Randomized, Controlled Trial of a Best-Practice Individualized Behavioral Program for Treatment of Childhood Overweight: Scottish Childhood Overweight Treatment Trial (SCOTT)
OBJECTIVE. The objective of this study was to determine whether a generalizable best-practice individualized behavioral intervention reduced BMI z score relative to standard dietetic care among overweight children.
METHODS. The design consisted of an assessor-blinded, randomized, controlled trial involving 134 overweight children (59 boys, 75 girls; BMI ≥ 98th centile relative to United Kingdom 1990 reference data for children aged 5–11 years) who were randomly assigned to a best-practice behavioral program (intervention) or standard care (control). The intervention used family-centered counseling and behavioral strategies to modify diet, physical activity, and sedentary behavior. BMI z score, weight, objectively measured physical activity and sedentary behavior, fat distribution, quality of life, and height z score were recorded at baseline and at 6 and 12 months.
RESULTS. The intervention had no significant effect relative to standard care on BMI z score from baseline to 6 months and 12 months. BMI z score decreased significantly in both groups from baseline to 6 and 12 months. For those who complied with treatment, there was a significantly smaller weight increase in those in the intervention group compared with control subjects from baseline to 6 months. There were significant between-group differences in favor of the intervention for changes in total physical activity, percentage of time spent in sedentary behavior, and light-intensity physical activity.
CONCLUSIONS. A generalizable, best-practice individualized behavioral intervention had modest benefits on objectively measured physical activity and sedentary behavior but no significant effect on BMI z score compared with standard care among overweight children. The modest magnitude of the benefits observed perhaps argues for a longer-term and more intense intervention, although such treatments may not be realistic for many health care systems.
There has been a rapid rise in the prevalence of childhood overweight in recent years. Overweight is now 1 of the most common pediatric health problems,1 and it has significant adverse effects on physical and psychosocial health in childhood and adulthood.2
Systematic reviews3–5 reported that previous randomized, controlled trials (RCTs) of childhood overweight treatment had methodologic flaws, such as small sample sizes, high dropout rates, short-term follow-up, lack of details about the randomization process, lack of blinding, and failure to use intention-to-treat analysis. Intensive, behavioral treatment programs for overweight children have proved successful in clinical studies from 1 center in the United States3,4; however, because these interventions were intense and required a number of health professionals, they may not be readily generalizable to all health care systems. Systematic reviews have therefore concluded that high-quality studies that test more generalizable interventions to treat childhood overweight are needed urgently.3–5
These systematic reviews identified a number of promising treatment strategies that can now be regarded as evidence based: use a theoretical basis for treatment, direct treatment at motivated families, involve the entire family in treatment, aim for weight maintenance rather than weight loss, be more intensive in treatment (more frequent and longer appointments), and combine dietary modification with changes in physical activity and/or sedentary behavior (eg, television viewing).3–5 We developed a behavioral intervention, based on recommendations from systematic reviews3–5 and expert committee guidance,6 that was intended to treat overweight children within a health care setting with limited resources. Our intervention used a family-centered approach7,8 and behavior change techniques8–10 to manage childhood overweight and can be regarded as a best-practice approach on the basis of current evidence. To our knowledge, this is the first RCT to implement best-practice recommendations in a relatively low-intensity, office-based setting using a single health professional, thus making the intervention practical and generally applicable to a range of health care settings.
The primary aim of this assessor-blinded RCT was to test the efficacy of our generalizable, best-practice individualized behavioral intervention relative to standard dietetic care (control condition) to reduce BMI z score among overweight children (BMI ≥ 98th centile relative to United Kingdom 1990 reference data) aged between 5 and 11 years. We also determined the impact of the intervention relative to the control condition on weight, objectively measured physical activity and sedentary behavior, fat distribution, growth, and quality of life (QoL). Design, conduct, and reporting of the trial followed Consolidated Standards of Reporting Trials (CONSORT) guidelines.11
This study was conducted at the Royal Hospitals for Sick Children in Glasgow and Edinburgh, Scotland. Eligibility criteria were overweight children (BMI ≥ 98th centile relative to United Kingdom 1990 reference data, usually referred to as obesity in the United Kingdom)3,4 who were aged 5 to 11 years and attending a standard elementary school and had at least 1 parent who perceived the child's weight as a problem and was willing to make lifestyle changes. We excluded children who had an underlying medical cause for their overweight or serious comorbidity that required urgent treatment or who had received treatment for overweight in the past year. Overweight children were recruited from dietetic waiting lists and were referred from hospital doctors, family physicians, school nurses, community dietitians, and community pediatricians in Glasgow and Edinburgh. Ethical approval was obtained from the research ethics committee at each participating site. Written informed consent was obtained from all children and their parents/guardians.
Randomization and Concealment
Children attended a baseline assessment, where the researcher obtained consent, recorded baseline measurements, and assigned a study code. For ensuring concealment, the study code was sent to a statistician, who produced a computer-generated randomization list and allocated participants to the intervention or control group. Randomization was in blocks of 10 (ratio 1:1) and was stratified by gender and study center (Edinburgh or Glasgow). The statistician informed the research dietitians, who were delivering the intervention of the group allocation and who then informed participants of their groups. Participants commenced intervention or control treatments within 3 weeks of the baseline measurements.
The intervention has been published elsewhere.7 Briefly, this is a practical, best-practice behavioral program delivered by experienced pediatric dietitians who are trained in behavior change counseling8,12 on a 1-to-1 basis (ie, 1 dietitian saw 1 family). The program consisted of 8 appointments (7 outpatient visits and 1 home visit) during 26 weeks with a total patient contact time of ∼5 hours. The program used a family-centered approach whereby the child (and family) took control of his or her own lifestyle changes.8,13 We used various behavioral change techniques, guided by models of behavior change, to enhance the child's motivation for making lifestyle changes: exploring motivation to make changes, exploring pros and cons of change, identifying barriers to change, problem-solving barriers, goal-setting, rewards, self-monitoring, social support, and preventing relapse.8–10,12,14 Although these behavioral techniques were developed for adults, they are increasingly being used to elicit lifestyle changes in children15; however, the dietitians had to modify their explanation of these strategies to the children, particularly with younger children. The strategies were directed at the children, although parents and the dietitian helped the child (especially younger children) understand and engage with the behavioral techniques. An article describing our intervention in more detail has been published.7
Children were encouraged to alter their diet by using a modified traffic-light approach [reduce intake of foods high in fat and sugar (red), increase intake of fruit and vegetables (green)]5,16 increase their physical activity,17 and restrict their sedentary behavior (television viewing and playing computer/video games) to no more than 2 hours per day or the equivalent of 14 hours per week as is widely recommended.4,6 Because the intervention focused on behavior change rather than weight change, children were weighed only 3 times during the 6-month program.
All written materials were reviewed by several pediatric dietitians throughout Scotland. The dietitians received 18 hours of training on behavioral change counseling, conducted selected reading,8 and practiced the behavior change techniques and written materials with a small number of obese children before delivering the intervention in the main trial. For ensuring that the core skills and behavioral techniques involved in the intervention were being correctly implemented by the dietitians, 6 patient consultations from the main trial were assessed by 2 independent experts on behavior change counseling (not involved in the training). Assessments were scored 1 (minimum) to 7 (maximum); scores were 5 to 6 for global rating (patient-centeredness), 6 to 7 for core conditions (empathy, genuineness, and acceptance), 5 to 6 for guiding principles (client responsibility, social influence, and collaboration), 5 to 7 for affirmation, and 4 to 7 for pace of interviews. The assessment concluded that both dietitians were highly skilled in these techniques.
Children who were randomly assigned to the control group received typical dietetic care currently are offered for overweight individuals by hospital and community dietetic services in Scotland. This involved 3 to 4 outpatient appointments delivered by pediatric dietitians during 6 to 10 months with a total patient contact time of ∼1.5 hours. Standard care did not reflect best practice, because it was very low intensity, concentrated on dietary change with minimal focus on physical activity or sedentary behavior, and involved a didactic “medical model” rather than a behavioral, client-centered approach. In addition, advice on weight management was mainly directed toward the parents rather than the child, as described elsewhere.7 A process evaluation of the dietitians who conducted standard care confirmed this description of typical dietetic treatment. Table 1 describes the differences between standard care and our best practice, behavioral intervention. Weight maintenance was the aim of both treatment conditions.3,4,6
Outcomes and Blinding
Outcome measures were recorded at baseline and then at 6 months and 12 months after the start of treatment by the same trained researcher, who was blinded to group allocation throughout the trial. Measures were put in place to ensure blinding, and the researcher had to report incidents of possible unblinding, which occurred in <5% of participants.
Our primary outcome was BMI z score. BMI was calculated by measurement of height to 0.1 cm and weight to 0.1 kg in duplicate using a portable stadiometer (Leicester Height Measure; Child Growth Foundation, London, England) and portable scales (TANITA 300GS; Cranlea & Co, Birmingham, England) with children in light indoor clothing and no socks and shoes. BMI and height were expressed relative to United Kingdom 1990 population reference data as z scores.18 Waist circumference was expressed relative to United Kingdom reference data19 as a z score to provide an index of fat distribution.
We measured habitual physical activity and sedentary behavior objectively for 7 days during all waking hours using the CSA/MTI WAM-7164 accelerometer (Manufacturing Technology Inc, Fort Walton Beach, FL), as previously described.20,21 Activity data were summarized as total physical activity (accelerometer count per minute) and percentage of waking hours in sedentary behavior, light-intensity physical activity, and moderate to vigorous physical activity using cut points validated against direct observation and energy expended during free-living activities for children.22,23
We assessed QoL of participating children by using the Pediatric Quality of Life Inventory 4.0, which provides valid and reliable assessments24,25 and includes parallel parent proxy reports and child self-report versions for those aged 5 to 7 and 8 to 12 years. We categorized socioeconomic status using a standard area-based measure, the deprivation category based on Carstairs scores from the 2001 Scottish census.26 The amount of time spent by the intervention and control dietitians interviewing children, administration after the interviews, and traveling for home visits (intervention only) and training costs (intervention only) were recorded to estimate economic costs of novel treatment and standard care.
Sample Size and Power
We used a between-group difference of −0.25 in the change in BMI z score over 6 months. Before the study, we assumed an SD of change in BMI z score of 0.4, but observed SD for change in BMI z score was 0.21, giving a δ of 1.15 (0.25/0.21). Thus, with power 0.90, 0.05 significance level, and a 2-sided test, the required sample size was 34 per group. For allowing for dropout, 134 participants were entered at baseline; dropout at 6 months was 27.6%, thus 97 participants were included in the analysis of the primary outcome (BMI z score) at 6 months. Our achieved power was 0.9999 for detection of a difference in BMI z score between groups of −0.25 over 6 months.
Not all outcome measures were normally distributed. Differences between groups for the median change in outcomes from baseline to 6 and 12 months were analyzed using Mann-Whitney tests. Categorical data were analyzed by using χ2 tests. The primary analysis was performed on an intention-to-treat basis for each outcome measure and involved all participants who attended for follow-up measures, regardless of whether they completed the treatment. We also performed a planned per-protocol analysis for BMI z score and weight (kilograms) using only intervention and control participants who complied well with the assigned treatment, defined as those who attended ≥75% of scheduled appointments.
Flow of Participants Through the Trial
Figure 1 describes the flow of participants through the trial. Participants were recruited from June 2003 to June 2004, and outcome measures were recorded at baseline and then at 6 months (median: 27 weeks; interquartile range [IQR]: 26–28) after the start of treatment and 12 months after the start (median: 54 weeks; IQR: 52–55). Of the 237 children assessed for eligibility, 134 (56.5%) consented and were randomly assigned to the intervention (n = 69) or control group (n = 65). Of the 134 children who were randomly assigned, 97 (72.4%) attended the 6-month follow-up and 86 (64.2%) attended at 12 months (Fig 1); the primary outcome (BMI z score) and weight and height z score were available for all of these participants. Follow-up in each group was similar at 6 months (χ2 = 0.1, degrees of freedom [df] = 1, P = .7) and 12 months (χ2 = 0.07, df = 1, P = .8). Baseline characteristics of study participants are shown in Table 2; there were no significant differences between groups at baseline. The median BMI z score at baseline was >3 in both groups.
Physical activity and sedentary behavior measurements were obtained in 117 participants at baseline and 67 at 6 months; however, compliance with wearing the accelerometer was very poor at 12 months; therefore, we did not have sufficient data to perform an analysis at 12 months. QoL scores from baseline to 6 months were analyzed for 94 parent proxy reports and 92 child self-reports.
Of the 69 participants assigned to the intervention, 44 (63.8%) complied well with treatment, defined as attendance at ≥75% of scheduled appointments. Of the 65 participants assigned to the control condition, 46 (70.8%) attended ≥75% of scheduled appointments (χ2 = 0.75, df = 1, P = .39).
Primary Outcome: BMI z Score
There were no significant differences between the intervention and control groups for changes in BMI z score and weight (kg) from baseline to 6 and 12 months (Table 3). BMI z score significantly decreased in both groups from baseline to 6 months (intervention: 95% confidence interval [CI]: −0.18 to −0.07; control: 95% CI: −0.16 to −0.03) and baseline to 12 months (intervention: 95% CI: −0.22 to −0.04; control: 95% CI: −0.26 to −0.08). Weight (kg) significantly increased in both groups from baseline to 6 months (intervention: 95% CI: 2.2–3.6; control: 95% CI: 2.8–4.4) and 12 months (intervention: 95% CI: 5.4–7.8; control: 95% CI 5.5–7.7). Table 4 presents actual median values for BMI z score and weight at baseline and at 6 and 12 months.
There were no significant between-group differences for changes in waist circumference z score from baseline to 6 and 12 months (Table 5).
Objectively Measured Physical Activity and Sedentary Behavior
At baseline, the proportion of monitored time spent in sedentary behavior was high and participation in moderate to vigorous physical activity was low in both groups (Table 2). The median duration of activity monitoring over 7 days was 11.3 hours/day (IQR: 10.4–12.2) in the intervention group and 11.5 hours/day (IQR: 10.6–12.7) in the control group. There were significant between-group differences for the change in total activity (mean counts per minute) and percentage of time spent in sedentary behavior and light-intensity activity from baseline to 6 months in favor of the intervention group (Table 5).
Other Secondary Outcomes
There were no significant between-group differences for the change in height z score from baseline to 6 and 12 months (Table 5). Similarly, no significant between-group differences were found for changes in QoL scores for the child self-report or parent proxy report from baseline to 6 months (Table 5). Parent-reported QoL scores significantly improved from baseline to 6 months in both groups (Table 5). The cost (for 1 patient) of delivering the novel intervention was £108 ($192 US) and £29 ($52 US) for the standard treatment.
We also performed a planned per-protocol analysis for BMI z score and weight using only intervention and control participants who complied with the assigned treatment, defined as those who attended ≥75% of scheduled appointments; 84 participants were included in analysis at 6 months, and 77 were analyzed at 12 months. The intervention had no significant effect on BMI z score relative to control subjects from baseline to 6 months (Mann-Whitney test, 95% CI: −0.02 to 0.14; P = .2) and 12 months (Mann-Whitney test, 95% CI: −0.14 to 0.11; P = .8). There was a significantly smaller increase in weight in the intervention group compared with the control subjects from baseline to 6 months (Mann-Whitney test, 95% CI: 0.05–2.25; P = .04) for participants who complied with treatment, but this effect was not evident at 12 months (Mann-Whitney test, 95% CI: −1.5 to 1.9; P = .8). Median weight increase was 2.9 kg (IQR: 1.1–4.1) in the intervention group and 4.0 kg (IQR: 2.4–5.6) in the control group.
The generalizable best-practice individualized behavioral program that was tested in this study had modest benefits for BMI z score, for objectively measured physical activity and sedentary behavior, and for QoL. Furthermore, for participants who complied well with treatment, weight gain was significantly lower in the intervention group compared with the control subjects from baseline to 6 months. However, we observed no significant differences between the best-practice approach and standard care for BMI z score, weight, and fat distribution over 12 months among overweight children of elementary school age. The significant benefits in physical activity and sedentary behavior in favor of the best-practice intervention group may reflect differences in treatment targets: our intervention focused on diet, physical activity, and sedentary behavior, whereas standard care had minimal emphasis on physical activity and did not target sedentary behavior. This study therefore provides some evidence that inclusion of these behavioral targets as part of treatment is worthwhile, although the changes in activity and sedentary behavior actually observed were modest. Information on dietary intake was not collected in this study but may have added to our understanding of the lifestyle changes observed.
There is widespread concern, particularly from parents, that treating overweight children may increase the risk for adverse effects; however, research in this area is limited.4,27 We found that our best-practice program, which was family-centered and intensive, did not adversely affect the child's growth or QoL. There were significant improvements in health-related QoL over time in both groups in this study. We previously showed that QoL is impaired in our clinical sample of overweight children compared with lean control subjects matched for age, gender, and socioeconomic status.28 Improvements in QoL in this study moved overweight children's QoL scores into the range for healthy lean children.28
It is possible that both treatments may have been successful relative to a no-treatment control group and that our decision to use a standard care control group attenuated differences between groups. Previous studies in the United Kingdom showed that BMI z scores continued to increase in overweight children who did not receive treatment.29,30 In contrast, in this study, we found small but significant decreases in BMI z scores over 12 months in both groups; however, the clinical significance of these changes is unclear because evidence suggests that a decrease in BMI z score of 0.5 over 1 year may be needed to improve cardiovascular and metabolic risk factors in overweight children.31 It should also be noted that other studies that used a waiting list control group found significant improvements in anthropometric outcomes in the control condition.32 It is possible that the detection of obesity through recruitment and baseline assessment may make families aware of the issue and motivate them to change behavior. The mean BMI z score of our sample was >3, indicating the degree of overweight was extreme and, therefore, families in this study may have been more resistant to treatment.
Systematic reviews have recommended that future treatment programs be both generalizable and evidence-based, using the elements of treatment likely to be most effective from previous studies.3–5,33 In addition, treatment programs that use a more behavioral approach to changing lifestyle in children are more likely to be successful in the treatment of overweight and other chronic childhood diseases than more traditional, didactic approaches to treatment6,9,10; therefore, we used these recommendations to develop a generalizable, best-practice behavioral intervention delivered by a single pediatric dietitian in an office-based setting, thereby making the manpower burden and treatment costs generalizable (less than $200 per patient). We also ensured that the behavioral approach used in the intervention was of a very high quality. The intervention group dietitians were highly trained in behavior change counseling, and several consultations were assessed by 2 independent experts in health behavior change counseling.
We rigorously tested the efficacy of this intervention in a high-quality RCT that followed Consolidated Standards of Reporting Trials guidelines. In contrast to many previous studies included in systematic reviews of childhood obesity treatment,3–5,33 this study was adequately powered, follow-up was longer than most previous trials, compliance to both intervention and standard care treatments was good and retention rates at follow-up assessments were high, outcomes were recorded by a blinded researcher, habitual physical activity and sedentary behavior were measured objectively, and we used intention-to-treat analysis and properly conducted randomization procedures. We maximized generalizability by recruiting obese children from a number of health professionals in the 2 major cities in Scotland. In addition, the intervention tested was generalizable (as previously discussed), and the control condition accurately reflected current treatment of overweight children in Scotland.
It is possible that a more intense and/or longer duration intervention than that used in this study may have been more successful, but our aim was to test an intervention that was practical and so likely to be generally applicable. Alternatively, the family-centered approach used in the intervention may have allowed families to set lifestyle goals that were not substantial enough to have a marked effect on BMI z score.
The generalizable, best-practice individualized behavioral intervention that was tested in this study had modest benefits on objectively measured physical activity and sedentary behavior. Furthermore, the intervention had a positive effect on weight for those who complied with the program. Both treatments had a small but significant effect on BMI z score over the 12 months. The modest magnitude of the benefits observed perhaps argues for a longer term and even more intense approach to treatment of pediatric overweight, although such treatments may not be realistic for many health care systems. These findings may be useful in the development of future treatment programs.
This work was supported by a grant from the Scottish Executive Health Department. The funders had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript. The funder's role was limited to peer review of the original grant application.
We are immensely grateful to the study participants and their families. We thank the following people for assisting with patient recruitment: the doctors and dietitians at the Royal Hospitals for Sick Children in Glasgow and Edinburgh; community dietitians; school nurses and health visitors in Glasgow and Edinburgh; and doctors and dietitians at St John's Hospital in Livingston, Wishaw General Hospital, and Falkirk Royal Infirmary.
- Accepted August 15, 2007.
- Address correspondence to Adrienne R. Hughes, PhD, Department of Sports Studies, University of Stirling, Scotland FK9 4LA, United Kingdom. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
Dr Reilly is guarantor; Drs Reilly and Hughes had full access to data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis; all authors were involved in obtaining funding, conception, and design of the study, interpretation of the data, critical revision of the manuscript, and final approval of the submitted manuscript; Dr Reilly was involved in developing the intervention, study supervision, and drafting the manuscript; Dr Hughes was involved in developing the intervention, drafting the manuscript, collecting all outcome measures, and analyzing the data; Ms Stewart and Ms Chapple developed and delivered the intervention and drafted the manuscript; and Mr McColl was the trial statistician and participated in randomization procedures and supervising data analysis.
- ↵Reilly JJ, Methven E, McDowell ZC, et al. Health consequences of obesity. Arch Dis Child.2003;88 (9):748– 752
- ↵Summerbell CD, Ashton V, Campbell KJ, Edmonds L, Kelly S, Waters E. Interventions for treating obesity in children. Cochrane Database Syst Rev.2003;(3):CD001872
- ↵Scottish Intercollegiate Guidelines Network (SIGN). Management of obesity in children and young people: a national clinical guideline. SIGN 69:2003. Available at: www.sign.ac.uk/guidelines/fulltext/69/index.html. Accessed September2006
- ↵Rollnick S, Mason P, Butler C. Health Behavior Change: A Guide for Practitioners. London, United Kingdom: Churchill Livingston;1999
- ↵Rollnick S, Butler C, McCambridge J, Kinnersley P, Elwyn G, Rescnicow K. Consultations about changing behaviour. BMJ.2005;331 (7522):961– 963
- ↵Cavill N, Biddle S, Sallis JF. Health enhancing physical activity for young people: statement of the United Kingdom Expert Consensus Conference. Pediatr Exerc Sci.2001;13 :12– 25
- ↵Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, 1990. Arch Dis Child.1995;73 (1):25– 29
- ↵McLoone P. Carstairs Scores for Scottish Postcode Sectors From the 2001 Census. Glasgow, United Kingdom: Public Health Research Unit, University of Glasgow;2004
- ↵Quattrin T, Liu E, Shaw N, Shine B, Chiang E. Obese children who are referred to the pediatric endocrinologist: characteristics and outcome. Pediatrics.2005;115 (2):348– 351
- ↵Rudolf M, Christie D, McElhone S, et al. WATCH IT: a community-based programme for obese children and adolescents. Arch Dis Child.2006;91 (9):736– 739
- ↵Reinehr T, Andler W. Changes in the atherogenic risk factor profile according to degree of weight loss. Arch Dis Child.2004;89 (5):419– 422
- ↵Golley RK, Magarey AM, Baur LA, Steinbeck KS, Daniels LA. Twelve-month effectiveness of a parent-led, family-focused weight-management program for prepubertal children: a randomized controlled trial. Pediatrics.2007;119 (3):517– 525
- ↵Reilly JJ, Wilson ML, Summerbell CD, Wilson DC. Obesity: diagnosis, prevention, and treatment—evidence based answers to common questions. Arch Dis Child.2002;86 (6):392– 394
- Copyright © 2008 by the American Academy of Pediatrics