Psychological Predictors of Success in a Pediatric Cognitive-Behavioral Weight-Control Program

Texas Medicine Magazine Logo  

Symposium on Adult Diseases in Children - February 2009  


Tex Med. 2009;105(2):25-32.  

By Carmen Mikhail, PhD; A. Scott Raynaud, PhD; Vernisha Y. Shepard, Med; Peter Nieman, FRCP(C); Diana Arceo, MD; and William J. Klish, MD  


The incidence of childhood obesity has increased dramatically in recent years, 1-3  leading some experts to use words such as "epidemic proportions." 4 The most recent National Health and Nutrition Examination Survey estimates that 16.3 percent of American children and adolescents are now overweight. 5 The prevalence of obesity in children has tripled between 1980 and 2000 and continues to increase across all age groups. 6 Another alarming trend is the increased prevalence of obesity across all racial-ethnic groups. 7 Additionally, childhood obesity leads to a higher risk for adult obesity, and adults will have more severe obesity if their obesity begins in childhood. 8-10

The increase in morbidity associated with the current progressive incidence of childhood obesity has lead to serious public health concerns. 11 In addition, the psychological costs need to be addressed. These consequences pertain to both the obese child and his or her family.

Obese children encounter stigmatization because of the existing cultural aversion toward fatness and preference for a slim physique. They suffer from depression and poor self- and body-esteem, and their mothers are in greater psychological distress. 12-14 However, some studies show that obese children may not have any more psychological problems than normal-weight children. 15-17 Assessment of psychological profiles of obese children and their families might help to elucidate factors related to successful participation in weight-control programs.

To date, research on psychological predictors of success in pediatric weight control is sparse. Greater compliance has been related to obese adolescents' beliefs regarding personal control over weight 18 and to positive parental expectations. 19 Program dropout has been linked to subjects' self-reported depression and poorer self-concept. 20 Data are needed from researching unselected subjects, i.e., where subjects are randomly entered in a study. Identifying predictors of success may help the practicing clinician focus on factors that determine a successful outcome for children and their families enrolled in a weight-control program. Therefore, the purpose of this study is to explore the relationship between psychological profiles and success in a pediatric cognitive-behavioral weight-control program.



Initially considered for analysis were 477 overweight subjects who completed questionnaires for the "A Weigh of Life" weight-control program between 1992 and 1999. Overweight was defined as being at 110 percent ideal body weight for height (percent IBW) or greater. Primary care physicians, staff physicians, and school nurses referred the subjects. Demographically, the sample consisted of 297 girls and 180 boys. Of the total subjects, 250 were white, 123 African American, 92 Hispanic, and 12 other. The mean age was 11 years (range, 6.0-19.2 years). Patients were on average 177.7 ± 40.8 percent IBW with an average BMI of 34.1 ± 10.4 at the beginning of the program.  Figure 1 displays the initial mean percent IBW by race and sex.

Before starting the program, obese children and their mothers or guardians completed questionnaires that assessed the child's psychological and behavioral profile and child and parental attitudes toward dieting and exercise. The measures were selected based on clinical experience and literature review.

Children completed the following questionnaires:

  • Children's Depression Inventory, 21  assessing several indices of childhood depression including mood, vegetative functions, self-evaluation, and interpersonal behaviors. The depression inventory has been shown to have good reliability and validity. 21,22
  • Body-Esteem Scale, 23 measuring the child's attitudes toward his or her physical appearance. The scale has an odd-even split-half reliability of r  =.85 and shows concurrent validity with the Physical Attributes subscale of the Piers-Harris Self-Concept Scale ( r =.67).
  • Piers-Harris Children's Self-Concept Scale, 24 evaluating six aspects of self-esteem: behavior, intellectual status, physical appearance and attributes, anxiety, popularity, and happiness and satisfaction. The internal consistency of the test is high, with alpha coefficients of .90 to .91. Additionally, reliabilities of .88 to .93 have been cited using the Kuder-Richardson formula.
  • Children's Health Locus of Control Scale, 25  measuring the extent to which the child believes that his or her health is the result of self-actions versus external forces such as medical personnel, fate, luck, or random events. Those with an "internal locus of control" believe that health is the result of self-actions whereas those with an "external locus of control" believe that external forces determine their health. Findings provide evidence that the Children's Health Locus of Control Scale has acceptable levels of reliability, internal consistency, and construct validity. A higher score indicates a more internal locus of control.
  • Dieting Beliefs Scale for Children, based on the Dieting Beliefs Scale, 26 measuring locus of control specific to weight loss and adapted for children for this study by simplifying the language. This scale assesses the expectancy that one can control one's weight. Validity and reliability have not been determined for a pediatric population.

Parents completed the following questionnaires:

  • Family Adaptability and Cohesion Evaluation Scales-III, 27 in which parents evaluate how they would like the family to be ideally and how they currently perceive the family. The factors assessed are adaptability and cohesion. The alpha reliability is .62 for adaptability, .77 for cohesion, and .68 for the total scale.
  • Child and Adolescent Adjustment Profile, 28 in which parents evaluate their child's adjustment in areas of peer relations, dependency, hostility, productivity, and withdrawal.
  • Dieting Beliefs Scale, 26 which measures locus of control specific to weight loss.
  • Restrained Eating Scale, 29 which assesses dietary restraint, or the extent to which one vigilantly monitors and restricts one's food intake, leading to feelings of deprivation.

Weights and heights were measured by using an electronic strain gauge scale and stadiometer, respectively. Percentage of ideal body weight for children was calculated by using height and weight norms that were applicable at the time the data were collected. 30



Overweight subjects (n=477) completed questionnaires and had a screening interview by a clinical psychologist. Following this, the subjects entered the A Weigh of Life weight-control program at the Eating Disorders Clinic, Section of Pediatric Gastroenterology, Hepatology, and Nutrition at Texas Children's Hospital in Houston. The program consisted of 15 weekly lessons, given as one-on-one sessions conducted by a master's level psychotherapist, a registered dietitian, and an exercise therapist.

The program emphasized cognitive-behavior therapy, dietary education, and an exercise plan. Behavior modification methods included self-monitoring, identifying behavioral chains, goal setting and reinforcement, stimulus control, and using inhibitors.

Cognitive strategies included increasing awareness of hunger, cognitive restructuring, problem solving and planning, and eliciting social support. Subjects were given treatment manuals and weekly homework assignments. They were offered optional weight maintenance sessions upon completion of the program.

The program is a fee-for-service program, although insurance typically covered the screening interview by a psychologist.


Statistical Analyses  

Data were analyzed by Stata 31 and SAS 32 statistical software. Because it was desired to examine ethnic differences, only those ethnic groups with adequate data were retained (African-American, Hispanic, and white). Cases with excessive missing data were deleted and for the remainder, missing data were imputed using an algorithm in Stata. Of those with missing data, less than 5 percent of cases received imputed values.

For the final analysis, 413 subjects were retained. Of those, all the parents had questionnaire data, but only 386 children had questionnaire data; some were too young or mentally unable to complete the questionnaires. Parents provided report of the child's ability to understand and complete questionnaires independently.

Both dependent and independent variables were considered:

  • Independent variables for the analyses were the following: Children's self-rated variables included depression, body-esteem, self-esteem, health locus of control, and dieting beliefs. Family variables assessed included perceived cohesion, ideal cohesion, perceived adaptability, and ideal adaptability. Children's variables as rated by the parent included peer relations, dependency, hostility, productivity, and withdrawal. Parent variables included dieting beliefs and dietary restraint.
  • The dependent variable of interest is weight loss success as measured by percent change in ideal body weight. Covariates included sex, age, race/ethnicity, and baseline percent IBW. Age was chosen because developmental processes are thought to be ongoing in children and result in cognitive, behavioral, and physical changes that could affect the children's responses to the questionnaires. Age is related to child locus of control, child depression, and child body self-esteem and should therefore be controlled in any analysis. 33



Physical Findings (Weight Loss)
Of the 477 persons completing questionnaires, 390 appeared for a screening interview, 337 began the program, and 99 finished the program.

For those with complete entry data (n = 413), boys (percent IBW = 186.5 ± 42.8) were significantly heavier than girls (percent IBW = 171.7 ± 35.8) ( P <.001) at baseline. African-Americans (percent IBW = 194.3 ± 46.8) and Hispanics (percent IBW = 184.2 ± 38.9) were significantly heavier than whites (percent IBW = 166.3 ± 34.3) ( P <.0005) at baseline.

Of the 99 finishers, 88 met the inclusion criteria for the study (complete data). For this group of finishers, the mean percent IBW was 175.7 ± 38.0 at baseline and 154.9 ± 32.7 at post-treatment. Subjects who completed the entire program lost an average of 20.7 ± 13.5 percent IBW. This loss was highly significant ( P <.00005). Mean rate of weight loss was -0.31 ± 0.27 kg/wk or -0.15 ± 0.10 BMI units. There were no sex or race differences in weight loss.

Many finishers continued with additional sessions beyond the 15 lessons of the program. Those who continued into the maintenance phase of the program attended a total average of 21.2 ± 9.8 sessions and achieved a final decrease in percent IBW of 22.2 ± 16.5 percentage points. Of all the starters (finishers plus dropouts) with complete questionnaire data (n = 262), patients attended an average of 10.5 ± 9.6 sessions and achieved a drop of 10.9 ± 15.7 percentage points in their percent IBW.

To evaluate the weight loss of dropouts, we examined the reduction in percent IBW by number of sessions completed (1-13). Those who only completed two sessions or less gained weight but after that, during treatment, there was a steady pattern of weight loss by dropouts.  Figure 2 presents those results. It shows that many of those dropping out before the end of treatment were still losing weight.

In summary, 29 percent of starters finished the entire program. Patients attended the majority of sessions and were able to lose weight whether or not they finished. The achievements were impressive for those who finished and continued beyond the completion of the program. Those subjects who lost weight successfully did not differ by sex or race.

Psychological Findings
The  Table shows raw data findings from psychological questionnaires completed by children and parents. Test scores are presented for females versus males, as well as for African-Americans versus Hispanics versus whites. Our subjects were generally higher than the norm on depression and lower than the norm on self-esteem and body esteem. They were slightly higher in health locus of control, i.e., more internal in their beliefs. Our parents were higher than the norm in restrained eating practices.

To explore the effect of psychological variables on program completion, participants were divided into three groups representing their final status in the study:

  1. Nonstarters, who completed questionnaires and/or screening only.
  2. Nonfinishers, who attended some of the A Weigh of Life sessions.
  3. Finishers, who completed all the sessions. We found an effect for parental restraint (F[2,410] = 3.93, P <.05). Those finishing the program had parents who were less restrained in their own eating than those not starting the program, when controlling for percent IBW.

Two regression analyses determined which variables predicted decrease in percent IBW. The first analysis included 247 subjects who started treatment and had complete questionnaire data, predicting decrease in percent IBW with variables obtained from child questionnaires, age, and number of treatment plus maintenance sessions completed. The regression was run on ranked data since the distributions were skewed. Correlations between decrease in percent IBW and psychological variables are:

  • Age: r = 0.01, P >.05;
  • Depression: r = -0.03, P >.05;
  • Self-esteem: r = -0.11, P >.05;
  • Health locus of control: r = 0.05, P >.05;
  • Body esteem: r = -0.06, P >.05;
  • Dieting beliefs: r = 0.02, P >.05;
  • Number of sessions: r = 0.40, P <.01.

The analysis revealed a model with self-esteem and number of sessions as the best model (F[2,244] = 82.13, P <.0001, R 2 = .40). The coefficients for both self-esteem (t[1] = 2.48, P <.01) and the number of sessions (t=[1] = 12.55, P <.0001) were significantly different from zero. Those attending more sessions (β = .62) and those with lower self-esteem (β = -.10) achieved better outcomes. The regression equation was: Decrease in percent IBW = 218.23 + .62 x number of sessions − .10 x self-esteem.

The next analysis predicted change in percent IBW with variables obtained from parent questionnaires, along with the child's age and number of sessions completed, and included 262 subjects. Correlations between decrease in percent IBW and psychological variables are:

  • Age: r = 0.07, P >.05;
  • Perceived cohesion: r = 0.02, P >.05;
  • Ideal cohesion: r = 0.04, P >.05;
  • Perceived adaptability: r = -0.12, P >.05;
  • Ideal adaptability: r = -0.04, P >.05;
  • Dieting beliefs: r = -0.02, P >.05;
  • Peer relations: r = -0.04, P >.05;
  • Dependency: r = 0.06, P >.05;
  • Hostility: r = 0.08, P >.05;
  • Productivity: r = -0.00, P >.05;
  • Withdrawal: r = 0.00, P >.05;
  • Restraint: r = -0.03, P >.05;
  • Number of sessions: r = 0.41, P <.01.

The regression was run on ranked data since the distributions were skewed. This analysis revealed that a model including ideal adaptability, hostility, and number of sessions completed was the best model (F[3,258] = 61.73, P <.0001; R 2 = .41). The coefficients for ideal adaptability (t[1] = -2.58, P <.01) and number of sessions completed (t[1] = 13.25, P <.0001) were significant. The coefficient for hostility (β =1.89) was not significant (t[1] =0.07, P =.06). Those completing more sessions (β =.64) and wanting less adaptability, i.e., more structure at home (β = -.10) had better outcomes. The regression equation was: Decrease in percent IBW = 233.08 + .64 x number of sessions − 0.09 x ideal adaptability + 0.07 x hostility.



The present study examined treatment outcome as well as psychological predictors of success for children enrolled in a cognitive-behavioral weight-control program. Roughly one-third of our sample completed the program. If we define weight loss by a change in ideal body weight for height (percent IBW), then finishers achieved a significant percent IBW loss during treatment of 20.8 percentage points. Even when including dropouts, percent IBW fell 10.9 percentage points. Thus, treatment appears to have positive effects regardless of whether patients finish the entire program.

Parents of children completing the program were less restrained in their own eating than were parents of children not starting the program. Parents of children not starting the program scored in the restrained range, suggesting that they are hyper-vigilant about their own food intake. Thus, it appears that when parents have their own weight and food concerns, their children are less likely to start the program. Clinicians must make special efforts to encourage this subpopulation to pursue weight control in their children, as well as deal with their own issues surrounding food and weight. Although there have been reports of an association between parental feeding style and child eating and weight, 34 we are not aware of reports of the relation between parental restraint and child weight-control program dropout.

The most important predictor of decreased percent IBW was the number of sessions completed. This result is consistent with findings in the adolescent literature indicating that therapist contact is an important predictor of weight loss. 35 Thus, one way of improving treatment outcome is to extend treatment length. The results might also suggest that self-help aids may be less effective in weight control. However, in this nonexperimental study, the relation between length of treatment and weight loss could be due to factors inherent in the subjects before they started the program, such as motivation, ability to persevere, or the ability to work for long-term goals.

Weight loss in the program was linked to two psychological variables: self-esteem and family structure. Children with lower self-esteem initially achieved better outcomes. A speculative explanation is that poor self-esteem may reflect greater distress over one's obesity, which may act as a catalyst for change. In one study, adolescents who were initially rated by parents as being in more psychological distress, as measured by anxiety and depression, showed greater weight reductions in treatment. 35 However, in another study of pediatric weight loss, global self-esteem did not predict weight loss. 36 Additionally, a positive relation between self-esteem and successful weight loss has been demonstrated in adults. 37 It may be that self-esteem impacts motivation for weight change differentially for children than for adults - a reminder that we cannot generalize results across these two populations.

Parents who wished to have more structure at home had children who did better in our program. This is consistent with our clinical observations that parents who are capable of setting limits achieve better results. It is also consistent with results of studies indicating that structured parenting is related to a lower body mass index in children 38  and that children of permissive mothers are twice as likely to be overweight as those of authoritative mothers. 39  

Although these psychological variables predicted clinical outcomes, the best predictor was number of sessions attended. It could be that psychological variables are not as important in predicting success in pediatric weight control. However, our therapists, who have backgrounds in psychology or counseling, may have addressed psychological issues as they arose during treatment so that they did not impede success in the program.

A limitation of the present study is its generalizability. Subjects were all presenting for obesity treatment and as such may be more distressed about the child's weight and more motivated to change their lifestyle than those in a general population. Therefore, results can be generalized only to subjects self-selecting for cognitive-behavioral weight-control treatments. Additionally, there was no control group to test the outcome of not seeking treatment.




Obese children can lose significant weight in a cognitive-behavioral weight-control program . Even those dropping out prematurely may experience some benefit. Program completion is related to lack of parental dietary restraint. Weight loss is related most strongly to number of sessions attended. Weight loss also is related to psychological variables within the child and the family, including self-esteem of the child and adaptability of the family. These findings may direct clinicians to address these factors in their work with this population.




  1. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999-2004. JAMA . 2006;295(13):1549-1555.
  2. Li C, Ford ES, Mokdad AH, Cook S. Recent trends in waist circumference and waist-height ratio among US children and adolescents. Pediatrics . 2006;118(5):e1390-e1398.
  3. Ogden CL, Yanovski SZ, Carroll MD, Flegal KM. The epidemiology of obesity. Gastroenterology . 2007;132(6):2087-2102.
  4. Strauss RS, Pollack HA. Epidemic increase in childhood overweight, 1986-1998. JAMA . 2001;286(22):2845-2848.
  5. Ogden CL, Carroll MD, Flegal KM. High body mass index for age among US children and adolescents, 2003-2006. JAMA . 2008;299(20):2401-2405.
  6. Wyatt SB, Winters KP, Dubbert PM. Overweight and obesity: prevalence, consequences, and causes of a growing public health problem. Am J Med Sci . 2006;331(4):166-174.
  7. Whitaker RC, Orzol SM. Obesity among US urban preschool children: relationships to race, ethnicity, and socioeconomic status. Arch Pediatr Adolesc Med . 2006;160(6):578-584.
  8. Guo SS, Wu W, Chumlea WC, Roche AF. Predicting overweight and obesity in adulthood from body mass index values in childhood and adolescence. Am J Clin Nutr . 2002;76(3):653-658.
  9. Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH. Predicting obesity in young adulthood from childhood and parental obesity. N Engl J Med . 1997;337(13):869-873.
  10. Thompson DR, Obarzanek E, Franko D, et al. Childhood overweight and cardiovascular disease risk factors: the National Heart, Lung, and Blood Institute Growth and Health Study. J Pediatr . 2007;150(1):18-25.
  11. Wyatt SB, Winters KP, Dubbert PM. Overweight and obesity: prevalence, consequences, and causes of a growing public health problem. Am J Med Sci . 2006;331(4):166-174.
  12. Young-Hyman D, Schlundt DG, Herman-Wenderoth L, Bozylinski K. Obesity, appearance, and psychosocial adaptation in young African American children. J Pediatr Psychol . 2003;28(7):463-472.
  13. Young-Hyman D, Tanofsky-Kraff M, Yanovski SZ, Keil M, Cohen ML, Peyrot M, Yanovski JA. Psychological status and weight-related distress in overweight or at-risk-for-overweight children. Obesity Res . 2006;14(12):2249-2258.
  14. Zeller MH, Saelens BE, Roehrig H, Kirk S, Daniels SR. Psychological adjustment of obese youth presenting for weight management treatment. Obes Res . 2004;12(10):1576-1586.
  15. Rumpel C, Harris TB. The influence of weight on adolescent self-esteem. J Psychosom Res . 1994;38(6):547-556.
  16. Klesges R, Haddock CK, Stein R, Klesges LM, Eck LH, Hanson CL. Relationship between psychosocial functioning and body fat in preschool children: a longitudinal investigation. J Consult Clin Psychol . 1992;60(5):793-796.
  17. Epstein LH, Klein KR, Wisniewski L. Child and parent factors that influence psychological problems in obese children. Int J Eat Disord . 1994;15(2):151-158.
  18. Uzark KC, Becker MH, Dielman TE, Rocchini AP. Psychosocial predictors of compliance with a weight control intervention for obese children and adolescents. J Compliance Health Care . 1987;2(2):167-178.
  19. Uzark KC, Becker MH, Dielman TE, Rocchini AP, Katch V. Perceptions held by obese children and their parents: implications for weight control intervention. Health Educ Q . 1988;15(2):185-198.
  20. Zeller M, Kirk S, Claytor R, Khoury P, Grieme J, Santangelo M, Daniels S. Predictors of attrition from a pediatric weight management program. J Pediatr . 2004;144(4):466-470.
  21. Kovacs M. Children's Depression Inventory Manual . North Tonawanda, NY: Multi-Health System; 1992.
  22. Romano BA, Nelson RO. Discriminant and concurrent validity of measures of children's depression. J Clin Child Psychol . 1988;17(3):255-259.
  23. Mendelson BK, White DR. Relation between body-esteem and self-esteem of obese and normal children. Percept Mot Skills . 1982;54(3 pt 1):899-905.
  24. Piers EV. Piers-Harris Children's Self-Concept Scale . Los Angeles, CA: Western Psychological Services; 1984.
  25. Parcel GS, Meyer MP. Development of an instrument to measure children's health locus of control. Health Educ Monogr . 1978;6(2):149-159.
  26. Stotland S, Zuroff D C. A new measure of weight locus of control: The Dieting Beliefs Scale. J Pers Assess . 1990;54(1-2):191-203.
  27. Olson DH. Faces III (Family Adaptation and Cohesion Scales). St. Paul, MN. University of Minnesota; 1985.
  28. Ellsworth RB. Child and Adolescent Adjustment Profile (CAAP Scale). Bellevue, WA: Ellsworth Krebs Associates; 1997.
  29. Herman CP, Polivy J. Restrained eating. In: Stunkard AJ. Obesity. Philadelphia, PA: WB Saunders Co; 1980:208-225.
  30. Hamill PV, Drizd TA, Johnson CL, Reed RB, Roche AF. NCHS growth curves for children : birth-18 years, United States. Hyattsville, MD: US Dept of Health, Education, and Welfare, National Center for Health Statistics; 1977.
  31. Stata Corporation. Stata reference manual. Release 6. College Station, TX: Stata Corp; 1999.
  32. SAS Institute. Getting started with the SAS system using SAS/ASSIST software. Version 6. Cary, NC: SAS Institute; 1996.
  33. Friedman MA, Brownell KD. Psychological correlates of obesity: Moving to the next research generation. Psychol Bull . 1995;117(1):3-20.
  34. Faith MS, Scanlon KS, Birch LL, Francis LA, Sherry B. Parent-child feeding strategies and their relationships to child eating and weight status. Obes Res . 2004;12(11):1711-1722.
  35. Germann JN, Kirschenbaum DS, Rich BH, O'Koon JC. Long-term evaluation of multi-disciplinary treatment of morbid obesity in low-income minority adolescents: La Rabida Children's Hospital's FitMatters program. J Adolesc Health . 2006;39(4):553-561.
  36. Braet C. Patient characteristics as predictors of weight loss after an obesity treatment for children. Obes Res . 2006;14(1):148-155.
  37. Nir Z, Neumann L. Self-esteem, internal-external locus of control, and their relationship to weight reduction. J Clin Psychol . 1991;47(4):568-575.
  38. Chen J-L, Kennedy C. Family functioning, parenting style, and Chinese children's weight status. J Fam Nurs. 2004;10(2):262-279.
  39. Rhee KE, Lumeng JC, Appugliese DP, Kaciroti N, Bradley RH. Parenting styles and overweight status in first grade. Pediatrics . 2006;117(6):2047-2054.




Back to article  



Back to article  




Back to article  



Dr. Mikhail, Dr. Raynaud, Ms. Shepard, Ms. Arceo, and Dr. Klish are from the Department of Pediatrics, Baylor College of Medicine, Texas Children's Hospital, Houston; Dr Nieman is from the Department of Pediatrics, Faculty of  Medicine, University of Calgary, Calgary, Alberta, Canada.  



February 2009 Texas Medicine Contents
Texas Medicine Main Page