2 54.5 56.4 58.6 58.7 .006** 24301 55.9 Reference Grade 9 th 64.7 68.5 64.6 67.1 68.9 66.3 .002 16454 66.6 Reference 10 th 63.0 65.7 66.4 65.8 66.8 66.1 .003* 14742 65.6 1.03 0.98, 1.08 11 th 62.4 61.9 65.3 64.9 66.6 65.7 .004** 12975 64.5 1.03 0.98, 1.09 12 th 62.4 63.9 63.2 65.2 64.5 67.8 .005** 11996 64.6 1.04 0.99, 1.10 Race/Ethnicity AI/AN 49.6 67.0 66.7 65.9 71.4 64.6 .014* 470 64.2 0.92 0.79, 1.07 Asian 63.4 58.6 67.1 64.4 65.6 66.3 .005 1815 64.3 0.92 0.85, 0.99 AA 58.4 56.0 55.1 59.2 57.3 57.8 .001 6819 57.4 0.69 0.66, 0.71 NH/PI 58.3 65.6 75.7 66.7 68.6 71.7 .008 450 68.6 1.11 0.94, 1.31 White 63.4 66.9 66.1 66.2 68.1 66.9 .003** 34951 66.2 Reference Hispanic 66.4 67.3 67.4 70.5 68.8 70.5 .004* 6011 68.5 1. 11 1.06, 1.16 Multiple H 63.4 72.8 69.3 72.2 70.4 70.0 .003 2755 70.1 1.20 1.11, 1.28 Multiple NH 68.5 66.2 62.5 64.0 71.6 69.3 .001 2434 67.5 1. 06 0.98, 1.14 BID 1.NW NBID 51.4 53.9 54.5 53.7 56.4 56.6 .005** 26846 54.4 Reference 2.NW BID 89.6 91.2 90.7 89.3 88.7 89.9 .001 8017 89.9 7.47 6.95, 8.02 3.OW BID 63.7 68.1 69.1 70.7 74.6 70.8 .008** 5032 69.4 1.90 1.80, 2.00 4.OW NBID 87.7 88.3 86.8 88.0 88.7 88.0 .001 13081 87.9 6.09 5.78, 6.42 Note: *: p<.05; **: p<.01

PAGE 54

54 CHAPTER 3 RISK FACTORS OF UNHE ALTHY WEIGHT CONTROL BEHAVIORS AMONG HIGH SCHOOL STUDENTS Background Weight control behaviors are common among adolescents. According to the 2009 Youth Risk Behavioral Surveillance System (YRBSS), 59 .3 % of female and 3 0.5% of male students in grades 9 12 reported trying to lose weight (Eaton et al., 2010) Unhealthy Weight Control Behaviors (UWCB), such as fasting, using diet pills/powders/liquids, vomiting, or taking laxatives (Eisenberg et al., 2005; Haines et al., 2006; D. NeumarkSztainer et al., 2003) we re found common among adolescents Recent YRBSS indicated that in 2009, 1 0.6% of high school students did not eat for 24 or more hours, 5 .0 % took diet pills, powders, or liquids without a doctors advice, and 4 .0 % vomited or took laxatives to lose weight or to keep from gaining weight during the 30 days before the survey (Eaton et al., 2010) UWCB predicts th e onset of more severe disordered eating behaviors and eating disorders, including anorexia nervosa, bulimia nervosa, and binge eating disorder, as well as depression and other medical and psychosocial morbidity (Fonseca et al., 2009) Contrary to the projected weight loss, particularly during adolescence, UWCB has been shown to predict weight gain over time (Fonseca et al., 2009) ; (D. N eumark Sztainer et al., 2009) In addition, adolescents who had UWCB have approximately 3 times greater risk for being overweight and 6 times for binge eating with loss of control 5 years later (D. Neumark Sz tainer et al., 2006) Besides, UWCB can impact adolescents psychological health, particularly during a time of rapid physical, psychological, and social development (Daee et al., 2002) Adolescents who were engaged in UWCB exhibited the most negative patterns of psychosocial and health behaviors, such as alcohol or marijuana use, school delinquency, unprotected sex, and suicide attempts (D. Neumark Sztainer et al., 1996) Other health consequences caused by UWCB include fatigue,

PAGE 55

55 dizziness, lack of energy, sore throat, abdominal pain, constipation, loss of dental enamel, and pancreatitis (Pritts & Susman, 2003) These medical compli cations may be irreversible and may simultaneously cause growth retardation, structural brain changes, and pubertal delay or arrest (Golden et al., 2003) Recent studies have suggested several significant predictors of UWCB, including sex (K. Boutelle et al., 2002; Fonseca et al., 2009; Lowry et al., 2002; McCab e & Ricciardelli, 2009; D. Neumark Sztainer et al., 2006; Striegel Moore et al., 2009) race/ethnicity (Haley et al., 2010; Silverman et al., 2001) overweight status (K. Boutelle et al., 2002; Foti & Lowry, 2010) sedentary behaviors, physical activities (moderateto vigorous physical activity) dietary behaviors (fruit/vegetable intake, fast food, and sugar sweetened beverages) (D. R. Neumark Sztainer et al., 2007) and Body Image Distortion (BID), which refers to a discrepancy between perceived and actual weight status (Liechty, 2010) Howeve r, these aforementioned studies have a similar limitation the lack of sample representativeness and generalizability. In addition, these aforementioned predictors have not been examined on the same horizon in one same sample, without mentioning the fact that race/ethnicity and vegetable and fruit consumption are seldom linked with UWCB. Thus, i mportant pieces of the puzzle were never made into one whole picture. Hence, t he purpose of this study was to investigate the predictive factors of UWCB among Ameri can adolescent s using recent national representative data. Results can help identify high risk populations and provide upto date information in designing tailored prevention programs specifically targeting adolescents. M ethods T he Division of Adolescent a nd School Health at the CDC established the nations biennial epidemiologic surveillance system in 1991 t o monitor health risk behaviors among youth and

PAGE 56

56 young adults. The Youth Risk Behavior Surveillance System (YRBSS) monitors the prevalence of obesity an d asthma as well as the six categories of priority high risk behaviors, including behaviors that contribute to unintentional injuries and violence, tobacco use, alcohol and other drug use, sexual behaviors, unhealthy dietary behaviors, and physical inactivity. The YRBSS employed a threestage cluster sampling design to surveys students in grades 9 12 from nationally sampled sc hools across the United States and applies a weighting system, thus providing the most representative data for youth health practices The Youth Risk Behavior Survey (YRBS) has been give n biennially since 1999. O fficial brief descriptive statistics were reported in the Morbidity and Mortality Weekly Report in the year following the survey and YRBSS website The data collected by the YR BSS is an important source of information in the research of weight control behaviors and was used in this study as a basis to identify the risk factor s of UWCB For the 2009 national YRBS, the school response rate was 81%; the student response rate was 88 %; the overall response rate was 71%. Sample Demographics The 2009 survey included 16, 410 participants; 47.8% were females. Grades were fairly evenly distributed with slightly more (28%) 9th grade rs and less (22.2%) 12th grade rs Most students were Caucasian (58.7%), followed by Hispanic and Multiple/Hispanic (18.6%), African American (14.4%), Multiple/non Hispanic (3.6%), Asian (3.4%), Native Hawaiian/Other Pacific Islander (0.8%), and American Indian/Alaska Native (0.6%). Measures This study exami ned the UWCB predictors which were selected based on the available YRBS items that are demographic s psychological ly related to UWCB and lifestyle behaviors that directly related to energy intake/expenditure. The candidate predictors include sex, grade, r ace/ethnicity, BID, current weight control tendency, sedentary lifestyle, participation in

PAGE 57

57 moderate physical activity, participation in vigorous physical activity, participation in daily physical education class, vegetable, fruit and sugar sweetened bever age consumption. As described in the 2009 YRBS methodology, UWCB was assessed based on participants self identification from at least one of the following 3 questions: During the past 30 days, did you go without eating for 24 hours or more (also called fasting) to lose weight or to keep from gaining weight? (Response options: Yes, No), During the past 30 days, did you take a ny diet pills, powders, or liquids without a doctors advice to lose weight or to keep from gaining weight? (Do not include meal replacement products such as Slim Fast.) (Response options: Yes, No), and During the past 30 days, did you vomit or take laxa tives to lose weight or to keep from gaining weight? (Response options: Yes, No). In this study, o verweight was defined as a Body Mass Index ( BMI ) being equal to or larger than the 85th percentile of the age (Kuczmarski et al., 2002) P articipant s who had their BMI below their corresponding 85th percentiles w ere classified as non overweight. As presented in the literature, perception about weight seems to play a more significant role in weight control. Foll owing this track, Body Image Distortion (BID) was selected to predict UWCB. BID is the variable generated by comparing the perceived body weight and actual weight of each participant I n YRBS, p articipants were asked How do you describe your weight? The response option s include Very underweight Slightly underweight About the right weight Slightly overweight Very overweight The responses in the first three options were pool ed as perceived nonoverweight and the last two were considered perceived overweight. The body weight perception was cross tabulated with the respondents overweight status based on their actual BMI percentiles The participants whose BMI was equal to or greater than the 85th percentile and perceived themselves as over weight and the ones whose BMI was less than the

PAGE 58

58 85th percentile and perceived themselves as nonoverweight were defined as no BID. The participants whose BMI was equal to or larger than the 85th percentile and perceived themselves as nonoverweight and the ones whose BMI was less than the 85th percentile and perceived themselves as overweight were defined as having BID. Consequently four weight BID categories were tabulated in this study : nonoverweight without BID, nonoverweight with BID, overweight withou t BID, and overweight with BID. Current weight control tendency was assessed by asking the following question: Which of the following are you trying to do about your weight? (Response options: lose weight, gain weight, stay the same weight, or I am not trying to do anything about my weight). P articipants who checked the lose weight option were classified as trying to lose weight while others were pooled as not trying to lose weight. Sedentary behavior s were assessed by the sum of more than 2 hours pe r day of TV viewing time, video or computer games playing, or computer using for something that is not school work on an average school day (D. R. Neumark Sztaine r et al., 2007; US Dept of Health and Human Services, 2000) This measure was created by adding the responses from these 2 questions: On an average school day, how many hours do you watch TV? (Response options: I do not watch TV on an average school day, less than 1 hour per day, 1 hour per day, 2 hours per day, 3 hours per day, 4 hours per day, 5 or more hours per day), and On an average school day, how many hours do you play video or computer games or use a computer for something that is not school work? (Include activities such as Nintendo, Game Boy, PlayStation, Xbox, compute r games, and the internet. (Response options: I do not play video or computer games or use a computer for something that is not school work, less than 1 hour per day, 1 hour per day, 2 hours per day, 3 hours per day, 4 hours per day, 5 or more hours per day).

PAGE 59

59 Regular participation in moderate physical activity was assessed by asking the following question: On how many of the past 7 days did you participate in physical activity for at least 30 minutes that did not make you sweat or breathe hard, such as f ast walking, slow bicycling, skating, pushing a lawn mower, or mopping floors? (Response options are 0 to 7 days). The participation in vigorous physical activity was assessed by asking the question On how many of the past 7 days did you exercise or part icipate in physical activity for at least 20 minutes that made you sweat and breathe hard, such as basketball, soccer, running, swimming, laps, fast bicycling, fast dancing, or similar aerobic activities? (Response options are 0 to 7 days). Participation in daily school PE was assessed by asking, In an average week when you are in school, on how many days do you go to physical education (PE) classes? (Response options are 0 to 5 days). The participation in moderate and vigorous physical activity, and da ily PE class were analyzed as dichotomous variables according to their corresponding national health objectives and recommendations. These cut off points include 30 minute moderate physical activity in 5 or more days per week, 20 minute vigorous physical a ctivity in 3 or more days per week, and attendance in PE class 5 days per week (US Dept of Health and Human Services, 2000) The 2009 YRBS data contain 2 variables QNVEG and QNFRUIT, to dete rmine if participants met the d ietary recommendations by American Heart Association for children and adolescents, which suggested at least 3 times of vegetables and 2 time s of fruit per day for age s 14 to 18 years (Gid ding et al., 2005) QNVEG was assessed by combining 4 questions: During the past 7 days, how many times did you eat green salad?, During the past 7 days, how many times did you eat potatoes? (Do not count french fries, fried potatoes, or potato chips .), During the past 7 days, how many times did you eat carrots?, and During the past 7 days, how many

PAGE 60

60 times did you eat other vegetables? (Do not count green salad, potatoes, or carrots.). (Response options include I did not eat vegetables during the past 7 days, 1 to 3 times during the past 7 days, 4 to 6 times during the past 7 days, 1 time per day, 2 times per day, 3 times per day, 4 or more times per day). QNFRUIT was assessed by combining 2 questions: During the past 7 days, how many times did yo u drink 100% fruit juices such as orange juice, apple juice, or grape juice? (Do not count punch, Kool Aid, sports drinks, or other fruit flavored drinks.) and During the past 7 days, how many times did you eat fruit? (Do not count fruit juice.) (Respon se options include I did not eat fruit/drink fruit juice during the past 7 days, 1 to 3 times during the past 7 days, 4 to 6 times during the past 7 days, 1 time per day, 2 times per day, 3 times per day, 4 or more times per day). The consumption of s ugar sweetened beverage s w as assessed by asking: During the past 7 days, how many times did you drink a can, bottle, or glass of soda or pop, such as Coke, Pepsi, or Sprite? (Do not include diet soda or diet pop) (Response options include I did not soda or pop during the past 7 days, 1 to 3 times during the past 7 days, 4 to 6 times during the past 7 days, 1 time per day, 2 times per day, 3 times per day, 4 or more times per day) In 19992004, US youth ages 2 19 years, consumed an average of 224 kcal per capita per day from sugar sweetened beverages (Y. C. Wang, Bleich, & Gortmaker, 2008) A can, bottle, or glass of soda or pop was around the average. Therefore, the cut off point was 1 time per day. Data Analysis IB M SPSS Statistics 19.0 software was used to conduct all analyses to account for the complex sampling design and weighting of the data. Per YRBS instructions, the weighting factor was applied in the data analysis to ensure national representation. Statistical analysis was performed over the YRBS 2009 data downloaded from YRBSS website. To assure the quality of

PAGE 61

61 analysis, these descriptive statistics were fully compared against the Youth Online results, the official YRBSS statistics. The outcome variable of th is study, UWCB is in a dichotomous categorical format Considering the attribute of this outcome variable, logistic regression analysis was identified as the most rational statistical method. Stepwise m ultiple logistic regression with backward elimination allows all the candidate predictors placed into a regression model and deletes insignificant variables step by step from the model until reaching a point where the remaining variables make significant partial contribution in predicting the outcome variabl e (Agresti & Finlay, 1997) Hence, a fter the backward elimination the predictive variables can be identified. To determine the interaction, two way interactions of predictive variables were added to the model to test the significance. Similar backward elimination was applied to all the two way interaction terms The f inal model was determined by Akaike Information Criterion (AIC), which judges the model by how close its fitted values tend to be to the true expected values (Agresti, 2007) R esults Table 31 presented the sample description and the distribution of UWCB by the candidate predictors. In 2009, the prevalence of trying to lose weight was 44.4% and prevalence of engaging in at least one UWCB was 14.2% among high school students. About 57.4 % of adolescents had sedentary behaviors while 28.7% participated in moderate physical activity, 33.3% participated in daily school physical education, and 67.7% participated in vigorous physical activity. Regarding dietary behavior s only 33.9% of adolescents consume d the recommended fruit, and 1 3.8% consumed enough vegetables while 29.2% dr a nk sugar sweetened beverage s daily. Nine percent of adolescents wer e nonoverweight with BID, 10 percent were overweight with BID. The UWCB distribution was similar in 5 variable s, i.e.,

PAGE 62

62 grade, moderate physical activity, daily physical education, sedentary lifestyle, and fruit consumption. The other 7 variables exhibited statistical significance Table 32 presented the seven variables identified after stepwise multiple logistic regression analysis T wenty one possible two way interaction terms were entered into the second block of the regression model. The backward elimination removed 13 no n significant two way interaction terms. Although either model seemed reasonable, the complex model with 8 interaction terms only reduced the AIC value by 1.5% (9468 to 9318). Therefore, the model with seven main effects without in teractions was adopted The final logistic regression model revealed that after control other significant variables, the strongest predictive factor for UWCB was current weight lose tendency Adolescents who were trying to lose weight demonstrated approximately 4. 7 times significantly higher risk to engag e in UWCB in reference to those who were not (OR= 4.66, 95% C.I.: 4.06 5.35). In addition, compared with adolescents who were nonoverweight without BID, all other weight perception groups exhibited signi ficantly elevated risk to engag e in UWCB : adolescents who were non overweight and had BID (OR= 1.98, 95% C.I.: 1.692.32), overweight with BID (OR= 1.44, 95% C.I.: 1.201.73), and overweight without BID (OR= 1.38, 95% C.I.: 1.201.58). Like most of the rel ated research, our findings suggested female adolescents tend ed to engage more in UWCB compared to male peers (OR=1.62, 95% C.I.: 1.451.82). On the contrary, another demographic factor, race/ethnicity has not been well investigated in the literature. Th e results of the present study can bridge this gap. Several significant differences among races /ethnicity were uncovered C ompared with Caucasian adolescents the American Indian/Alaska Native adolescents (OR=2.63, 95% C.I.: 1.564.45) adolescents of multiple race/ethnicity with n on Hispanic origin (OR= 1.55, 95% C.I.: 1.192.02), and adolescents of

PAGE 63

63 multiple race/ethnicity with Hispanic origin (OR= 1.31, 95% C.I.: 1.081.59) exhibited higher odds of engaging in UWCB while Asian adolescents (OR=0.35, 95 % C.I.: 0.22 0.55) tend ed not to engage in this behavior. Among the candidate predictors that increase energy intake, vegetable and sugar sweetened beverage consumption increased the risk of UWCB. Adolescents who ate recommended or greater amounts of vege table s exhibited significant higher odds of engaging in UWCB compared to those who did not (OR=1.91, 95% C.I.: 1.662.18). In addition, a dolescents who drank 1 or more sugar sweetened beverage s per day tend ed to engage in UWCB compared to those who repor ted less than 1 (OR=1.59, 95% C.I.:1.421.78). When the candidate predictors that increase energy expenditure were examined only the lack of participation in vigorous physical activity appeared statistical ly significant. The results showed the a dolescent s who did not participate in vigorous physical activity tended to engage in UWCB, compared to those who participated 20 minutes of vigorous physical activity 3 or more days per week (OR=1.14, 95% C.I.:1.02 1.27). Three m ultiple logistic analysis w ere con ducted for each of the UWCBs. The results showed the same significant predictors and similar strengths of association in odds ratios. Thus, the significant predictors of each UWCB and their odds ratios can be considered approximately equivalent. D iscussion This study aimed to explore the risk factors of UWCB among adolescents using the nationally representative YRBS data. Our findings suggested UWCB is moderately prevalent among female adolescents. This result is consistent from agreement among the existin g studies that female adolescents are at higher risk for engaging in UWCB (K. Boutelle et al., 20 02; Fonseca et al., 2009; Lowry et al., 2002; McCabe & Ricciardelli, 2009; D. NeumarkSztainer et

PAGE 64

64 al., 2006; Striegel Moore et al., 2009) Besides, our findings confirmed Haley s findings in the relationship s between UWCB and race/ethnicity in which Ame rican Indian/Alaska Native adolescents were at the highest risk and Asian adolescents were at the lowest risk (Haley et al., 2010) Based on the strength of this study to analyze the nationally representative data, ou r findings also expanded the generalizability to the national level. However, the latent effect s behind these associations, such as social and culture influences are suggested for future research. Even though our findings revealed weight control attempts w ere quite prevalent among adolescents, the prevalence of obesity is still increasing in the United States (Y. Wang & Beydoun, 2007) Nearly a half of the adolescents in the US were trying to lose weight. Among those who attempted to lose weight, approximately a quarter engaged in UWCB. Moreover, this study confirmed that Body Image Distortion increases the risk of UWCB (Liechty, 2010) and further demonstrated that the risk of UWCB elevates particularly among the nonoverweight adolescents with BID Yet the underlying factors of BID have not been explored in current literature. F urther studies investigating the mechani sm of how BID interacts with UWCB to provide insight into the foundation of UWCB prevention programs will be helpful. Findings in this study also verified that vigorous physical activity was a strong pr otect ive factor against UWCB (D. R. Neumark Sztainer et al., 2007) while daily sugar sweetened beverage consumption was identified as a risk factor Although frequent vegetable consumption was suggested as the key element of a healthy weight loss progr am (C. N. Hart et al., 2010) our findings demonstrated adolescents who consumed recommended amounts of vegetable (3 times/day) tend ed to engage in UWCB. Whether this discrepancy reflects a lack of knowledge about weig ht control methods or an inability to implement the healthy method effectively cannot be answered in this study. Further studies in monitoring the weight control behaviors, weight

PAGE 65

65 control knowledge dissemination, and obstacles to healthy weight control beh aviors are necessary to provide insight for weight control program planning specifically targeting adolescents. Despite the need of further research, findings of this study have significant implications for prevention/intervention programming with adolesce nts in a number of important areas. First, UWCB is prevalent in some populations. As a result, early identification of eating disorders is recommended as it leads to more successful treatment outcomes (Bulik et al., 2007) Second, b ased on our finding s it is rational to believe a b rief assessment of weight status and BID can help identify adolescents at risk By strengthen ing the prevention efforts to such an at risk population, the lower risk of UWCB can be projected Finally, a combination of encouraging vigorous physical activity and decrease of sugar sweetened beverage consumption can yield great benefits in promoting healthy weight control behaviors and preventing the unhealthy weight control behaviors.

PAGE 66

66 Table 34 UWCB distribution by the candidate predictors Characteristics T otal UWCB Frequency Percent Frequency Percent All participants 16035 100 2277 14.2 Sex Female 7816 47.8 1464 19.1 Male 8537 52.2 799 9.6 Grade 9 th grade 4570 28.0 595 13.5 10 th grade 4273 26.2 577 13.7 11 th grade 3843 23.5 578 15.4 12 th grade 3628 22.2 507 14.3 Race/Ethnicity American Indian/Alaska Native 101 0.6 27 26.7 Asian 546 3.4 33 6.3 Black 2320 14.4 310 13.8 Native Hawaiian/Other Pacific Islander 130 0.8 20 15.9 Hispanic 1766 11.0 238 13.8 Multiple Hispanic 1223 7.6 232 19.4 Multiple Non Hispanic 572 3.6 108 19.5 White 9452 58.7 1274 13.8 BID Nonoverweight without BID 9413 62.4 822 8.8 Nonoverweight with BID 1356 9.0 400 29.8 Overweight with BID 1507 10.0 223 15.0 Overweight without BID 2804 18.6 599 21.5 Current Weight Control Tendency To lose weight 7157 44.4 1757 24.7 Not to lose weight 8976 55.6 516 5.8 Vegetable C onsumption Above recommendation 2239 13.8 456 20.7 Below recommendation 13980 86.2 1803 13.1 Fruit C onsumption Above recommendation 5486 33.9 818 15.2 Below recommendation 10703 66.1 1432 13.6 Sugar sweetened Beverage Daily Yes 4736 29.2 782 16.8 No 11503 70.8 1479 13.1 Sedentary L ifestyle Yes 9270 57.4 1344 14.8 No 6868 42.6 906 13.4 Vigorous Physical Activity Yes 10285 67.7 1322 13.1 No 4916 32.3 798 16.5 Moderate Physical Activity

PAGE 67

67 Yes 4296 28.7 597 14.0 No 10696 71.3 1521 14.3 Daily School P hysical Education Yes 5239 33.3 674 12.9 No 10514 66.7 1548 14.8 Table 35 Results of stepwise multiple logistic regression Characteristics O dds R atio 95% C onfidence I nterval Sex Female 1.62*** (1.45, 1.82) Male Reference Race/Ethnicity American Indian/Alaska Native 2.63** (1.56, 4.45) Asian 0.35*** (0.22, 0.55) Black 1.03 (0.88, 1.21) Native Hawaiian/Other Pacific Islander 1.17 (0.63, 2.17) Hispanic 0.86 (0.72, 1.02) Multiple Hispanic 1.31** (1.08, 1.59) Multiple Non Hispanic 1.55*** (1.19, 2.02) White Reference BID Nonoverweight with BID 1.98*** (1.69, 2.32) Overweight with BID 1.44*** (1.20, 1.73) Overweight without BID 1.38*** (1.20, 1.58) N onoverweight without BID Reference Current Weight Control Tendency To l ose w eight 4.66*** (4.06, 5.35) Not to lose weight Reference Vegetable Consumption Above recommendation 1.91*** (1.66, 2.18) Below recommendation Reference Consume Sugar sweetened Beverage Daily Yes 1.59*** (1.42, 1.78) No Reference Vigorous Physical Activity No 1.14* (1.02, 1.27) Yes Reference Note: *: p<.05, **: p<.01, ***: p<.001

PAGE 68

68 CHAPTER 4 THE RELATIONSHIP BETWEEN UNHEALTHY WEIGHT CONTROL BEHAVIORS AND SELECTED HEALTH RISK BEHAVIORS AMONG HIGH SCHOOL STUDENTS Background Unhealthy Weight Control Behaviors (UWCB) such as fasting, using diet pills/powders/liquids, vomiting, or taking laxatives are prevalent among U.S. adolescents (Eisenberg et al., 2005; Haines et al., 2006; D. NeumarkSztainer et al., 2003) A r ecent Youth Risk Behavior Survey (YRBS) sugges ted, in 2009, 10.6% of high school students did not eat for 24 or more hours ; 5% had taken diet pills, powders, or li quids without a doctors advice ; and 4% had vomited or taken laxatives to lose weight or to keep from gaining weight during the 30 days bef ore the survey (Eaton et al., 2010) Although UWCB did not project immediate life threat en ing dangers, but studies have suggested that UWCB can lead to more complicated ailments such as eating disorders, anorexia nervo sa, and bulimia nervosa (Daee et al., 2002; Patton et al., 1999) as well as other health compromising behaviors including suicide attempts, tobacco, alcohol, and marijuana use, unprotected sexual i ntercourse, and multiple sexual partners (Haley et al., 2010; D. NeumarkSztainer et al., 1996) To further explore the relationship between each health risk behavior, we conducted a comp rehensive literature review and summarized the findings by the risk behaviors below. Researchers who examined UWCB and its association with substance use among adolescents and found that alcohol, tobacco, and other drug use (ATOD) w as associated with adol escents UWCB. (Cance et al., 2005; Croll et al., 2002; French et al., 2001; Haley et al., 2010) Studies found a lcohol users had about 2.5 times higher ri sk than nonus ers. In addition to alcohol, tobacco use also was associated with UWCB among adolescents. Researchers found that some students smoked cigarettes as a method of weight control (Johnson et al., 2009) especially

PAGE 69

69 female adolescents (Lowry et al., 2002) The literature also show ed an increased likelihood of UWCB among adolescent smokers. In terms of risk adolescent smokers were significantly more likely to engage in U WCB (Strauss & Mir, 2001) and adolescent smokers odds ratio was approximately 2 times higher than non smokers (Johnson et al., 2009) In addition female adolescents exhibi ted a higher association between UWCB and ATOD use than male peers (Croll et al., 2002) In regard to the relationship between UWCB and other substance misuse among adolescents, research also indicated that those who engaged in UWCB reported more health compromising behaviors such as substance use in the past year (McGuire et al., 2002) A study analyzing 2003 YR B S data, which included illicit drug use, suggested juvenile delinquents were more likely than nondeli nquents to engage in UWCB (Ho et al., 2006) In addition, adolescent females who used laxatives or vomited to lose weight in the past year were more likely than those who did not do so to use substances including ecstasy, inhalants, nonmedical psychotherapeutics, marijuana, cigarettes, and alcohol during the past year (Cance et al., 2005) The same study also reported th at girls who used illicit drugs or tobacco were 4 times more likely to engage in UWCB than nonusers. Among both sexes, UWCB was associated significantly with substance use including illicit drug, tobacco, and binge drinking (Haley et al., 2010) Conversely, from the prevention standpoint, adolescents who reported abstinence from alcohol or drug s were less likely to report purging (French et al., 2001) Suicide represents the thir d leading cause of death among youth aged 15 19 years in 2010. A prior suicide attempt was identified as one of the most significant risk factors for a fatal adolescent suicide case (Centers for Disease Control and Prevention (CDC), 2010b) Similarly, adolescents who engaged in UWCB were more likely to report a history of suicide attempts in

PAGE 70

70 the past year (McGuire et al., 2002) Furthermore, adolescents engaging in multiple UWCB were at greater risk of having suicidal thoughts (Kim et al., 2009) In terms of risk, adolescents with suicidal ideation exhibited more than a 3 times higher risk of engaging in U WCB when compared to their counterparts (Haley et al., 2010) As mentioned previously, female adolescents were more likely to engage in UWCB. One study revealed that female adolescents who engaged in UWCB exhibit ed 4. 2 times higher odds of suicidal thoughts and 3.4 times higher odds of suicide attempts than their counterparts (Cook et al., 2007) R elationships between UWCB and sex related behaviors also have been reported. For exa mple, researchers suggested that female adolescents who engaged in UWCB were more likely to report being sexually active (McGuire et al., 2002) The study indicated that adolescent girls who engaged in UWCB exhibite d a 1.6 times higher odds of ever had vaginal intercourse compared to the girls who did not engage in UWCB (Cook et al., 2007) When examin ing relationships between UWCB and sexual behaviors from the prevention perspectives, researchers suggested that values related to abstinence from alcohol and sex appear to be protective factors against unhealthy eating behaviors and may reflect a general res ilience that buffers against a broad range of health risk behaviors (French et al., 2001) S tudies also showed that physical and sexual dating violence against adolescen ts was associ ated with a 2 to 5 times increased risk of UWCB (Haley et al., 2010; Silverman et al., 2001) Although no literature directly indicated bullying victimization wa s associated with UWCB, bullying in schools wa s associated with psychological distress and poor physical health (Rigby, 2003) Thus, a potential linkage between bullying and UWCB has yet to be explored. The preceding literature review highlight s the potential linkages between UWCB and the selected health risk behaviors. However, most of the investigations were based on researcher -

PAGE 71

71 initiated studies with relatively smaller sample sizes. In addition, most of the aforementioned associations relied on bivariate statistics. Collectively, t hese findings suggested paired associations between UWCB and the selected risk behaviors among adolescents and strength of association varied from study to study. Thus, more timely, in depth, and inclusive investigations with multivariate statistical an alysis are warranted. As health behavior researcher s suggested, practitioners and researchers should attempt to address multiple behaviors, not merely one health practice at a time, and consider contextual and environmental factors necessary to understand the broader context and more effectively influence health behaviors (Werch et al., 2009) Therefore, this study aimed to elucidate the association s between UWCB and he alth risk behaviors based on national ly representative participants. T he results can help provide discerning information for planning prevention and intervention programs. M ethods Study Design Data for this study were obtained from the national YRBS administered in 2009. C onducted every 2 years YRBS track s the prevalence of health risk behaviors that can lead to morbidity and mortality among US high school students. A three stage cluster sampling method and a weighting system were applied to represent all students attending high school in the country In the data set, several variables were assessed by more than one item. A ll results presented in this study are based on analyses of weighted data. A detailed description of survey methodology is available elsewhere ( Brener et al., 2004) Measures Variables examined in this study were originated from 25 YRBS items and recoded according to YRB S definitions (Eaton et al., 2010) UWCB was assessed by participant s Yes

PAGE 72

72 respons e to at least one of the following 3 questions: During the past 30 days, did you go without eating for 24 hours or more (also called fasting) to lose weight or to keep from gaining weight? (Response options: Yes, No) ; During the past 30 days, did you ta ke any diet pills, powders, or liquids without a doctors advice to lose weight or to keep from gaining weight? (Do not include meal replacement products such as Slim Fast .) (Response options: Yes, No) ; and During the past 30 days, did you vomit or take laxatives to lose weight or to keep from gaining weight? (Response options: Yes, No). Hence, UWCB is considered a dichotomous variable in this study Alcohol use in the past 30 days was assessed by 2 questions. Adolescents who responded with at least 1 day to the question During the past 30 days, on how many days did you have 5 or more drinks of alcohol in a row, that is, within a couple of hours? were classified as binge drink ing S tudents who were not binge drinking but responded at least 1 day to the question During the past 30 days, on how many days did you have at least one drink of alcohol? (Response options: 0 day s 1 or 2 days, 3 to 5 days, 6 to 9 days, 10 to 19 days, 2039 days, all 30 days) we re classified as casual drink ing The student s who did not drink in the past 30 days were coded as abstinence Cigarette smoking w as assessed by the following question: During the past 30 days, on how many days did you smoke cigarettes? (Response options: 0 day s 1 or 2 days, 3 to 5 days, 6 to 9 days, 10 to 19 days, 2029 days, all 30 days). The adolescents who responded 0 day were classified as non smokers. The s tudents who respond 20 or more days were classified as frequent smokers. The rest were classified as occasional smokers. Drug use was c oded into 3 levels: never used drugs, used drugs beyond 30 days, and used drugs in the past 30 days. The participants who responded one or more times in either one of the

PAGE 73

73 following two questions were classified as used drug in 30 days. During the past 30 days, how many times did you use marijuana? and During the past 30 days, how many times did you use any form of cocaine, including powder, crack, or freebase? (Response options: 0 times, 1 or 2 times, 3 to 9 times, 10 to 19 times, 20 to 39 times, 40 or more times). Participants who responded 0 times in the aforementioned questions, but answered at least once in one of the 8 lifetime drug use questions, were coded as used drugs beyond 30 days. The 8 questions included During your life, how many times ha ve you used marijuana?; During your life, how many times have you used any form of cocaine, including powder, crack, or freebase?; During your life, how many times have you sniffed glue, breathed the contents of aerosol spray cans, or inhaled any paint s or sprays to get high?; During your life, how many times have you used heroin (also called smack, junk, or China White)?; During your life, how many times have you used methamphetamines (also called speed, crystal, crank, or ice)?; During your life how many times have you used ecstasy (also called MDMA)?; During your life, how many times have you taken steroid pills or shots without a doctor's prescription?, (Response options: 0 times, 1 or 2 times, 3 to 9 times, 10 to 19 times, 20 to 39 times, 40 or more times); and During your life, how many times have you used a needle to inject any illegal drug into your body? (Response options: 0 times, 1 time, 2 or more times). Only adolescents who responded 0 times to all 10 questions were classified as never used drugs. Suicidal behavior was categorized into three levels: never thought about it, suicidal ideation, and suicidal attempt. Suicidal attempt was assessed by asking During the past 12 months, how many times did you actually attempt suicide? ( Response options: 0 times, 1 time, 2 or 3 times, 4 or 5 times, 6 or more times). Adolescents who responded 1 or more times were classified as attempting suicide. Suicidal ideation was assessed by asking "During the past 12 months, did you

PAGE 74

74 ever seriously co nsider attempting suicide? Adolescents who responded yes, but answered no in suicidal attempts, were coded as having suicidal ideation but no attempts. Three questions assessed sexual behavior and sexual partners: Have you ever had sexual intercourse? (Response options: Yes, No); During your life, with how many people have you had sexual intercourse? (Response options: I have never had sexual intercourse, 1 person, 2 people, 3 people, 4 people, 5 people, 6 or more people); and During the past 3 mon ths, with how many people did you have sexual intercourse? (Response options: I have never had sexual intercourse, I have had sexual intercourse but not during the past 3 months, 1 person, 2 people, 3 people, 4 people, 5 people, 6 or more people). Student s who responded with 2 or more people either in the past 3 months or lifetime were classified having multiple sex partners. Adolescents who responded never having sexual intercourse in all 3 questions were classified as never had sex. The remainder were cl assified as having sex with one partner. In this study, violence included physical violence, sexual violence, and been bullied. Physical violence was assessed by asking During the past 12 months, did your boyfriend or girlfriend ever hit, slap, or physi cally hurt you on purpose? Sexual violence was assessed by asking Have you ever been physically forced to have sexual intercourse when you did not want to? Been bullied was assessed by asking During the past 12 months, have you ever been bullied on school property? All questions had 2 response options: Yes or No. Data Analysis T o explore relationships between UWCB and the selected risk behaviors including alcohol use, cigarette smoking, drug use, multiple sex partners, and suicidal behavior, Goodman and Kruskals Gamma statistics were used to determine bivariate correlation s due to the ordinal data format B ivariate relationship s between UWCB and physical violence, sexual violence, and been bullied were evaluated by Pearsons P hi coefficient due to binary data format Multiple logistic

PAGE 75

75 regression was conducted to determine the odds ratios for the main effects while controlling the other co varia te s. To assure the quality of analysis, the descriptive statistics were fully compared with the offici al YRBSS statistics on the Youth Online. All analys e s w ere conduc ted by gender because the literature suggested existence of significant differences by gender in each of th e se health risk behaviors. IBM SPSS Statistics 19.0 software was used to conduct the data analyses and to account for the complex sampling design and weighting of data R esults The 2009 YRBS survey included 16,410 participants, with 47.8% female. Grade levels were fairly evenly distributed with slightly more (28%) 9th graders and fewer (22.2%) 12th graders. Most students were Caucasian (58.7%), followed by Hispanic and Multiple races with Hispanic origin (18.6%), African American (14.4%), Multiple Races and non Hispanic origin (3.6%), Asian (3.4%), Native Hawaiian/Other Pacific Islander (0.8%), and American Indian/Alaska Native (0.6%). Table 1 presents the two way contingency table of UWCB and the selected health risk behaviors by gender. Proportions of UWCB were found in ascending order, while strength or severity of health risk behav iors was increased for both female and male adolescents. This finding implies a dose response relationship in the associations of UWCB and each of the selected risk behaviors. In addition, the Goodman and Kruskals Gamma statistics indicated that UWCB was highly correlated with each of the selected health risk behaviors ranging from .34 to .62 with very high statistical significance (p<.001). Among the selected risk behaviors, the highest UWCB prevalence rates were observed in both female (46.0%) and males (40.5%) adolescents who attempted suicide. Suicidal behaviors also showed the highest correlation with UWCB in female adolescents (r=.552, p<.001) and male adolescents (r=.621, p<.001). The Pearsons Phi coefficients indicated UWCB was associated with phys ical violence, sexual

PAGE 76

76 violence, and been bullied among both male and female adolescents (rranged from .11 to .19). All correlation coefficients were highly significant at p<.001 level. Table 2 presents relationships between UWCB and selected health risk behaviors by gender. The logistic regression model revealed that, after controlling for confounding risk behaviors, the strongest relationship between UWCB and the selected health risk behaviors was suicidal behavior in both male and female adolescents. A mong female adolescents, when compared with adolescents who never thought about suicide, those who had suicidal ideation exhibited significantly elevated risk to engage in UWCB (OR=2.04; 95% CI, 1.652.52), and with a 3.2 times significantly higher risk fo r those who attempted suicide (OR=3.19; 95% CI, 2.554.00). In addition, female adolescents who ever tried drugs beyond 30 days exhibited significantly higher risk (OR=1.78; 95% CI, 1.472.15), and those who used drugs in the past 30 days exhibited an even higher risk (OR=1.90; 95% CI, 1.512.37) compared with those who never used drugs. In comparison to those who never smoked, both female occasional smokers and frequent smokers demonstrated an elevated risk of engaging in UWCB (OR=1.71; 95% CI, 1.382.10 f or occasional smokers; OR=1.73; 95% CI, 1.312.27 for frequent smokers). Compared with abstinence, casual drinking females tended to engage in UWCB with OR at 1.37 (95% C.I.: 1.12 1.67), and binge drinking females were at higher risk (OR=1.53, 95% C.I.: 1. 25 1.86). One additional significant risk behavior was been bullied. Female adolescents who have ever been bullied exhibited a higher risk of UWCB compared to those who had not been bullied (OR= 1.35; 95% CI, 1.141.60) Conversely, the logistic regression model revealed 7 significant health risk behaviors among male adolescents. After controlling for other health risk behaviors, male adolescents who reported suicidal ideation had a 1.69 times higher risk of engaging in UWCB (95% C.I.: 1.20-

PAGE 77

77 2.38), and those who attempted suicide had a 3.55 times higher risk (95% C.I.: 2.515.03) when compared with those who never think about suicide. In reference to abstinence, both causal drinking and binge drinking male adolescents had a higher risk of engaging in UWCB (OR = 1.47; 95% CI, 1.082.01 for casual drinking, and OR=2.16; 95% CI, 1.632.85 binge drinking). In addition, having one sex partner showed a 1.44 times higher risk of UWCB (95% C.I.: 1.131.86), and multiple sex partners showed 1.64 times higher risk (95% C .I.: 1.22 2.22), when compared to those who never had sex. Both male occasional smokers (OR=1.65; 95% CI, 1.24 2.19) and frequent smokers (OR=1.41; 95% CI, 1.011.99) had a higher risk of engaging in UWCB when compared to those who never smoke. Lastly, mal e adolescents who were victims of physical violence (OR=1.47; 95% CI, 1.10 1.95), sexual violence (OR=1.71; 95% CI, 1.142.57), and been bullied (OR=1.85; 95% CI, 1.47 2.32) also showed a higher risk of engaging in UWCB when compared to those who were not victims. D iscussion Our multivariate analyses based on a nationally representative sample confirm ed some of the bivariate findings in the literature and p rovided more in depth statistic al analyses C o occurrences of UWCB and the selected risk behaviors w ere found in both male and female adolescents. However, our findings also suggest ed different degrees of the association Although our bivariate analysis revealed significant associations between UWCB and all of the selected health risk behaviors in both genders the multivariate analys e s suggest ed a different list of significant risk behavior associations between males and females. Among female adolescents, alcohol drinking, cigarette smoking, drug use, suicidal behaviors, and been bullied were associated with UWCB, while male adolescent s UWCB exhibited significant association s with alcohol drinking, cigarette smoking, multiple sex partners suicidal behaviors, physical violence, sexual violence and been bullied

PAGE 78

78 Consistent with the findings from previ ous studies (Cance et al., 2005; Croll et al., 2002; Haley et al., 2010) adolescents in this study who used substances such as alcohol, cigarette, and drugs were more likely to engage in UWCB. Although Haley and colleagues bivariate analysis (2010) showed that girls who used alcohol, illicit drugs, and tobacco were 3 4 times more likely to engage in UWCB than nonusers, our multivariate analysis with an adjustment over the othe r 7 risk behaviors showed female substance users had only about 1.7 times elevated risk compared to nonusers. The strength of the association was similar with Cance and colleagues multivariate analysis controlling for demographics and other substance use In addition, findings from this study suggest ed a higher risk of UWCB among female adolescents when they used more substances in frequency or quantity. However, this phenomenon was only observed with alcohol use among male adolescents In regard to the relationship between UWCB and sexual behaviors, previous studies showed female adolescents who engaged in UWCB were approximately 1.6 times more likely to report being sexually active (Cook et al., 2007; McGuire et al., 2002) Although our bivariate analysis support ed this association among female adolescents, the multivariate analysis did not suggest a statistically significant association. Yet our findings suggested male adoles cents who had one sex ual partner were 1.4 times more likely to engag e in UWCB than male adolescents who had not have sex. Moreover, male adolescents who ha d multiple sexual partners exhibited an even higher likelihood (1.6 times) to engage in UWCB when com par ed to those who never had sex. In addition, t he strong association between UWCB and suicidal behavior echoed similar strength (2 4 times) of the association found in previous studies (Cook et al., 2007; Kim et al., 2009; Silverman et al., 2001) However the association between UWCB and suicidal behavior

PAGE 79

79 was in a different direction. O ur findings suggest ed that a dolescents who attempted suicide exhibited a higher risk of UWCB than those who only contemplated (thought about suicid e) wh ile Cook and colleagues multivariate analysis (2007) found ed th at adolescents with suicidal ideation exhibited higher risk of UWCB than those who attempted suicid e Such diffe rence s warranted further investigation Our multivariate findings also confirmed the association between UWCB and violen ce behaviors. M ale adolescents who experienced physical and sexual violence (Haley et al., 2010; Silverman et al., 2001) and had been bullied in school tended to be more likely to engage in UWCB, while female only adolescent s exhibited a significant association between UWCB and the experience of being bull ied Our findings did not show multivariate significance in the association between UWCB and physical and sexual violence among females. This result differ s from Silverman and colleagues multivariate analysis (2001) using 1997 and 1999 Massachusetts YRBS whic h show ed strong associations between UWCB and physical and sexual violence among female adolescents. Whether such differences occurred due to sample size (7 300s vs. 2, 000s ), year of study (2009 vs. 1997 and 1999), coverage (U.S. vs. 1 state), outcome vari able in the regression (UWCB vs. lifetime violence from dating partners), selection of covariates in the regression or other methodological considerations can shed light for future studies. Moreover, the study ha d shown that bullying occur s when conflict exists involving and imbalance of power in which the victim is less powerful than the aggressor (Rigby, 2003) Findings in the current study may imply that male students with UWCB were relatively less powerful than thei r male counterpart s and their intimate sexual partner. Results from the study suggest prevention efforts should be extended to male adolescents who experienced p hysical and sexual violence and had been bullied o n school property

PAGE 80

80 R esults from this study clearly demonstrate link s between UWCB and the selected health risk behaviors, future research should determine the underlying psychological and social mechanisms behind the co occurrence of these health risk behaviors for use in develop ing mor e effective interventions that address multiple risk behaviors. Such research initiatives may include the development of more detailed, multiple item instrument s with sound psychometric propertie s to assess each of the selected health risk behaviors among adolescents. Some study limitations should be noted Due to the cross sectional design, the direction of causality between UWCB and selected health risk behaviors cannot be determined. In addition, though the data include d representative samples of high s chool adolescent s across the U.S. which allows for the extrapolations of findings from the study population to a larger extent the r esults do not represent adolescents who did not attend school. With the limitations and the needs for additional research duly noted, findings from this study still have significant implications for intervention and prevention program for adolescents in a number of ways. First, health risk behaviors were prevalent and highly associated with each other among high school adolescents. As a result, school health professionals should routinely screen UWCB and the related health compromised behaviors to help identifying high risk adolescents, especially those with more frequ ent or a greater degree of exposure, such as heavy alcohol drinkers, frequent smokers, recent drug users, and those who had multiple sex partners and have attempted suicide. Second, although female adolescents were more prevalent in UWCB, male adolescents who engaged in UWCB also need attention because health risk behaviors such as suicidal behaviors, substance use, multiple sexual partners, physical and sexual violence, and having been bullied can co occur with their UWCB. Lastly, adolescents who engaged i n UWCB face a higher risk of engaging in many different health compromising

PAGE 81

81 behaviors. Therefore, practitioners working in school, community, and clinical sites to combat substance abuse, sexual risk behavior, suicidal behavior, and violence are encouraged to include UWCB in the range of their prevention programs. Intervention programs with comprehensive strategies that cover multiple risk behaviors are highly recommended.

PAGE 82

82 Table 46 UWCB d istribution b y s ex and s elected h ealth r is k b ehaviors Health Risk Behaviors Female Male Yes No Correlation Yes No Correlation n % n % n % n % Alcohol .417*** .437*** Abstinence 487 12.2 3519 87.8 233 5.1 4304 94.9 Casual d rink ing 274 21.2 1018 78.8 96 8.6 1019 91.4 Binge d rink ing 540 31.1 1194 69.9 335 16.7 1666 83.3 Tobacco .506*** .466*** Non s moker 795 14.2 4823 85.8 317 5.9 5097 94.1 Occasional s moker 369 33.5 733 66.5 229 13.7 1437 86.3 Frequent s moker 183 38.9 287 61.1 127 20.3 499 79.7 Drug .449*** .370*** Never used d rug 536 12.2 3869 87.8 275 6.0 4298 94.0 Used d rug beyond 30 days 385 25.3 1134 74.7 158 11.2 1252 88.8 Drug use in 30 days 446 33.8 875 66.2 281 15.1 1580 84.9 Sexual Risk Behavior .338*** .423*** Never had s ex 537 13.7 3381 86.3 223 5.3 3974 94.7 Had s ex one partner 491 21.3 1815 78.7 218 10.4 1883 89.6 Had s ex m ultiple partners 326 32.9 666 67.1 246 17.4 1170 82.6 Suicidal Behavior .552*** .621*** Never thought about it 854 14.6 4980 85.4 470 7.0 6271 93.0 Suicidal i deation but never attempted 216 32.0 460 68.0 61 12.9 424 87.1 Attempted 263 46.0 309 54.0 137 40.5 201 59.5 Physical Violence .112*** .120*** Yes 233 32.9 475 67.1 170 20.1 676 79.9 No 1226 17.7 5698 82.3 625 8.4 6819 91.6 Sexual Violence .141*** .191*** Yes 271 35.1 500 64.9 129 35.8 231 64.2 No 1131 17.1 5497 82.9 653 8.4 7088 91.6 Been bullied .114*** .111*** Yes 428 27.8 1112 72.2 245 16.6 1235 83.4 No 962 16.8 4770 83.2 529 8.1 5990 91.9 Note: ***: p<.001

PAGE 83

83 Table 47 Relationships between UWCB and s elected h ealth r isk b ehaviors by s ex Health Risk Behaviors Female Male Odds Ratio 95% Confidence Interval Odds Ratio 95% Confidence Interval Alcohol Casual d rink ing 1.37** 1.12, 1.67 1.47* 1.08, 2.01 Binge drinking 1.53*** 1.25, 1.86 2.16*** 1.63, 2.85 Abstinence Reference Tobacco Occasional s moker 1.71*** 1.38, 2.10 1.65*** 1.24, 2.19 Frequent s moker 1.73*** 1.31, 2.27 1.41* 1.01, 1.99 Non s moker Reference Drug Used d rug beyond 30 days 1.78*** 1.47, 2.15 1.15 0.87, 1.53 Drug use in 30 days 1.90*** 1.51, 2.37 0.92 0.68, 1.26 Never used d rug Reference Sexual Risk Behavior Had s ex one partner 1.01 0.84, 1.20 1.44** 1.13, 1.86 Had s ex m ultiple partners 1.18 0.93, 1.49 1.64*** 1.22, 2.22 Never had s ex Reference Suicidal Behavior Suicidal i deation but never attempted 2.04*** 1.65, 2.52 1.69** 1.20, 2.38 Attempted 3.19*** 2.55, 4.00 3.55*** 2.51, 5.03 Never thought about it Reference Physical Violence Yes 1.20 0.95, 1.52 1.47** 1.10, 1.95 No Reference Sexual Violence Yes 1.23 0.98, 1.55 1.71** 1.14, 2.57 No Reference Been bullied Yes 1.35*** 1.14, 1.60 1.85*** 1.47, 2.32 No Reference Note: *: p<.05, **: p<.01, ***: p<.001

PAGE 84

84 CHAPTER 5 SUMMARY, CONCLUSIONS RECOMMENDATIONS AND IMPLICATIONS Summary The present study focused on the Unhealthy Weight Control Behaviors (UWCB) among high school students. Specifically, the study served three purposes: 1) to estimate the trends of UWCB and Hea lthy Weight Control Behavior (HWCB). 2) to identify the predictive factors of UWCB. 3) to explore the relationship between UWCB and the selected health risk behaviors. The study in Chapter 2 analyzed YRBS 19992009 data to examine the possible linear and quadratic trends of HWCB and UWCB among high school students Findings suggested weight control behaviors were prevalen t among high school students. Moreover, a significant linear descending trend of UWCB and a significant linear ascending trend of HWCB we re identified S ubgroup analyses also showed the presence of similar trends of HWCB and UWCB by sex, ethnic groups, and BID status. The bivariate analys e s on the pooled data revealed that female adolescents and nonoverw eight adolescents with BID tended to engage more in both HWCB and UWCB. Compared with Caucasian Asian and African American adolescents tended not to engage in UWCB, while adolescents of Hispanic, m ultiple race with Hispanic origin and m ultiple race without Hispan i c origin tended to engage more in UWCB. African American and Asian adolescents practiced less HWCB, while adolescents in m ultiple race with Hispanic origin and Hispanic engaged in more HWCB than Caucasian s Chapter 3 used the most recent (2009) YRBS data to identify the predictive factors of UWCB among high school students The stepwise multiple logistic regression with backward elimination modeled 7 predictive factors. Two way interactions of significant variables were considered and the backward elimination retained 8 significant 2 way interaction effect s H owever the complex model with 8 interaction effect s only reduced 1.5% of Akaike

PAGE 85

85 Information Criterion (AIC), which judges the model by how close its fitted values tend to be the true expected values Th erefore, the model with 7 main effects without interactions was adopted. The final logistic regression revealed that trying to lose weight being American Indian/Alaska Native or multiple race/ethnicity nonover weight but having body image distortion (BID) eating the recommended amount of vegetables being female drinking sugar sweetened beverage daily and being not vigorous in physical activ ity tended to engage more in UWCB. Chapter 4 also used 2009 YRBS data to explore the relationship s between UWCB and selected health risk behaviors. The bivariate analysis revealed significant association between UWCB and alcohol use, cigarette smoking, drug use, suicidal behaviors, multiple sexual partners, physical and sexual violence and been bullied among both ma le and female adolescents F indings from this analysis also indicat ed a dose response relationship in the associations of UWCB and eac h of the selected risk behaviors. Moreover, t he multivariate analysis identified significant associations between UWCB and suicidal behaviors, drug use, cigarette smoking, alcohol use, and been bullied among female adolescents, while male adolescents demonstrated more significant associations, including suicidal behaviors, alcohol use, multiple sexual partners, cigarette smoking, physical and sexual violence, and been bullied. Conclusions These investigations represented the first efforts to examine trends of weight control behaviors, predictive factors of UWCB, associations between UWCB and selected health risk behavior among high school adolescents based on national representative survey data. Findings from this line of research suggested one out of seven adolescents engaged in UWCB in 2009 and confirmed female adolescents were at higher risk of engaging in UWCB (K. Boutelle et al., 2002; Fonseca et al., 2009; Lowry et al., 2002; McCabe & Ricciardelli, 2009; D. NeumarkSztainer et al., 2006; Striegel Moore et al., 2009) T his study also confirmed that Body Image

PAGE 86

86 Distortion increase d the risk of UWCB (Liechty, 2010) and further dem onstrated the elevate d risk of UWCB among the nonoverweight adolescents with BID Findings in this study also verified that vigorous physical activity was a strong pr otect ive factor against UWCB (D. R. Neumark Sztainer et al., 2007) while daily sugar sweetened beverage consumption was identified as a risk factor. In regards to the associations between UWCB and selected health risk factors, this study confirmed adolescents who used substances, such as alcohol, cigarette, and drug were more likely to engage in UWCB (Cance et al., 2005; Croll et al., 2002; Haley et al., 2010) In additio n, t he strong association between UWCB and suicidal behavior echoed similar strength (2 4 times) of the association found from the previous studies (Cook et al., 2007; Kim et al ., 2009; Silverman et al., 2001) However the association between UWCB and suicidal behavior was found in different direction after multivariate analysis O ur findings suggest ed that a dolescents who attempted suicide exhibited higher risk of UWCB than those who only contemplated (thought about suicid e) wh ile Cook and colleagues multivariate analysis (2007) showed the adolescents with suicidal ideation exhibited higher risk of UWCB than those who attempted suicid e Lastly, M ale adolescents who experienced physical and sexual violence (Haley et al., 2010; Silverman et al., 2001) and been bullied in school tended to be more likely to engage in UWCB, while female adolescent s only exhibited significant association between UWCB and the experience of being bull ied Recommendations for Future Research The findings of this study suggest the following actions: To investigate t he latent effect s such as social and culture influences behind associations between UWCB and sex, race/ethnicity, and BID

PAGE 87

87 To investiga te the mechanism s of how BID interacts with UWCB in order to provide insight s to the effectiveness of UWCB prevention programs. T o monitor the weight control behaviors, weight control knowledge dissemination, and obstacles to healthy weight control behaviors to provide insight s for weight control program planning specifically targeting adolescents. T o determine the underlying psychological and social mechanisms behind the co occurrence of these health risk behaviors and to develop more effective interventions with coverage of multiple risk behaviors. Such research initiatives may include the development of more detailed, multiple item instrument s with sound psychometric propertie s to assess each of the selected health risk behaviors among adolescents. I mplications Despite the need for further research, findings of this study have significant implications for prevention/intervention programming with adolescents in a number of impo rtant areas. First, UWCB is prevalent in some populations. As a result, early identification of eating disorders is recommended as it had been shown to lead to more successful treatment outcomes (Bulik et al., 2007) Thus, school health professionals should routinely screen UWCB and the related health compromised behaviors to help identify ing high risk adolescents, especially the ones with more frequent and/or greater degree of exposure, such as heavy alcohol drinker, frequent smokers, recent drug users, and those who had multiple sex partners and have attempted suicide Second, b ased on our finding s it is reasonable to believe a brief assessment of weight status and BID can help identify adolescents at risk By streng then ing the prevention efforts for such an at risk population, the lower risk of UWCB can be projected Third, although UWCB appeared more prevalent among female adolescents, male adolescents who engaged in UWCB also need

PAGE 88

88 attention s because health risk behaviors, such as suicidal behaviors, substance use, multiple sexual partners, physical/sexual violence and been bullied could co occur with their UWCB Fourth a combination of encouraging vigorous physical activity and discouraging sugar sweetened beverag e consumption can yield great benefits in promoting healthy weight control behaviors and prevent ing the unhealthy weight control behaviors. Last ly, adolescents who engaged in UWCB are at high er risk of engaging in many different health compromis ing behaviors T herefore the practitioners working in school, community, and clinical sites to combat substance abuse, sexual risk behavior, suicidal behavior and violence are encouraged to include UWCB in the range of their prevention programs Interventio n programs with comprehensive strategies that cover multiple risk behaviors are highly recommended.

PAGE 89

89 LIST OF REFERENCES Adam, S., Westenhofer, J., Rudolphi, B., & Kraaibeek, H. K. (2009). Effects of a combined inpatient outpatient treatment of obese children and adolescents. Obesity Facts, 2(5), 286 293. doi:10.1159/000234415 Agresti, A. (2007). An introduction to categorical data analysis (2nd ed.). Hoboken, NJ: Wiley Interscience. Agresti, A., & Finlay, B. (1997). Statistical methods for the social sciences (3rd ed.). Upper Saddle River, N.J.: Prentice Hall. Austin, S. B., Spadano Gasbarro, J., Greaney, M. L., Richmond, T. K., Feldman, H. A., Osganian, S. K., et al. (2011). Disordered weight control behaviors in early adolescent boys and girls of color: An under recognized factor in the epidemic of childhood overweight. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 48(1), 109112. doi:10.1016/j.jadohealth.2010.05.017 Batch, J. A. (2005). Benefits of physical activity in obese adolescents and children. Internal Medicine Journal, 35(8), 446. doi:10.1111/j.14455004.2005.00876.x Boslaugh, S. (2007). Secondary data sources for public health : A practical guide. Cambridge ; New York: Cambridge University Press. Retrieved from Publisher description http://www.loc.gov/catdir/enhancements/fy0703/2006034301d.html; Table of contents only http://www.loc.gov/catdir/toc/ecip073/2006034301.html Boutelle, K., Neumark Sztainer, D., Story, M., & Resnick, M. (2002). Weight control behaviors among obese, overweight, and nonoverweight adole scents. Journal of Pediatric Psychology, 27(6), 531 540. Boutelle, K. N., Libbey, H., Neumark Sztainer, D., & Story, M. (2009). Weight control strategies of overweight adolescents who successfully lost weight. Journal of the American Dietetic Association, 109(12), 20292035. doi:10.1016/j.jada.2009.09.012 Bray, G. A. (2009). Medications for obesity: Mechanisms and applications. Clinics in Chest Medicine, 30(3), 52538, ix. doi:10.1016/j.ccm.2009.05.014 Breitkopf, C. R., & Berenson, A. B. (2004). Correlates of weight loss behaviors among low income african american, caucasian, and latina women. Obstetrics and Gynecology, 103(2), 231239. Brener, N. D., Kann, L., Kinchen, S. A., Grunbaum, J. A., Whalen, L., Eaton, D., et al (2004). Methodology of the youth risk behavior surveillance system. MMWR.Recommendations and Reports : Morbidity and Mortality Weekly Report.Recommendations and Reports / Centers for Disease Control, 53(RR 12), 113.

PAGE 90

90 Bulik, C. M., Berkman, N. D., Brownl ey, K. A., Sedway, J. A., & Lohr, K. N. (2007). Anorexia nervosa treatment: A systematic review of randomized controlled trials. The International Journal of Eating Disorders, 40(4), 310320. doi:10.1002/eat.20367 Cance, J. D., Ashley, O. S., & Penne, M. A. (2005). Unhealthy weight control behaviors and MDMA (ecstasy) use among adolescent females. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 37(5), 409. Catenacci, V. A., Hill, J. O., & Wyatt, H. R. (2009) The obesity epidemic. Clinics in Chest Medicine, 30(3), 41544, vii. doi:10.1016/j.ccm.2009.05.001 Centers for Disease Control and Prevention (CDC). (2010a). 2009 national YRBS data user manual. Retrieved 08/15, 2010, from http://www.cdc.gov/HealthyYouth/yrbs/pdf/national_usersmanual_yrbs.pdf Centers for Disease Control and Prevention (CDC). (2010b). Youth risk behavior survey (YRBS) standard 2011 high schoo l questionnaire item rationale. Retrieved 02/23, 2011, from http://www.cdc.gov/HealthyYouth/yrbs/pdf/questionnaire/2011_standard_itemrationale.pd f Chen, M. Y., Fan, J. Y., Jane, S. W., & Wu, J. Y. (2009). Do overweight adolescents perceive the need to reduce weight and take healthy actions? The Journal of Nursing Research : JNR, 17(4), 270277. Cheung, P. C., Ip, P. L., Lam, S. T., & Bibby, H. (2007). A study on body weight perception and weight control behaviours among adolescents in hong kong. Hong Kong Medical Journal = Xianggang Yi Xue Za Zhi / Hong Kong Acade my of Medicine, 13(1), 1621. Clarke, S. P., & Cossette, S. (2000). Secondary analysis: Theoretical, methodological, and practical considerations. The Canadian Journal of Nursing Research = Revue Canadienne De Recherche En Sciences Infirmieres, 32(3), 109 129. Cook, S. J., MacPherson, K., & Langille, D. B. (2007). Far from ideal: Weight perception, weight control, and associated risky behaviour of adolescent girls in nova scotia. Canadian Family Physician Medecin De Famille Canadien, 53(4), 67984, 678. Croll, J., Neumark Sztainer, D., Story, M., & Ireland, M. (2002). Prevalence and risk and protective factors related to disordered eating behaviors among adolescents: Relationship to gender and ethnicity. The Journal of Adolescent Health : Official Publica tion of the Society for Adolescent Medicine, 31(2), 166 175. Crow, S., Eisenberg, M. E., Story, M., & Neumark Sztainer, D. (2006). Psychosocial and behavioral correlates of dieting among overweight and nonoverweight adolescents. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 38(5), 569574. doi:10.1016/j.jadohealth.2005.05.019

PAGE 91

91 Daee, A., Robinson, P., Lawson, M., Turpin, J. A., Gregory, B., & Tobias, J. D. (2002). Psychologic and physiologic effects of diet ing in adolescents. Southern Medical Journal, 95(9), 10321041. Eaton, D. K., Kann, L., Kinchen, S., Ross, J., Hawkins, J., Harris, W. A., et al. (2006). Youth risk behavior surveillance --united states, 2005. MMWR.Surveillance Summaries : Morbidity and Mo rtality Weekly Report.Surveillance Summaries / CDC, 55(5), 1 108. Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Ross, J., Hawkins, J., et al. (2008). Youth risk behavior surveillance --united states, 2007. MMWR.Surveillance Summaries : Morbidity and M ortality Weekly Report.Surveillance Summaries / CDC, 57(4), 1 131. Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Ross, J., Hawkins, J., et al. (2010). Youth risk behavior surveillance united states, 2009. MMWR.Surveillance Summaries : Morbidity and Mortality Weekly Report.Surveillance Summaries / CDC, 59(5), 1 142. Eisenberg, M. E., Neumark Sztainer, D., Story, M., & Perry, C. (2005). The role of social norms and friends' influences on unhealthy weight control behaviors among adolescent girls. Soci al Science & Medicine (1982), 60(6), 1165 1173. Epstein, L. H., Roemmich, J. N., Paluch, R. A., & Raynor, H. A. (2005). Physical activity as a substitute for sedentary behavior in youth. Annals of Behavioral Medicine : A Publication of the Society of Beha vioral Medicine, 29(3), 200209. doi:10.1207/s15324796abm2903_6 Epstein, L. H., Valoski, A., Wing, R. R., & McCurley, J. (1994). Ten year outcomes of behavioral family based treatment for childhood obesity. Health Psychology : Official Journal of the Divi sion of Health Psychology, American Psychological Association, 13(5), 373383. Epstein, L. H., Valoski, A. M., Vara, L. S., McCurley, J., Wisniewski, L., Kalarchian, M. A., et al. (1995). Effects of decreasing sedentary behavior and increasing activity on weight change in obese children. Health Psychology : Official Journal of the Division of Health Psychology, American Psychological Association, 14(2), 109 115. Flegal, K. M., Carroll, M. D., Ogden, C. L., & Curtin, L. R. (2010). Prevalence and trends in obesity among US adults, 19992008. JAMA : The Journal of the American Medical Association, 303(3), 235241. doi:10.1001/jama.2009.2014 Fonseca, H., Matos, M. G., Guerra, A., & Pedro, J. G. (2009). Are overweight adolescents at higher risk of engaging in unhealthy weight control behaviours? Acta Paediatrica (Oslo, Norway : 1992), 98(5), 847 852. Foti, K., & Lowry, R. (2010). Trends in perceived overweight status among overweight and nonoverweight adolescents. Archives of Pediatrics & Adolescent Medicine, 164(7), 636642. doi:10.1001/archpediatrics.2010.90

PAGE 92

92 French, S. A., Leffert, N., Story, M., Neumark Sztainer, D., Hannan, P., & Benson, P. L. (2001). Adolescent binge/purge and weight loss behaviors: Associations with developmental assets. The Journal of A dolescent Health : Official Publication of the Society for Adolescent Medicine, 28(3), 211221. Fulton, J. E., McGuire, M. T., Caspersen, C. J., & Dietz, W. H. (2001). Interventions for weight loss and weight gain prevention among youth: Current issues. S ports Medicine (Auckland, N.Z.), 31(3), 153 165. Gidding, S. S., Dennison, B. A., Birch, L. L., Daniels, S. R., Gillman, M. W., Lichtenstein, A. H., et al. (2005). Dietary recommendations for children and adolescents: A guide for practitioners: Consensus statement from the american heart association. Circulation, 112(13), 20612075. doi:10.1161/CIRCULATIONAHA.105.169251 Golden, N. H., Katzman, D. K., Kreipe, R. E., Stevens, S. L., Sawyer, S. M., Rees, J., et al. (2003). Eating disorders in adolescents: Po sition paper of the society for adolescent medicine. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 33(6), 496503. Grunbaum, J. A., Kann, L., Kinchen, S., Ross, J., Hawkins, J., Lowry, R., et al. (2004). Youth risk behavior surveillance --united states, 2003. MMWR.Surveillance Summaries : Morbidity and Mortality Weekly Report.Surveillance Summaries / CDC, 53(2), 1 96. Grunbaum, J. A., Kann, L., Kinchen, S. A., Williams, B., Ross, J. G., Lowry, R., et al. (2002). Youth risk behavior surveillance --united states, 2001. MMWR.Surveillance Summaries : Morbidity and Mortality Weekly Report.Surveillance Summaries / CDC, 51(4), 1 62. Guo, S. S., & Chumlea, W. C. (1999). Tracking of body mass index in children in relation to overweight in adulthood. The American Journal of Clinical Nutrition, 70(1), 145S 8S. Haines, J., Neumark Sztainer, D., Perry, C. L., Hannan, P. J., & Levine M. P. (2006). V.I.K. (very important kids): A school based program designed to reduce teasing and unhealthy weight control behaviors. Health Education Research, 21(6), 884 895. Haley, C. C., Hedberg, K., & Leman, R. F. (2010). Disordered eating and unhe althy weight loss practices: Which adolescents are at highest risk? The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 47(1), 102 105. Harring, H. A., Montgomery, K., & Hardin, J. (2010). Perceptions of body we ight, weight management strategies, and depressive symptoms among US college students. Journal of American College Health : J of ACH, 59(1), 43 50. Hart, C. N., Jelalian, E., Raynor, H. A., Mehlenbeck, R., Lloyd Richardson, E. E., Kaplan, J., et al. (2010). Early patterns of food intake in an adolescent weight loss trial as predictors of BMI change. Eating Behaviors, 11(4), 217222. doi:10.1016/j.eatbeh.2010.05.001

PAGE 93

93 Hart, K. E., & Warriner, E. M. (2005). Weight loss and biomedical health improvement on a v ery low calorie diet: The moderating role of history of weight cycling. Behavioral Medicine (Washington, D.C.), 30(4), 161170. Heywood, S., & McCabe, M. P. (2006). Negative affect as a mediator between body dissatisfaction and extreme weight loss and mus cle gain behaviors. Journal of Health Psychology, 11(6), 833 844. Ho, C. H., Kingree, J. B., & Thompson, M. P. (2006). Associations between juvenile delinquency and weight related variables: Analyses from a national sample of high school students. The Int ernational Journal of Eating Disorders, 39(6), 477 483. doi:10.1002/eat.20271 Hutchinson, D. M., & Rapee, R. M. (2007). Do friends share similar body image and eating problems? the role of social networks and peer influences in early adolescence. Behaviou r Research and Therapy, 45(7), 15571577. Johnson, J. L., Eaton, D. K., Pederson, L. L., & Lowry, R. (2009). Associations of trying to lose weight, weight control behaviors, and current cigarette use among US high school students. The Journal of School He alth, 79(8), 355 360. Kann, L., Kinchen, S. A., Williams, B. I., Ross, J. G., Lowry, R., Grunbaum, J. A., et al. (2000). Youth risk behavior surveillance --united states, 1999. MMWR.CDC Surveillance Summaries : Morbidity and Mortality Weekly Report.CDC Sur veillance Summaries / Centers for Disease Control, 49(5), 1 32. Kim, D. S., Cho, Y., Cho, S. I., & Lim, I. S. (2009). Body weight perception, unhealthy weight control behaviors, and suicidal ideation among korean adolescents. The Journal of School Health, 79(12), 585 592. Kuczmarski, R. J., Ogden, C. L., Guo, S. S., Grummer Strawn, L. M., Flegal, K. M., Mei, Z., et al. (2002). 2000 CDC growth charts for the united states: Methods and development. Vital and Health Statistics.Series 11, Data from the Nation al Health Survey, (246)(246), 1 190. Liechty, J. M. (2010). Body image distortion and three types of weight loss behaviors among nonoverweight girls in the united states. The Journal of Adolescent Health : Official Publication of the Society for Adolescen t Medicine, 47(2), 176182. Lowry, R., Galuska, D. A., Fulton, J. E., Wechsler, H., & Kann, L. (2002). Weight management goals and practices among U.S. high school students: Associations with physical activity, diet, and smoking. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 31(2), 133144. Malinauskas, B. M., Raedeke, T. D., Aeby, V. G., Smith, J. L., & Dallas, M. B. (2006). Dieting practices, weight perceptions, and body composition: A comparison of norm al weight, overweight, and obese college females. Nutrition Journal, 5, 11.

PAGE 94

94 Mathias, J. L., & Kent, P. S. (1998). Neuropsychological consequences of extreme weight loss and dietary restriction in patients with anorexia nervosa. Journal of Clinical and Exp erimental Neuropsychology, 20(4), 548564. McCabe, M. P., & Ricciardelli, L. A. (2009). Extreme weight change behaviours: Are overweight and normal weight adolescents different, and does this vary over time? European Eating Disorders Review : The Journal of the Eating Disorders Association, 17(4), 301314. McGuire, M. T., Story, M., Neumark Sztainer, D., Halcon, L., Campbell Forrester, S., & Blum, R. W. (2002). Prevalence and correlates of weight control behaviors among caribbean adolescent students. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 31(2), 208211. National Research Council (US) and Institute of Medicine (US) Committee on Developing a Strategy to Reduce and Prevent Underage Drinking. (2004). Neumark Sztainer, D. (1999). The weight dilemma: A range of philosophical perspectives. International Journal of Obesity and Related Metabolic Disorders : Journal of the International Association for the Study of Obesity, 23 Suppl 2, S317. Neumark Sztain er, D., Story, M., & French, S. A. (1996). Covariations of unhealthy weight loss behaviors and other highrisk behaviors among adolescents. Archives of Pediatrics & Adolescent Medicine, 150(3), 304308. Neumark Sztainer, D., Story, M., French, S. A., & Re snick, M. D. (1997). Psychosocial correlates of health compromising behaviors among adolescents. Health Education Research, 12(1), 3752. Neumark Sztainer, D., Story, M., Toporoff, E., Cassuto, N., Resnick, M. D., & Blum, R. W. (1996). Psychosocial predic tors of binge eating and purging behaviors among adolescents with and without diabetes mellitus. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 19(4), 289 296. Neumark Sztainer, D., Wall, M., Eisenberg, M. E., Story, M., & Hannan, P. J. (2006). Overweight status and weight control behaviors in adolescents: Longitudinal and secular trends from 1999 to 2004. Preventive Medicine, 43(1), 5259. doi:10.1016/j.ypmed.2006.03.014 Neumark Sztainer, D., Wall, M., Guo J., Story, M., Haines, J., & Eisenberg, M. (2006). Obesity, disordered eating, and eating disorders in a longitudinal study of adolescents: How do dieters fare 5 years later? Journal of the American Dietetic Association, 106(4), 559 568. Neumark Sztaine r, D., Wall, M., Story, M., & Sherwood, N. E. (2009). Five year longitudinal predictive factors for disordered eating in a population based sample of overweight adolescents: Implications for prevention and treatment. The International Journal of Eating Dis orders, 42(7), 664 672.

PAGE 95

95 Neumark Sztainer, D., Wall, M. M., Story, M., & Perry, C. L. (2003). Correlates of unhealthy weight control behaviors among adolescents: Implications for prevention programs. Health Psychology : Official Journal of the Division of Health Psychology, American Psychological Association, 22(1), 88 98. Ne umark Sztainer, D. R., Wall, M. M., Haines, J. I., Story, M. T., Sherwood, N. E., & van den Berg, P. A. (2007). Shared risk and protective factors for overweight and disordered eating in adolescents. American Journal of Preventive Medicine, 33(5), 359 369. Ogden, C. L., & Carroll, M. D. (2010). Prevalence of obesity among children and adolescents: United states, trends 19631965 through 20072008CDC/National Center for Health Statistics. Ogden, C. L., Carroll, M. D., Curtin, L. R., Lamb, M. M., & Flegal, K. M. (2010). Prevalence of high body mass index in US children and adolescents, 20072008. JAMA : The Journal of the American Medical Association, 303(3), 242 249. doi:10.1001/jama.2009.2012 Ogden, C. L., & Flegal, K. M. (2010). Changes in terminology fo r childhood overweight and obesity. National Health Statistics Reports, (25)(25), 1 5. Patton, G. C., Selzer, R., Coffey, C., Carlin, J. B., & Wolfe, R. (1999). Onset of adolescent eating disorders: Population based cohort study over 3 years. BMJ (Clinica l Research Ed.), 318(7186), 765768. Position of the american dietetic association: Optimal weight as a health promotion strategy.(1989). Journal of the American Dietetic Association, 89(12), 18141817. Pritts, S. D., & Susman, J. (2003). Diagnosis of ea ting disorders in primary care. American Family Physician, 67(2), 297 304. Rahman, M., & Berenson, A. B. (2010). Self perception of weight and its association with weight related behaviors in young, reproductive aged women. Obstetrics and Gynecology, 116( 6), 12741280. doi:10.1097/AOG.0b013e3181fdfc47 Reinehr, T., Widhalm, K., l'Allemand, D., Wiegand, S., Wabitsch, M., Holl, R. W., et al. (2009). Two year follow up in 21,784 overweight children and adolescents with lifestyle intervention. Obesity (Silver Spring, Md.), 17(6), 11961199. doi:10.1038/oby.2009.17 Rigby, K. (2003). Consequences of bullying in schools. Canadian Journal of Psychiatry.Revue Canadienne De Psychiatrie, 48(9), 583 590. Rippe, J. M., & Hess, S. (1998). The role of physical activity in the prevention and management of obesity. Journal of the American Dietetic Association, 98(10 Suppl 2), S31 8. Serdula, M. K., Mokdad, A. H., Williamson, D. F., Galuska, D. A., Mendlein, J. M., & Heath, G. W. (1999). Prevalence of attempting weight los s and strategies for controlling weight. JAMA : The Journal of the American Medical Association, 282(14), 13531358.

PAGE 96

96 Silverman, J. G., Raj, A., Mucci, L. A., & Hathaway, J. E. (2001). Dating violence against adolescent girls and associated substance use, unhealthy weight control, sexual risk behavior, pregnancy, and suicidality. JAMA : The Journal of the American Medical Association, 286(5), 572579. Smolak, L., & Levine, M. P. (2001). A two year follow up of a primary prevention program for negative body image and unhealthy weight regulation. Eating Disorders, 9(4), 313 325. Stice, E., Cameron, R. P., Killen, J. D., Hayward, C., & Taylor, C. B. (1999). Naturalistic weight reduction efforts prospectively predict growth in relative weight and onset of obes ity among female adolescents. Journal of Consulting and Clinical Psychology, 67(6), 967974. Story, M., Hauck, F. R., Broussard, B. A., White, L. L., Resnick, M. D., & Blum, R. W. (1994). Weight perceptions and weight control practices in american indian and alaska native adolescents. A national survey. Archives of Pediatrics & Adolescent Medicine, 148(6), 567571. Strauss, R. S., & Mir, H. M. (2001). Smoking and weight loss attempts in overweight and normalweight adolescents. International Journal of Obesity and Related Metabolic Disorders : Journal of the International Association for the Study of Obesity, 25(9), 1381 1385. StriegelMoore, R. H., Rosselli, F., Perrin, N., DeBar, L., Wilson, G. T., May, A., et al. (2009). Gender difference in the preval ence of eating disorder symptoms. The International Journal of Eating Disorders, 42(5), 471 474. doi:10.1002/eat.20625 Talamayan, K. S., Springer, A. E., Kelder, S. H., Gorospe, E. C., & Joye, K. A. (2006). Prevalence of overweight misperception and weigh t control behaviors among normal weight adolescents in the united states. TheScientificWorldJournal, 6, 365373. doi:10.1100/tsw.2006.70 The, N. S., Suchindran, C., North, K. E., Popkin, B. M., & Gordon Larsen, P. (2010). Association of adolescent obesity with risk of severe obesity in adulthood. JAMA : The Journal of the American Medical Association, 304(18), 20422047. doi:10.1001/jama.2010.1635 US Dept of Health and Human Services. (2000). Healthy people 2010 (2nd ed.). Washington, DC: Government Print ing Office. van den Berg, P., Neumark Sztainer, D., Hannan, P. J., & Haines, J. (2007). Is dieting advice from magazines helpful or harmful? five year associations with weight control behaviors and psychological outcomes in adolescents. Pediatrics, 119(1) e30 7. Wang, Y., & Beydoun, M. A. (2007). The obesity epidemic in the united states -gender, age, socioeconomic, racial/ethnic, and geographic characteristics: A systematic review and meta regression analysis. Epidemiologic Reviews, 29, 628. doi:10.1093/epirev/mxm007

PAGE 97

97 Wang, Y. C., Bleich, S. N., & Gortmaker, S. L. (2008). Increasing caloric contribution from sugar sweetened beverages and 100% fruit juices among US children and adolescents, 19882004. Pediatrics, 121(6), e160414. doi:10.1542/peds.20072834 Weinstock, H., Berman, S., & Cates, W.,Jr. (2004). Sexually transmitted diseases among american youth: Incidence and prevalence estimates, 2000. Perspectives on Sexual and Reproductive Health, 36(1), 610. doi:10.1363/psrh.36.6.04 Werch, C. E., Ames, S., Moore, M. J., Thombs, D., & Hart, A. (2009). Health behavior insights: The transtheoretical/stages of change model: Carlo C. DiClemente, PhD. Health Promotion Practice, 10(1), 41 48. doi:10.1177/1524839908323519

PAGE 98

98 BIOGRAPHICAL SKETCH Chung Bang Ben Weng was born in 1974 in Chia Yi, Taiwan. Ben grew up in southern Taiwan and graduated from Chia Yi High School in 1992. Ben attended the Fu Jen Catholic University (FJU) and earned a Bachelor of Physical Education degree in 1996. In summe r 1999, Ben enrolled at the Central Michigan University pursuing Master of Art in e xercise s cience. Upon graduation, Ben continued his study at the Illinois Institute of Technology and earned a Master of Science in c omputer s cience degree in 2002. Ben star ted teaching website design, database, and computer network at Northern Taiwan Institute of Technology since 2003. Ben also taught badminton, golf, volleyball, and several other P.E. courses in FJU and Ming Chun University since 2004. Bens teaching experi ence in Taiwan led him to pursue the Ph.D. degree in health related field. Ben entered the Ph.D. program in the Department of Health Education and Behavior at the University of Florida (UF) in 2007. During his first year at UF, Ben explored a range of he alth behaviors and clarified his research focus. He developed a line of research investigating weight control behaviors among young people. Bens doctoral dissertation includes three scientific manuscripts readily to submit for peer review in scholarly jou rnals. Ben receive d a Doctor of Philosophy degree in h ealth and h uman performance with an emphasis in h ealth b ehavior and a minor in s tatistics through the College of Health and Human Performance in December of 2011.