- Click here for explanation of classification scheme.

To evaluate an intervention designed to prevent obesity by modifying Girl Scout troop meeting environments and by empowering girls to improve the quantity and quality of family meals in their home environment. 

•  Criteria for the troops:
  • Troop needed to be a registered Girl Scout Junior troop consisting of girls in the fourth and fifth grades
  • Troop leaders had to complete Girl Scout leader training and pass a criminal background check
  • Troops needed to meet at least twice per month at meeting facilities capable of allowing physical activity and food preparation
  • Troops needed to have initial agreement of leaders and parents for the troop to participate in a research study.
• Criteria for the Girl Scouts and parents:
  • Girls had to be attending members of Girl Scouts in one of the included troops
  • All girls of participating troops were included for direct observation variables, and those with parental consent were included for the individual variables understudy
  • Parents were included if they agreed to complete a questionnaire for each child attending one of the troops.

• Criteria for the troops
  • Troops not primarily composed of Girl Scout Juniors
  • Troops not meeting during the study period
  • Troops not having leader and parental consensus approval for troop participation.
• Criteria for the Girl Scouts and parents
  • Girl was not an attending member of the included Girl Scout troops
  • Inability to speak or read English.

Recruitment

• Troops were chosen from three adjacent Midwestern towns
• Participating troops earned $10 per girl and individual families earned $10 for taking part in the study.

Design

• Group-randomized controlled trial
• Nested cohort design with troops being the unit of randomization
• Troops stratified into large and small troops; then according to a random number generator were randomized within strata to the control or intervention conditions.

Blinding Used

Research assistants were blind to condition of each troop.

Intervention

• Intervention troop leaders instituted the intervention components at the scheduled troop meetings after being trained on the curriculum and supporting policies
  • Troop leaders underwent two hours of training 
  • Regular and ongoing e-mail and phone support took place throughout the intervention time period.
• A trained research assistant observed each troop during seven full meetings between time one and time two assessments to record troop meeting environmental variables, including leader health-promotion behaviors
• Intervention was based on core components of Social Cognitive Theory
  • Role modeling by peers, troop leaders and parents
  • Skill-building through active mastery experiences
  • Enhancement of self-efficacy and proxy-efficacy through role playing and active mastery experiences
  • Reinforcement of behavior through verbal praise and merit badges.
• Intervention consisted of three main components:
  • Interactive educational curriculum delivered by troop leaders
    • Eight modules, delivered over the course of about four months
    • Each module consisted of:
      • A discussion of intervention target behaviors
      • Worksheet for goal setting and self-monitoring
      • Physically active recreation session
      • Fruit or vegetable snack recipe preparation
      • Family meal role-playing
      • Clean-up period
      • Description of the take-home assignment.
    • Modules were designed to require 60 to 90 minutes to deliver.
  • Troop meeting policies implemented by troop leaders:
    • Providing 15 minutes per meeting for physically active recreation
    • Troop leaders participating in physically active recreation with girls
    • Provision of a fruit or vegetable snack prepared by girls
    • Troop leaders eating fruit of vegetable snack with girls
    • Troop leaders verbally promoting physical activity, fruit and vegetable consumption in troop meetings and for home, and verbally promoting family meals for home
    • Prohibition of sugar-sweetened beverages, candy and TV watching during meetings.
  • Badge assignments were completed at home by Girl Scouts with parental assistance.
• Target behaviors of the intervention included:
  • Frequent family meals
  • Parent-child shared physical activity
  • Elimination of TV during mealtime
  • Drinking water instead of sugar-sweetened beverages at mealtime
  • Including fruits and vegetables in family meals
  • Practicing good manners during family meals
  • Helping parents prepare family meals and cleaning up afterwards.  

Statistical Analysis

• Multiple measures were used to evaluate the intervention including process measures and individual-level outcome measures
• Intervention effects on individual outcome measures were analyzed consistent with other randomized controlled site-based interventions where the statistical design of the study is complicated by the lack of independence of data
• To adjust for the clustered data structure, a mixed-model design structure was used
• SAS 9.1 software package was used for the mixed-model analyses
• SPSS 15.0 was used to compute descriptive statistics and univariate analyses
• To assess individual-level intervention effects, general linear model analyses were run on difference scores with girls nested within the troop as random effect (to address clustering of girls within troops) and weight status (overweight or not), authoritarian parenting level (median split), socio-economic status (free/reduced or not) and race/ethnicity as fixed effects
• To assess differences in objectively monitored physical activity by condition, general linear model analyses were run on moderate-to-vigorous physical activity, with girls nested within the troop as random effect and weight status, socio-economic status and race/ethnicity as fixed effects
• To assess outcomes for troop meeting environments and percentages, descriptive statistics and chi-square analyses were used. 

Timing of Measurements

• Time one assessment occurred during a two-week period in October 2007
• Time two assessment occurred during a two-week period in April 2008
• Troop meeting observations occurred during seven full meetings between time one and time two assessments
• Parents completed the questionnaire before and after the intervention period. 

Dependent Variables

• Primary outcome: Child BMI Z-score
• Secondary outcomes:
  • Accelerometer-measured physical activity
  • Direct observation of snack offerings and consumption
    • A research assistant attended troop meetings on seven occasions and continuously directly observed aspects of the troop environment and activities, recording observations in a customized logbook 
    • During snacks, all foods and drinks accessible to girls and troop leaders were recorded along with the number of girls and adults actually consuming each food product.
  • Time spent in physically active meeting content: Research assistant recorded the amount of physically active time during meetings
  • Leader encouragement of physical activity and healthful eating. Assessed by observer reports by the research assistant of the meetings.  

Independent Variables

Intervention troop leaders implemented policies promoting physical activity and healthful eating at meetings and implemented a curriculum promoting obesity-prevention behaviors at home.

Control Variables

• Parents' socio-economic status
• Parents' parenting style determined from parent survey:
  • Questionnaire completed outside of troop meeting times, before and after intervention period
  • Assessed seven areas:
    • Family meal frequency
    • Fruit and vegetable servings typically consumed by parent
    • Habitual physical activity of parent was assessed using four items from the Behavioral Risk Factor Surveillance System, assessing frequency and duration (in 10-minute bouts or longer) of moderate and vigorous physical activity
    • Frequency of eating while watching TV
    • Sugar-sweetened beverage consumption 
    • Parent's height and weight were self-reported
    • Parenting style was measured withe the Parental Authority Questionnaire–Revised, containing 30 items with three sub-scales indicating authoritative, permissive and authoritarian
    • Family  cohesion scale.
• Girl's ethnicity
• Number of children per household
• Leader self ratings of intervention implementation: Leaders responded on a three-point scale
• Influences on BMI by girls' behaviors and characteristics from girls' survey: 
  • Questionnaires were completed at times one and two administered according to a standardized script
  • Fruit servings typically consumed: Responses were given on a five-point scale ranging from "none" to "4 servings or more"
  • Vegetable servings typically consumed:
    • Responses were given on a five-point scale ranging from "none" to "4 servings or more"
    • Fruits and vegetables were summed to create a single index of fruit and vegetable consumption (FV).
  • Habitual physical activity:
    • Physical activity was defined as "Any play, game, sport, or activity that gets you moving and breathing harder" and was discussed with numerous examples provided both by researchers and participants
    • Two items assessed days in the past week "On how many days were you physically active for a total of at least 60 minutes per day?"
  • Sugar-sweetened beverage consumption:
    • "Over the past week, how often did you drink regular soda or sugar-sweetened beverages?"
    • Responses were given on an eight-point scale.
  • Frequency of eating with television: Assessed with three items from the Family Eating and Activity Questionnaire-Revised.
• Influences on BMI by parents' behaviors and characteristics.

Initial N

• Assessed for eligibility: 11 troops, 116 girls
• Randomized: Seven troops, 76 girls.

Attrition (Final N)

• Intervention: 33 girls
• Analyzed: 39 girls.

Age

• Intervention: 10.6 (1.1) years
• Control: 10.5 (1.3) years.

Ethnicity

• Non-Hispanic Caucasian girls
  • Intervention: 79.4%
  • Control: 75.0%
• Racial/ethnic minority girls
  • Intervention: 20.6%
  • Control: 25.0%.

Other Relevant Demographics

• Parents are college graduates:
  • Intervention: 56.3%
  • Control: 48.7%.
• Lower socio-economic status:
  • Intervention: 28.1%
  • Control: 35.0%.

Anthropometrics

BMI

• Intervention: 20.1 (4.4)
• Control 19.1 (2.9).

Location

Kansas, three towns ranging in population from about 4,000 to 50,000.

Key Findings

Primary Outcome

•  Girls' BMI Z-score from Time One to Time Two:
  • Intervention, mean (SD): 0.57 (0.94) to 0.55 (0.94)
  • Control, mean (SD): 0.38 (0.75) to 0.36 (0.74)
  • No significant difference (F_1,5=0.42, P=0.544).

Secondary Outcomes

• Accelerometer-measured physical activity accumulated in intervention troops:
  • Significantly less sedentary activity (x²=6.3, P=0.011)
  • Significantly more moderate activity (x²=8.2, P=0.004)
  • Significantly more moderate-to-vigorous activity (x²=18.4, P<0.001).
• Opportunity for healthful eating and physical activity were greater in intervention group relative to control (x²=210.8, P<0.001)
• Time spent in physically active meeting content:
  • Mean intervention time: 20.7 minutes
  • Mean control time: 2.1 minutes
  • Significant difference P<0.001.
• Intervention leader encouragement of physical activity and healthful eating:
  • Promoted physical activity more frequently (x²=23.46, P<0.001)
  • Promoted healthful eating more frequently (x²=18.14, P<0.001)
  • Discouraged healthful eating and physical activity less frequently (x²=9.63, P=0.002).

Other Findings

Individual Characteristics by Troop Assignment at Time One

Demographic and Psychosocial Variables	Intervention Mean (SD)	Control Mean (SD)
Percent parents that are college graduates	56.3	48.7
Percent lower socio-economic status	28.1	35.0
Percent non-Hispanic Caucasian girls	79.4	75.0
Percent racial/ethnic minority girls	20.6	25.0
Authoritarian Parenting Scale Scorea	31.1 (3.9)	32.9 (4.4)
Authoritative Parenting Scale Scorea	38.8 (3.3)	38.6 (3.6)
Permissive Parenting Scale Scorea	23.2 (3.9)	23.9 (4.1)
Family Cohesion Scale Scoreb	64.4 (10.4)	60.4 (12.2)
Children per household	2.7 (1.3)	2.4 (0.8)
Girls' age	10.6 (1.1)	10.5 (1.3)

^aScale scores 10 to 50, higher numbers possessing more of the trait
^bScale scores 16 to 80, higher number possessing more of the trait; no significant differences by condition.

Leader self-ratings of intervention implementation

• Three troop leader self-rating averages over the eight modules ranged from 1.52 to 1.86 (zero = no implementation to 2.0 = full implementation)
• Troops differed (F_2,18 = 21.5, P<0.001) in overall implementation with averages of 1.43, 1.86 and 1.84 (mean=1.71). 

Troop environmental observations

• Control group: Based on 2,328 minutes of direct observation data from 28 meetings, the percentage of meeting time devoted to:
  • Girl Scouting activities (67.3%)
  • Management (16%)
  • Snack (9.8%)
  • Active games (2.1%)
  • Other content (4.9%).
• Intervention group: Based on 1,952 minutes of direct observation data from 21 meetings, the percentage of meeting time devoted to: 
  • Girl Scouting activities (42.4%)
  • Active games (20.7%)
  • Management (17%)
  • Snack (18.3%)
  • Other content (1.5%).

Inter-Rater Reliability Statistics for Troop Environmental Variables (Based on 144 Observed Minutes)

		Percent Agreement	Intra-class correlation	Cohen's Kappa	Significance Level
Session Context	Free time or structured activity	100	1.000	1.000	P<0.001
	General meeting content	95.9	0.977	0.941	P<0.001
	PA educational content	88.8	0.199	0.099	P=0.006
	HE educational content	98.6	0.920	0.850	P<0.001
Troop Leader Behavior	PA verbal promotion	82.5	0.083	0.038	P=0.438
	PA physical promotion	97.2	0.954	0.911	P<0.001
	PA promotion out-of-troop	100	1.000	1.000	P<0.001
	Any PA promotion	98.6	0.980	0.960	P<0.001
	HE verbal promotion	94.4	-0.026	-0.012	P=0.820
	HE physical promotion	96.5	0.000	**	**
	HE promotion out-of-troop	97.2	0.746	0.588	P<0.001
	Any HE promotion	90.3	0.542	0.367	P<0.001
	No HE or PA promotion	89.5	0.857	0.746	P<0.001

PA=physical activity
HE=healthful eating
**Unable to compute values due to lack of variability in this observation.

Girl Scout Troop Meeting Time in Physically Active Content (4,280 Minutes Total Observed Time)

		Total Minutes Active Content	Total Observed Minutes	Percent of Minutes in Active Content	Mean Minutes Active Content per Meeting
Intervention Troops	INT-1	99	824	12.0	14.1
	INT-2	131	562	23.3	18.7
	INT-3	175	566	30.9	25.0
	Mean	135	650.7	20.7*	19.3
Control Troops	CON-1	0	394	0	0
	CON-2	8	585	1.4	1.1
	CON-3	30	742	4.0	4.3
	CON-4	10	607	1.6	1.4
	Mean	12	582	2.1*	1.7

INT=intervention troops
CON=control troops
*Significant difference by condition
x²=210.8
P<0.001.

Raw Frequency Count of Observed Food Accessibility and Actual Food Exposures* in Troop Meeting Snacks by Condition (41 Snack Observations from 49 Troop Meetings)

		INT Food Accessibility	CON Food Accessibility	INT Food Exposures	CON Food Exposures
Fruits and vegetables	All fruits (with juice)	53	17	359	68
	Fruit juices	14	9	94	39
	All vegetables (with juices)	33	6	225	30
	Vegetable juice	2	0	11	0
Drinks	Drinking water	12	0	89	0
	Sugar-sweetened beverages	2	9	2	48
	Other drinks	2	3	19	13
Other food items	Salty snacks	6	8	47	51
	Dairy products (with milk)	13	7	93	51
	Candy	1	13	6	72
	Cakes and cookies	2	23	10	182
	Breads	2	7	18	42
	Meat, nuts, legumes	6	10	48	60
	Condiments	4	0	30	0

INT=intervention troop
CON=control troop
Intervention troops offered snacks at all 21 meetings (100%), control troops offered snacks at 20 out of 28 meetings (71%)
*Accumulated number of girls tasting/eating a food during troop meetings, over 41 troop observations.

Troop Environment and Troop Leader Behavior by Condition

	Intervention troops	Control troops	X2	Significance
Meeting time was structured (percent)	97.4	90.3	5.44	0.020
PA knowledge content (percent)	6.0	0.3	6.38	0.012
HE knowledge content (percent)	11.7	0.4	13.64	<0.001
Family connection content (percent)	3.1	0.0	3.09	0.079
Any PA promotion (percent)	16.6	1.5	23.46	<0.001
Any HE promotion (percent)	18.9	0.4	18.14	<0.001
No PA or HE promotion (percent)	64.5	99.1	1167.7	<0.001
Any PA discouragement (percent)a	0.2	0.6	4.64	0.031
Any HE discouragement (percent)a	0	0.3	5.88	0.015
No PA or HE discouragement (percent)a	99.8	99.1	9.63	0.002

PA=physical activity
HE=healthful eating
*This x² calculated based on actual observed number, rather than percent.

Accelerometer-measured physical activity as percentage of meeting time

• Sedentary physical activity:
  • Intervention: 64.7
  • Control: 77
  • X²=6.3, P<0.05.
• Light physical activity:
  • Intervention: 27.9
  • Control 21.4.
• Moderate physical activity:
  • Intervention: 6.0
  • Control: 1.5
  • X²=8.2, P<0.05.
• Vigorous physical activity:
  • Intervention: 1.0
  • Control: 0.1.
• Moderate-to-vigorous physical activity:
  • Intervention: 7.4
  • Control: 1.5
  • X²=18.4, P<0..05.
• Based on mixed-model analyses of accumulated minutes of moderate-to-vigorous physical activity, there were no significant differences by:
  • Weight status (F_1,400=0.45, P=0.50)
  • Socio-economic status (F_{1, 400}=1.86, P=0.173)
  • Race/ethnicity (F_{1, 400}=0.01, P=0.924).
• No significant interactions between intervention and these categorical variables (F_{1, 400}=0.01 to 0.21, P+0.648 to 0.918). 

 Individual Outcomes at Time One and Time Two by Condition

Individual Outcome Variables	Time 1 INT Mean±SD	Time 2 INT Mean±SD	Time 1 CON Mean±SD	Time 2 CON Mean±SD
Girls' BMI	20.1±4.4	20.4±4.5	19.1±2.9	19.2±3.0
Girls' BMI percentile	65.2±27.0	64.8±26.9	64.5±23.8	62.2±23.2
Girls' BMI Z-score	0.57±0.94	0.55±0.94	0.38±0.75	0.36±0.74
Parent BMIa	29.1±6.4	29.5±6.9	29.1±6.7	30.0±7.5
Family meals per weeka (scale zero to 21)	11.2±4.2	10.9±3.6	11.4±5.0	12.1±4.7
Girl days per week of 60 minutes MVPAb (scale zero to seven)	4.2±1.8	3.9±1.7	4.5±1.9	3.5±1.8
Percent of parents meeting MVPA standarda	46.7	42.0	33.3*	36.4
Girl FV servings per dayb (scale zero to eight)	5.0±2.0	4.9±1.7	3.7±1.9*	3.7±1.8*
Parent FV servings per daya (scale zero to eight)	3.9±1.5	4.4±1.6	3.7±1.8	4.4±1.5
Girl eating with TV scaleb (scale zero to four)	1.1±0.7	0.8±0.7	1.1±0.8	1.1±0.7
Parent eating with TV scalea (scale zero to four)	1.8±0.7	1.7±0.7	2.0±0.8	2.1±0.7*
Girl days per week SSB consumptionb (scale zero to seven)	3.1±2.2	2.3±2.4	2.6±2.4	2.2±2.4
Parent days per week SSB consumptiona (scale zero to seven)	1.8±2.2	2.0±2.6	2.7±2.7	2.4±2.8

Means adjusted for troop clustering, weight status, authoritarian parenting, socioeconomic status and race/ethnicity.

INT=intervention
CON=control
MVPA=moderate-to-vigorous physical activity
FV=fruits and vegetables
SSB=sugar-sweetened beverages
^aFrom parent report
^bFrom child report
*Significant difference by condition within same time point; for all variables, there were no significant difference in change from time one to time two by condition (P>0.05).

Body Mass Index

• Intra-class correlation coefficient assessing the troop-level variance of BMI associated with our primary outcome variable was low (ICC=0.025)
• Mixed model analysis on difference scores (T2 minus T1) revealed:
  •  No significant main intervention effects for girl BMI z-scores (F_{1, 5}=0.42, P=0.0544)
  •  No significant main intervention effects for parent BMI (F_1,5=1.58, P=0.264)
  • Significant main effect of socio-economic status on parent BMI (F_{1, 35}=6.74, P=0.014)
  • Lower socio-economic status parents increased more than three BMI units from time one to time two.

Body Mass Index Influences, Girl

• Mixed-model analysis on difference scores revealed:
  • No significant main intervention effect for:
    • Girl fruit and vegetable servings (F_1,5=1.54, P=0.269)
    • Girl moderate-to-vigorous physical activity (F_1,5=0.09, P=0.779)
    • Girl sugar-sweetened beverage consumption (F_1,5=0.41, P=0.549)
    • Girl eating with TV (F_1,5=0.62, P=0.463).
  • Significant main effect for:
    • Socio-economic status on girl physical activity (F_1,50=8.18, P=0.006)
    • Lower socio-economic status girls decreased in physical activity from time one to time two across conditions.

Body Mass Index Influences, Parent

• Mixed-model analysis on difference scores revealed:
  • No significant main intervention effect for:
    • Parent fruit and vegetable consumption (F_1,5=1.94, P=0.223)
    • Parent eating with TV (F_1,5=1.55, P=0.269)
    • Parent moderate-to-vigorous physical activity (F_1,5=0.87, P=0.393)
    • Parent sugar-sweetened beverage consumption.
  • Significant main intervention effect for:
    • Socio-economic status on parent fruit and vegetable consumption (F_1,5=5.51, P=0.023): Lower socio-economic status parents increased fruit and vegetable consumption by two servings from time one to time two across conditions
    • Authoritarian parenting style on parent physical activity (F_1,46=7.47, P=0.009):
      • Parents lower in authoritarian parenting level reported higher physical activity at time two compared to time one.
  • Significant interactions:
    • Between socio-economic status and parent fruit and vegetable consumption (F_1,47=5.51, P=0.023): Parents of lower socio-economic status in the control condition reported significant increases of fruit and vegetable consumption across times 1 and 2
    • Between girl weight status and parent eating with TV (F_1,50=6.95, P=0.011): Parents of overweight girls in the control condition increased eating with TV across the two time periods
    • Between attendance level and intervention on parent physical activity (F_1,50=5.07, P=0.029): Parents with girls with irregular attendance increased significantly in physical activity across the two time points.

• Implementing a health promotion curriculum and supporting policies to provide more healthful environments in Girl Scout troop meetings appears feasible on a broader scale
• Additional work is needed to bridge health promotion from such settings to other environments if lasting individual-level behavior change and obesity prevention remain targeted outcomes.

University/Hospital:	Kansas State University and University of Colorado
Not-for-profit	Sunflower Foundation: Health Care for Kansas