Design and aims
This study will use a staggered-entry, waitlist controlled clinical trial to assess the impact of a multidisciplinary intervention aiming to change the trajectory of overweight adolescents and thus help them avoid poor health in adulthood—Curtin University's Activity, Food and Attitudes Program (CAFAP). The specific aims of CAFAP are to:
· compare sedentary and moderate/vigorous activity before and after participation. It is hypothesised that sedentary behaviour will be reduced and moderate/vigorous activity will be improved after the program and these will be maintained at 3, 6 and 12 months post intervention;
· compare food intake and eating behaviours before and after participation. It is hypothesised that intake of fruit and vegetables will be increased and intake of junk foods will be reduced after the program and this will be maintained at 3, 6 and 12 months post intervention;
· compare physical status before and after participation. It is hypothesised that BMI, waist circumference, cardiovascular fitness, muscle endurance, strength and power will be improved after the program and this will be maintained at 3, 6 and 12 months post intervention;
· compare mental well-being before and after participation. It is hypothesised that mental well-being will be improved after the program and this will be maintained at 3, 6 and 12 months post intervention; and
· compare perceived quality of life before and after participation. It is hypothesised that quality of life will be improved after the program and this will be maintained at 3, 6 and 12 months post intervention.
The study will also explore the influence of adolescent age, gender, baseline status, autonomy support, autonomous motivation, goal attainment and conflict, parental factors and family function on changes in adolescent outcomes. Process evaluation will also be conducted to assess program fidelity, satisfaction and adverse outcomes.
Participants and recruitment
96, 11–16 years of age) will be recruited via the health system, the education system and from the general community. Paediatric specialists, allied health professionals at a tertiary children’s hospital, general medical practices close to the study and nurses in schools close to the study will be informed about the study and asked to identify potentially suitable adolescents. Community newspapers and radio mass media will be used to inform the general community of the study. Volunteers, who meet the inclusion and exclusion criteria and are aware of the full study responsibilities, will be screened by a medical practitioner to ensure they are medically suitable to participate. Inclusion criteria are: males and females aged 11–16 years, BMI higher than the 85th
centile on the standard Centers for Disease Control (CDC) BMI-for-age growth charts [70
] (includes children who are typically classified as overweight or obese), and have passed medical screening. Children will be excluded if: their obesity is due to identified genetic, metabolic or endocrine disease, they are undergoing treatment for psychiatric disorders, they reside remotely or are unable to attend twice weekly sessions at the designated community intervention locations, or they are assessed unsafe to participate by GP or Paediatrician.
A sample size of 96 is required at post test in the intervention group to detect a 20% difference in the outcome variables at 80% power and 5% level of significance. A small effect size (Cohen’s d
0.2) is assumed for studies on behavioural effects due to the influence of extraneous variables and the subtleties of human performance.
The intervention consists of an intensive 8 week multidisciplinary program focussed on improving activity, food and attitude habits with tapered follow-up for 12 months. The intensive phase will run within school terms and be delivered by community health professionals following training in the philosophy and approach of the program. Groups of 12 to 15 adolescents and their parents will attend twice weekly. On each occasion adolescents will participate in a 45 minute exercise class involving aerobic, strength and skill stations. Adolescents will also participate in hour long education sessions covering healthy activity, healthy eating, energy balance, food labelling, preparing meals and snacks, goal setting, problem solving, dealing with mood, family activity and eating and relationship with family. Parents will participate in education sessions covering the same issues. They will also have sessions on understanding adolescence, providing support, relationships with adolescents, community resources, food budgeting and a supermarket visit. There will also be informal support from other parents and staff in ‘walk and talk’ sessions. The intensive phase will conclude with a healthy cooking celebration. During the 12 month follow-up maintenance phase, participants will be contacted, with decreasing frequency, to prompt individual activity and eating goals.
CAFAP will be evaluated using a multiple cohort, staggered-entry, waitlist controlled clinical trial design (see Figure
). Participants will be recruited and given an initial assessment then reassessed after 3 months waiting (just before starting the program). The dual pre-participation assessments will provide a within-subjects control period. Follow-up assessments will be conducted immediately after completing the intensive program (3 months after second pre-program assessment) and at 3, 6 and 12 months post program.
Trial flowchart. This figure provides an overview of the participant flow for the cohorts in each wave of the study.
Cohorts of 12–15 overweight adolescents and their parents will be progressively recruited over the study period. Cohorts are expected to start in February 2012, April 2012 and July 2012, and continue until early 2014. The staggered start will control for external seasonal and public event confounders to intervention effects.
Primary outcome measures for adolescent
Time spent in sedentary, light, and moderate-to-vigorous intensity physical activity will be assessed over 7 days using an Actical accelerometer worn on the hip. The MiniMitter Actical is a widely used and validated accelerometer in studies of children and adolescents [71
]. Seven days of accelerometer measurement are recommended for the purposes of acceptable measurement of moderate to vigorous PA [72
]. Intensity will be determined using the Actical cut-offs reported by Colley et al. [73
]. Total weekly time in all activity intensities will be calculated as well as weekend and after school weekday intensity times.
Food intake will be assessed using a 3-day dietary record using weighed or household measures. Collecting dietary data from overweight and obese adolescents and minimising underreporting poses many challenges. There is insufficient evidence in the literature to support the use of one dietary measurement tool over another [31
]. Given the small group size, variation in daily adolescent diet and burden of dietary data collection, a three day food diary has been selected as the most appropriate tool as it will provide data on meals and snacks which will inform the intervention [31
]. Adolescents will be asked to record everything they eat and drink over a period of three days, including one weekend day. They will be provided with written directions and verbal instruction from the research dietitian regarding estimating portion size. On returning the dietary record, the dietitian will clarify details and amounts of foods. The number of serves of fruit and vegetables will be determined according to the Australian Guide to Healthy Eating serve sizes [74
], and junk food will be identified according to the Food Criteria System devised by Rangan et al. [37
] and classified into 600 kJ servings as per the Australian Guide to Healthy Eating [36
]. The diet records will be analysed for specific nutrient intake using the AUSNUT database and Foodworks Professional edition version 3.02 software. Data will be extracted for intakes of total energy, macronutrients and percentage contribution to energy intake, as well as micronutrients. All nutrition analyses will be completed by a dietitian.
Secondary outcome measures for adolescent
Sedentary and physical activity behaviours
Weekly frequency and school day/weekend day duration of participation in common sedentary behaviours (TV viewing, playing electronic games, using a computer) and moderate/vigorous physical activity (sports, dance, active play) will be assessed using questions drawn from the Western Australian Child and Adolescent Physical Activity and Nutrition Survey and the Healthy Kids Queensland Survey [75
Specific food behaviours that have been shown in the research to be associated with unhealthy outcomes and greater levels of overweight status, such as eating in front of the television or missing breakfast, will be assessed using questions drawn from the NSW Schools Physical Activity and Nutrition Survey [77
] and the Healthy Kids Queensland Survey [75
Physical status-BMI and waist circumference
Weight, height and waist circumference measurements will be made using calibrated scales, stadiometer and inelastic tape measure and standardized protocols. Body mass index will be calculated and age and sex adjusted BMI z-scores (calculated against CDC reference data) will be determined to enable comparison with state-wide surveys relevant to the Australian context (Western Australia and Queensland).
Cardiovascular fitness will be assessed using the modified incremental shuttle walk test [78
]. This test is based on the widely used incremental shuttle run test designed for healthy individuals [79
] with the reduction of shuttle distance from 20 m to 10 m to provide a lower demand start to the test for those with health problems. Participants walk/run between markers in time with beeps. The beep interval decreases every minute, requiring increases in walking velocity. The test is terminated when the participant can no longer reach the end of the 10 m in time with the beep. In obese adolescents test-rest reliability is very high (ICC
0.92) and validity has been demonstrated with strong correlations with other aerobic indices including maximal oxygen uptake in bicycle ergometer tests (r
Physical status-muscle performance
Strength of the quadriceps, biceps and deltoid muscles will be assessed using standard ‘break’ manual muscle testing protocols [81
]. ‘Break’ tests require the participant to hold a position while the tester applies progressive resistance until the position is broken. Peak force will be recorded using a force transducer. All muscle testing will be performed on the non-dominant limb and each test will be performed 3 times, with an average score used for analysis. Hand-held force transducer tests of muscle strength have been shown to have good criterion validity in comparison to laboratory dynamometers, with good intra-tester and inter-tester reliability [82
Lower limb muscle power will be assessed using a vertical jump protocol [83
]. Three measures are taken and the highest recorded [84
]. Vertical jump has been shown to be appropriate for measuring the explosive power of the lower limbs in both sedentary and athletic populations [85
] and has been used to determine power outputs in obese adolescents [86
Abdominal muscle endurance will be tested via a partial curl-up test of the Canadian Standardised Test of Fitness [87
]. Measurement will be taken of the number of curl-ups performed correctly in 1 minute. This is a reliable and valid test for measuring the dynamic endurance of abdominal muscles [88
The primary mental health outcome measure will be depressive symptoms assessed using the Moods and Feelings questionnaire (MFQ; [89
]). The MFQ is a 33 item self-report measure which looks at depressive symptoms in children and adolescents. It has good test-retest reliability [90
] and has been validated with a clinical youth population [91
Perceived autonomy support and autonomous motivation
The Perceived Autonomy Support Scale for Exercise Settings (PASSES; [92
]) and the Perceived Environmental Supportiveness Scale (PESS; [93
]) will be modified to measure perceived autonomy support, structure, and involvement from parents and instructors in regard to physical activity and healthy eating behaviours. Autonomous motives for physical activity behaviours will be measured using the revised Behavioural Regulations and Exercise Scale (BREQ-2; [93
]) and the integrated regulation scale for exercise behaviour [94
]. Autonomous motives for healthy eating behaviours will be measured using an adapted version of Mullan, et al.’s BREQ [95
] and Ryan and Connell’s [96
] perceived locus of causality for diet (PLOC; [97
]) and the integrated regulation scale for exercise behaviour [94
Goal attainment and goal conflict
Adolescents will be asked to report their overall goals to achieve between each data collection point (end of 8 week program and 3, 6, and 12 month follow-up) in the areas of physical activity, sedentary behaviour, and healthy eating. The amount and frequency of a behaviour (e.g., be active for 30 minutes (moderate to vigorous) four times a week) will be detailed in the overall goal to provide an outline for breaking long-term goals into smaller more manageable weekly goals. Adolescents will be provided weekly goal setting sheets for use during the 8 week program and one year post-program period. Weekly goal sheets will include the following information for each goal area: specific weekly goal (e.g., to be moderately active for 15 minutes 4 times/week and do more than 8000 steps 4 times/week); challenge rating; goal content (i.e., intrinsic—health; extrinsic—appearance); goal motive (autonomous—inherent enjoyment; controlled—external pressure); goal details for each day of the week (activity, frequency, time of day to perform activity); daily goal steps and actual steps; weekly and overall goal progress. Participants will respond to the challenge rating using a scale ranging from 0 (easy
) to 10 (impossible
) to indicate their perceived goal difficulty. Goal progress ratings will be completed one week after goals have been set using a 0 to 10 scale. Weekly goal progress scores for each goal area will be summed across each week to form a single index for weekly goal attainment [99
]. Overall progress scores will be summed across all goal areas to form a single index for overall goal attainment [99
]. During adolescents’ first goal setting session, they will be asked to list their four most important life goals and rate how much they believe these goals will enhance or interfere with the goals they set for physical activity, sedentary behaviour, and healthy eating. At each follow up data collection, adolescents will be presented with their four life goals stated at baseline and asked to indicate how much these goals interfered with their behaviour change goals for physical activity, sedentary behaviour, and healthy eating [100
]. Responses will be indicated using a scale from 0 (made goal achievement very difficult) to 10 (made goal achievement very easy). Scores will be summed across all four life goals to form a single index for goal conflict.
Quality of life
Quality of life will be assessed with the Paediatric Quality of Life—Teen Report (PedsQL) [102
]. The PedsQL is a 23-item self-report quality of life measure for 13–18 year olds. This measure has been demonstrated to have good validity and reliability [102
Secondary outcome measures for parents
Parental sedentary and physical activity behaviours
As per the adolescents, parents will complete a brief questionnaire on sedentary and physical activity behaviours using questions drawn from the Western Australian Child and Adolescent Physical Activity and Nutrition Survey and the Healthy Kids Queensland Survey [75
Parental mental health
Parents will complete the adult version of questionnaires covering the moods and feelings area (DASS-21). The DASS-21 is a short-form of the Depression, Anxiety and Stress Scale (DASS; [104
]). The DASS has been shown to reliably distinguish between symptoms of depression, anxiety and stress in clinical [105
] and non-clinical samples [104
]. The DASS-21 has been demonstrated to have acceptable reliability and validity [106
Parental perceived autonomy and autonomous motivation
The adapted version of the PESS [93
] and PASSES [92
] used for adolescents, will be modified to measure parents’ perceptions of autonomy support, structure, and involvement provided by the instructor. The scale will be adapted to assess parents’ perceptions in regard to supporting their adolescents’ behaviour changes for physical activity and healthy eating. Autonomous motives for supporting adolescents’ physical activity will be measured using the revised BREQ-2 [93
] and the integrated regulation scale for exercise behaviour [94
] previously described for adolescents. Autonomous motives for supporting adolescents’ healthy eating behaviours will be measured using an adapted version of the PLOC [97
] and integrated regulation scale for exercise behaviour [94
] used to measure adolescents’ autonomous motives.
Parental goal setting and goal conflict
The overall and weekly goal setting sheets previously described for adolescents will be adapted for parents to use in setting their own goals to help support adolescents’ behaviour change goals. Goal progress and goal challenge scores will be reported using the same format previously described for adolescents [99
]. Goal conflict sheets for parents will be adapted from those previously described for adolescents in order to assess parents’ goal conflict regarding the goals they set for supporting their adolescents’ physical activity, sedentary behaviour, and healthy eating [100
Parents will also complete the general functioning scale of the McMaster Family Assessment Device (FAD; [107
]). The FAD is a 53 item self-report which assesses family functioning on five subscales; problem-solving, affective responsiveness, affective involvement, behaviour control and general functioning. The FAD is suitable to be administered to all members of a family over 12 years, and has adequate internal consistency, test-retest reliability and concurrent validity [107
There is substantial agreement that interventions for overweight young people need to be evaluated in terms of process and outcome and not just immediate impact [109
] to understand not only how much the intervention works but why and for whom.
A broad process evaluation will be conducted in line with the recent framework for mandatory evaluation to ensure well informed public policy decisions [110
] and recommendations by the recent Cochrane review for obesity interventions [12
]. Questionnaire surveys of participants, focus groups of participants and program facilitators are anticipated [111
]. Program fidelity will be assessed through observations of sessions, review of program notes and focus groups with program staff [112
]. Program dose will be assessed by attendance records. Barriers to participation will be explored through adolescent/parent surveys including reasons for non-attendance and gaining reasons for drop out from those who do not complete the program [113
]. Participant satisfaction will be assessed with a validated tool [114
] along with focus groups. Monitoring will also be conducted for potential adverse effects, specifically: changes in linear height and psychological well-being, as recommended [12
Immediate impact of the intervention will be assessed by comparing the changes over the 3 months between pre-program assessment and immediately post-program assessment with any changes over the 3 months waitlist period between initial assessment and pre-program assessment. The primary impact measures will be changes in: 1) leisure time spent in light and moderate/vigorous physical activity; 2) leisure time spent in sedentary activity; 3) number of weekly serves of fruit and vegetables and 4) number of weekly serves of junk foods. Secondary impact measures will include: physical status (BMI z-score, waist circumference, cardiovascular fitness, muscle performance), mental health, perceived autonomy, sedentary and diet behaviours, parental mental health and family functioning.
The longer term changes in behaviour (same variables as short term impact), as well as physical and mental well being and family functioning, will be assessed by comparing changes between assessments at baseline and at 3, 6 and 12 months post intervention.
Expected program outcomes
The primary program outcomes are expected to be improved adolescent behaviours including: an increase in physical activity (~10% change); reduction in sedentary behaviours (~10% change); an increase in fruit and vegetable serves (~10% change); a reduction in junk food serves (~10% change). It is anticipated there will be improved physical status; improved mental health including reduced symptoms of depression; improved quality of life and improved family functioning. Improved parental mental health is also expected.
Following medical screening, participants will be provided with a description of the study and given an opportunity to have responsibilities, risks and benefits of participation clarified. After informed consent/assent from parent and adolescent, baseline study entry assessments are completed. Participants then enter a 3 month waitlist period and are then reassessed pre-program. Participants will complete the 8 week intensive phase of the intervention and be re-assessed. During the maintenance phase participants will be assessed at 3, 6 and 12 months post intensive phase. Each cohort of 12 to 15 adolescents will follow this pattern, with new waves of cohorts starting with each new school term. Two cohorts will start in Waves 1 and 2, with three cohorts starting in Wave 3 (See Figure
Process evaluation will use descriptive statistics and qualitative analysis. Impact and outcome evaluation will use a multi-level mixed modelling approach for the hierarchical data collected over the 18 month (waitlist, intensive program and maintenance) period (repeated measurements of individuals nested within families). The degree of change in outcomes over time periods will be examined for associations with adolescent age, gender, baseline status, autonomy support, autonomous motivation, goal attainment and conflict, parental factors and family function, by examining time interactions with these variables. All change estimates will be presented as adjusted mean change with corresponding 95% confidence intervals. Non-normally distributed data will be appropriately transformed as necessary, and statistical models checked for unduly influential outliers. All data analysis will be performed utilising current statistical software (Stata/IC 12.0 for Windows, Statacorp). A critical alpha level of 0.01 will be used to balance type 1 and type 2 errors.
Ethical approval for this intervention has been received from Curtin University Human Research Ethics Committee (HR105/2011).