Dietary Habits, Eating Behavior, Health Behavior
Conditions
Brief summary
In this study, one year changes in healthy eating and gardening behaviors associated with the intervention were estimated using a group randomized controlled trial design, with the comparison arm receiving a delayed intervention in the following school year.
Detailed description
The small randomized controlled trial of an intensive curriculum and school garden intervention was implemented in elementary schools in two communities (Shiprock and Tsaile). The study began in year 1 with a period of engagement and recruitment of schools in each community; with the goal of recruiting three schools in each community to complete the run-in phase. The run-in period included time for greater engagement with the school community, an assessment of the potential garden site and an assessment of evaluation readiness. Depending on whether three schools in the community qualify for the study at the end of run-in, or only two schools per community, it was planned to randomize three or two schools to intervention or delayed intervention status in each community. Six schools were randomized, two intervention schools and 4 comparison (delayed intervention) schools. Baseline data collection was conducted at the beginning of the school year 2019-2020 in Year 2 with 3rd and 4th grade children. Students in the intervention schools received the curriculum and school garden during that academic school year (Study Year 2). At the end of the school year, 8-month follow-up assessment was conducted. The main evaluation period is between baseline and 8-month follow-up, and the statistical analysis methods estimate the differential change in outcome measures (healthy eating and gardening). Not part of the clinical trial, the study continued in the next school year (2020-2021). At the beginning of Study Year 3, the comparison schools were scheduled to receive the delayed intervention. During Study Year 3, it was planned to conduct quarterly community gardening workshops for adult family members in all schools. Due to the COVID-19 pandemic, a series of 10 video workshops were produced and made available via YouTube. Attendance/Views were tracked. A final assessment survey was conducted at the end of that school year (in 2021) in all 6 schools to allow additional descriptive analyses.
Interventions
BEHAVIORALBehavioral Gardening Exposure
Assistance provided with planting and maintaining the school garden
BEHAVIORALHealthy Eating and Gardening Curriculum
14 to 17 lessons (about 45 minutes each) throughout the school year, during the normal school day, that are focused on healthy eating and gardening, coordinated with the growing season
Sponsors
National Institutes of Health (NIH)
National Cancer Institute (NCI)
Fred Hutchinson Cancer Center
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
PREVENTION
Masking
SINGLE (Outcomes Assessor)
Masking description
Interviewers were not informed of school or school assignment at follow-up
Intervention model description
Group Randomized Trial
Eligibility
Sex/Gender
ALL
Age
8 Years to 12 Years
Healthy volunteers
No
Inclusion criteria
* Elementary schools in Tsaile and Shiprock will have been in existence for at least three years, school size (have between 40 to 100 students in grades 3 and 4 combined), have more than two-thirds Navajo students and satisfy a garden site checklist that can overcome garden barriers (e.g. have access to water) * Assessment is restricted to Male/female students in grades 3 and 4 attending school in either Tsaile or Shiprock
Exclusion criteria
* Students in grades other than 3rd or 4th are not eligible for assessment
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Estimated Change in Child Fruit & Vegetable Self-efficacy | Beginning to end of school year 2019-20 ( Baseline and ~ 8 months) | Self-efficacy for eating F&V score is average of responses to five 4-point Likert scale questions. Higher score is higher self-efficacy. Score range is 1 to 4. Change in self-efficacy is based on a model that uses all available data (from baseline and 8-month follow-up). The unit of analysis is participant-timepoint. The analysis accounts for nesting within schools. Using a repeated measures linear mixed model analysis, the results are estimated change in self-efficacy score from baseline to 8-month follow-up, contrasting intervention group to comparison group. From the model, the estimated change in scores from baseline to 8-month follow-up ranged from 0.14 to 0.62. The unit of measure is score on a scale. The results show the least squares estimate of change in that score. Positive estimated change in score indicates self-efficacy increased from baseline to follow-up. Positive differential change indicates that intervention group change was greater than comparison group. |
| Estimated Change in AHEI Healthy Foods Score | Beginning to end of school year 2019-20 ( Baseline and ~ 8 months). | The Healthy Foods Score is a subscale from the modified Alternative Healthy Eating Index (AHEI). Data were derived from a picture sort frequency tool estimating consumption frequencies of ten major food groups of the Navajo diet. Scoring used the published AHEI-2010. The subscale is the sum of scores for fruits (fresh & dried), vegetables (not salad), whole grains, beans, nuts, only. Values range from 0 to 36. Higher values represent a better outcome. Change in Score is based on a model that uses all available data (from baseline and 8-month follow-up). The model is from a repeated measures linear mixed model analysis. Change in Healthy Foods score is estimated from baseline to 8-mth follow-up, contrasting intervention group to comparison group. The change scores ranged from -0.63 to 3.66. Positive estimated change indicates an increase from baseline to follow-up. Positive differential change indicates that intervention group change was greater than comparison group change. |
| Estimated Ratio of Full Self-efficacy to Grow F&V at School | Beginning to end of school year 2019-20 ( Baseline and ~ 8 months) | Measure derived from Percent of participants at each time point reporting I know I can. Using generalized mixed models, specifying a binomial distribution, accounting for nesting within schools, the Odds ratio of 8-month follow-up to baseline percent reporting I know I can was estimated for both intervention and comparison groups. The overall range in these odds ratios was 0.95 to 3.63 . The contrast of intervention group to comparison group was also an odds ratio. |
Countries
United States
Participant flow
Recruitment details
Two areas of the Navajo Nation were selected. Elementary schools within 70 miles of the closest Diné College campus were identified. Eligibility criteria for schools included within 30 miles of one of the two campuses; number of 3rd and 4th grade students between 40 and 140. Ten schools were eligible for recruitment. 6 schools completed run-in.
Pre-assignment details
Recruitment-eligible schools were 5 in each area. These schools were listed in random order within area. 9 schools started run-in, one school was not contacted. Two schools failed run-in, one in each area, leaving 6 schools eligible to be randomized. In 2019, six schools completed run-in and were randomized to intervention (2) or delayed intervention (4).
Participants by arm
| Arm | Count |
|---|---|
| Intervention Group Intervention schools receive a raised bed school garden and a curriculum of healthy eating and gardening lessons.
3rd and 4th grade students in the intervention group received the curriculum and exposure to the school garden during the academic school year (2019-20)
Behavioral Gardening Exposure: Assistance provided with planting and maintaining the school garden
Healthy Eating and Gardening Curriculum: 14 to 17 lessons (about 45 minutes each) throughout the school year, during the normal school day, that are focused on healthy eating and gardening, coordinated with the growing season. | 100 |
| Comparison Group For the comparison group schools, no program was provided in school year (2019-2020).
After the trial, and the follow-up evaluation, comparison schools received delayed intervention components as follows: School gardens were created during the 2020-2021 school year (subject to delays due to pandemic restrictions). Curriculum materials were supplied to the school for use by teachers as they saw fit, after pandemic restrictions were lifted. | 194 |
| Total | 294 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 |
|---|---|---|---|
| Overall Study | Student failed to respond to follow-up survey | 28 | 53 |
Baseline characteristics
| Characteristic | Intervention Group | Total | Comparison Group |
|---|---|---|---|
| Age, Continuous | 8.6 Years STANDARD_DEVIATION 0.7 | 8.7 Years STANDARD_DEVIATION 0.8 | 8.7 Years STANDARD_DEVIATION 0.9 |
| AHEI Healthy Foods Score | 16.9 units on a scale STANDARD_DEVIATION 5.8 | 17.5 units on a scale STANDARD_DEVIATION 6.2 | 17.8 units on a scale STANDARD_DEVIATION 6.4 |
| BMI for age percentile | 73.1 percentile STANDARD_DEVIATION 31.2 | 70.9 percentile STANDARD_DEVIATION 30.3 | 69.7 percentile STANDARD_DEVIATION 29.8 |
| Full self-efficacy to grow F&V at school | 49.0 percentage of participants | 43.2 percentage of participants | 40.0 percentage of participants |
| Overweight or Obese | 50 Participants | 133 Participants | 83 Participants |
| Race (NIH/OMB) American Indian or Alaska Native | 91 Participants | 266 Participants | 175 Participants |
| Race (NIH/OMB) Asian | 3 Participants | 5 Participants | 2 Participants |
| Race (NIH/OMB) Black or African American | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) More than one race | 3 Participants | 11 Participants | 8 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 0 Participants | 0 Participants | 0 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 1 Participants | 9 Participants | 8 Participants |
| Race (NIH/OMB) White | 2 Participants | 3 Participants | 1 Participants |
| Region of Enrollment United States | 100 Participants | 294 Participants | 194 Participants |
| Self-efficacy to eat F&V | 3.2 units on a scale STANDARD_DEVIATION 0.65 | 3.3 units on a scale STANDARD_DEVIATION 0.6 | 3.4 units on a scale STANDARD_DEVIATION 0.57 |
| Sex: Female, Male Female | 42 Participants | 151 Participants | 109 Participants |
| Sex: Female, Male Male | 58 Participants | 143 Participants | 85 Participants |
| Understands Navajo | 49 Participants | 168 Participants | 119 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk |
|---|---|---|
| deaths Total, all-cause mortality | 0 / 100 | 0 / 194 |
| other Total, other adverse events | 0 / 100 | 0 / 194 |
| serious Total, serious adverse events | 0 / 100 | 0 / 194 |
Outcome results
Primary
Estimated Change in AHEI Healthy Foods Score
The Healthy Foods Score is a subscale from the modified Alternative Healthy Eating Index (AHEI). Data were derived from a picture sort frequency tool estimating consumption frequencies of ten major food groups of the Navajo diet. Scoring used the published AHEI-2010. The subscale is the sum of scores for fruits (fresh & dried), vegetables (not salad), whole grains, beans, nuts, only. Values range from 0 to 36. Higher values represent a better outcome. Change in Score is based on a model that uses all available data (from baseline and 8-month follow-up). The model is from a repeated measures linear mixed model analysis. Change in Healthy Foods score is estimated from baseline to 8-mth follow-up, contrasting intervention group to comparison group. The change scores ranged from -0.63 to 3.66. Positive estimated change indicates an increase from baseline to follow-up. Positive differential change indicates that intervention group change was greater than comparison group change.
Time frame: Beginning to end of school year 2019-20 ( Baseline and ~ 8 months).
Population: Participants were 3rd and 4th grade students responding to baseline and/or 8-month follow-up. The participant at specified time point was the unit of analysis.
| Arm | Measure | Value (LEAST_SQUARES_MEAN) | Dispersion |
|---|---|---|---|
| Intervention Group | Estimated Change in AHEI Healthy Foods Score | 1.98 score on a scale | Standard Error 0.85 |
| Comparison Group | Estimated Change in AHEI Healthy Foods Score | 0.58 score on a scale | Standard Error 0.61 |
Comparison: Results present estimated change in score from baseline to 8-month follow-up in each group from model.~Values from the 100 and 194 baseline respondents and from the 72 and 141 follow-up respondents of intervention and comparison groups respectively contributed to the repeated measures linear mixed model, where the units of analysis were participant-time. The nesting within schools was accounted for in the model.p-value: 0.1895% CI: [-0.66, 3.49]Mixed Models Analysis
Primary
Estimated Change in Child Fruit & Vegetable Self-efficacy
Self-efficacy for eating F&V score is average of responses to five 4-point Likert scale questions. Higher score is higher self-efficacy. Score range is 1 to 4. Change in self-efficacy is based on a model that uses all available data (from baseline and 8-month follow-up). The unit of analysis is participant-timepoint. The analysis accounts for nesting within schools. Using a repeated measures linear mixed model analysis, the results are estimated change in self-efficacy score from baseline to 8-month follow-up, contrasting intervention group to comparison group. From the model, the estimated change in scores from baseline to 8-month follow-up ranged from 0.14 to 0.62. The unit of measure is score on a scale. The results show the least squares estimate of change in that score. Positive estimated change in score indicates self-efficacy increased from baseline to follow-up. Positive differential change indicates that intervention group change was greater than comparison group.
Time frame: Beginning to end of school year 2019-20 ( Baseline and ~ 8 months)
Population: Participants were 3rd and 4th grade students responding to baseline and/or 8-month follow-up. The participant at specified time point was the unit of analysis.
| Arm | Measure | Value (LEAST_SQUARES_MEAN) | Dispersion |
|---|---|---|---|
| Intervention Group | Estimated Change in Child Fruit & Vegetable Self-efficacy | 0.47 score on a scale | Standard Error 0.08 |
| Comparison Group | Estimated Change in Child Fruit & Vegetable Self-efficacy | 0.24 score on a scale | Standard Error 0.05 |
Comparison: The unit of measure is score on a scale. The results show the least squares estimate of change in that score, from baseline to 8-month follow-up in each group from model.~Values from the 100 and 194 baseline respondents and from the 72 and 141 follow-up respondents of intervention and comparison groups respectively contributed to the repeated measures linear mixed model, where the units of analysis were participant-time. The nesting within schools was accounted for in the model.p-value: 0.01795% CI: [0.04, 0.41]Mixed Models Analysis
Primary
Estimated Ratio of Full Self-efficacy to Grow F&V at School
Measure derived from Percent of participants at each time point reporting I know I can. Using generalized mixed models, specifying a binomial distribution, accounting for nesting within schools, the Odds ratio of 8-month follow-up to baseline percent reporting I know I can was estimated for both intervention and comparison groups. The overall range in these odds ratios was 0.95 to 3.63 . The contrast of intervention group to comparison group was also an odds ratio.
Time frame: Beginning to end of school year 2019-20 ( Baseline and ~ 8 months)
Population: Participants were 3rd and 4th grade students responding to baseline and/or 8-month follow-up. The participant at specified time point was the unit of analysis.
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| Intervention Group | Estimated Ratio of Full Self-efficacy to Grow F&V at School | 1.92 odds ratio |
| Comparison Group | Estimated Ratio of Full Self-efficacy to Grow F&V at School | 1.48 odds ratio |
Comparison: Odds ratio of 8-month follow-up to baseline percent reporting I know I can was estimated from repeated measures generalized linear mixed model with binomial distribution.~Values included from the 100 and 194 baseline respondents and from the 72 and 141 8-month follow-up respondents of intervention and comparison groups respectively. The person-timepoint is the unit of analysis.p-value: 0.5195% CI: [0.6, 2.81]Mixed Models Analysis