Youth Chef Academy: Pilot Results From a Plant-Based Culinary and Nutrition Literacy Program for Sixth and Seventh Graders

Abstract

BACKGROUND

National data confirm that youth are not eating recommended amounts of fruits and vegetables (F/V), legumes, and whole grains (WGs). Establishing plant-based eating patterns early in life may positively impact long-term health through tracking of adolescent eating patterns into adulthood and through potential associations between adolescent dietary intake and adult disease risk. The study aim was to examine the effectiveness of Youth Chef Academy (YCA), a classroom-based experiential culinary and nutrition literacy intervention for sixth and seventh graders (11- to 13-year-olds) designed to impact healthy eating.

METHODS

Study used a nonequivalent control group design with 8 schools selected for similarity in: free/reduced-price lunch, race/ethnicity, and student mobility rate (N = 248). Primary outcomes were times per day of F/V, vegetable, and WG consumption. Students completed a survey to assess primary outcomes and other measures at baseline and post-intervention.

RESULTS

Significant increases in times per day of F/V (p =.022) and vegetable only (p =.015) consumption in the intervention group compared to the control group. Increases in WG consumption showed trended toward significance (p=.071). Student engagement and nutrition knowledge showed significant intervention effects.

CONCLUSIONS

YCA positively impacts behavioral and knowledge variables related to healthy eating and increases students’ engagement in their classrooms.

Childhood obesity rates have risen sharply in recent decades¹ and is responsible for over $14.1 billion in US health care costs annually.² Healthy eating can lower the risk of obesity and related conditions^3,4 and a critical component of a healthy diet is consumption of plant-based foods.⁵ Fruit and vegetable (F/V)^6,7 and whole grain (WG) consumption^8,9 have been linked to reduced risk of heart disease and all-cause mortality. Whole grain consumption has also been linked to reduced risk of cancer8 and type 2 diabetes.¹⁰ Establishing plant-based eating patterns early in life may impact diet-related chronic disease trends^11–13 through the tracking of adolescent eating patterns into adulthood^11–16 and through potential associations between adolescent dietary intake and adult disease risk.¹⁷

National data confirm that youth are not eating recommended amounts of F/V, legumes, and WG.^4,18–22 National Health and Nutrition Examination Survey (NHANES) data from 2003 to 2010 show that while whole fruit consumption increased across this time period, total vegetable and vegetable subgroup intake did not increase over time and no sociodemographic group met Healthy People 2020 targets for F/V consumption.²⁰ Similarly, 2001–2010 NHANES data show no significant increase in WG consumption for any age group.²² Effective strategies are needed to improve F/V and WG consumption among youth.

A growing body of evidence shows that nutrition education combined with culinary training improves diet.²³ The 2015 Dietary Guidelines for Americans recommends action strategies for professionals such as “teach skills like . . . cooking, meal planning, and label reading that help support healthy eating patterns.”⁴ Applying this recommendation to youth through cooking-based nutrition education programs may provide a promising approach to improve plant food consumption and prevent obesity.²⁴ Learning activities like cooking which feature fun, collaborative, authentic work are associated with higher student engagement and may promote learning across academic subjects.²⁵ Research examining eating behaviors of youth has shown a positive link between cooking family meals and eating a healthy diet.^12,26Because middle-school-aged adolescents may cook at home more frequently than high schoolers,²⁶ culinary education during early adolescence could leverage this time of increased cooking responsibilities to impact diet quality.

Although home economics classes once served to introduce students to the basics of cooking, these programs have declined by nearly 40% since 2003²⁷ and their impact on diet composition has not been studied. In the intervention literature, published accounts of youth nutrition programs with cooking components are limited and vary in the degree of hands-on cooking permitted and ages targeted.^28–37 Nearly all in-school interventions target elementary school students (5- to 11-year-olds) while middle school programs (sixth-eighth grade: 11- to 14-year-olds) take place in camps and afterschool settings, limiting the consistency in attendance that a classroom-based program affords. While not delivered in middle school classrooms, 3 programs with cooking components measured change in consumption. Two reported significant behavioral change^32,33 and the third reported significant change in plate waste.²⁸

While few culinary programs for youth have reported significant increases in F/V consumption, several have documented changes in other variables related to dietary behavior such as attitudes, awareness, knowledge, self-efficacy for cooking, and food preferences.^{28–30,32–36,38} Some of these variables have been established as precursors to changing or predicting dietary behavior in youth. For example, Bere and Klepp³⁹ reported that change in F/V preferences was a significant predictor of future change in F/V intake among sixth/seventh graders. Backman et al⁴⁰ showed that intention to eat a healthy diet among adolescents predicted their consumption of F/V. Additionally, a 2006 review of determinants of F/V consumption among youth reported that nutrition knowledge and self-efficacy showed positive associations with F/V consumption, although additional evidence is needed.⁴¹

Given the importance of diet to health and the stagnant trends in plant food consumption among youth, innovative and effective strategies addressing healthy eating are critically needed. The purpose of this pilot study was to examine the effectiveness of Youth Chef Academy (YCA), a classroom-based, hands-on culinary and nutrition literacy curriculum, for middle school-aged students in a large, urban public school system. YCA’s nutrition literacy focus refers to program components designed to deliver a systems level context (eg, historical, social, economic, cultural) for practicing healthy eating and developing culinary skills. YCA has the potential to add to evidence-based nutrition education. Its constellation of features, including linkages to academic standards, are not fully present in other published interventions. The primary hypothesis was that students assigned to the YCA curriculum would demonstrate greater improvement in times per day of F/V, vegetable only (VG), and WG consumption over the 6-week intervention compared to students in the control group.

METHODS

This study, implemented 2011–2013, employed a nonequivalent control group design with 8 public kindergarten-eighth grade (K-8) schools in one school district. One sixth- or seventh-grade classroom (11- to 13-year-olds) per school was included in the study. Four intervention classrooms received the YCA curriculum and 4 control classrooms received their usual academic curriculum. No in-depth nutrition education was offered as part of the usual academic curriculum. This design helped minimize threats to internal validity including history and testing effects. The outcome evaluation utilized a baseline and post test survey, administered the week after YCA ended, in all 8 classrooms.

Participants

With approval of the school system superintendent (director of school district) and working in partnership with the school system’s curriculum specialist for health and physical education, study investigators recruited 4 intervention schools by initiating one-on-one meetings with principals. With teacher engagement and principal approval, study investigators identified 1 classroom in each school for the intervention. The school system research office then selected the control schools for similarity on free/reduced-price lunch utilization, race/ethnicity, grade, and student mobility rate.

All students in the intervention classrooms experienced the YCA curriculum. Students’ parents provided passive consent for their child’s participation in the surveys and students provided written assent. Students without parental consent or personal assent completed work as directed by their classroom teacher during the surveys. The university’s Institutional Review Board and the public school system’s Research Review Board approved this study prior to implementation (ClinicalTrials.gov Identifier: NCT02463149).

Intervention

YCA was developed by one of the lead investigators through 4 iterative classroom demonstration projects with program revisions based on student, teacher, and parent feedback (unpublished data). It promotes intake of a whole-food, plant-based diet by giving youth the opportunity to (1) explore the connections between foods, the food system, and health and (2) create meals “from scratch” as young chefs in training.

Guided by the tenets of Social Cognitive Theory,^42,43 YCA is comprised of 6 2-hour sessions (Table S1, Supporting Information presents theoretical linkages). Each session begins with a nutrition lesson focusing on an aspect of health or the food system (Table 1 presents curriculum detail). Nutrition lessons were linked to academic standards in math, literacy, and social studies. For example, students calculate their dietary percentage from plant foods, locate the origin of various cuisines on a world map noting distinctive ingredients and flavorings, and compare local and global food systems. Following the nutrition lesson, students divide into smaller teams to learn cooking techniques, knife skills, and explore ingredients by preparing recipes from world cuisines. Each YCA lesson ends with students eating the foods they prepared while discussing the dishes in constructive and descriptive ways. Students take home parent newsletters and at-home cooking assignments.

Table 1.

Youth Chef Academy: Culinary and Nutrition Literacy Curriculum Overview

Lesson	Focus of the Food and Nutrition Lesson	Focus of the Cooking Lesson	Recipes
Lesson 1: Kitchen basics			Tossed white bean salad with lemon vinaigrette
	• Explore kitchen equipment • Identify kitchen dangers and learn to prevent accidents including safe food handling • Learn and use measurement tools and units	• Hand washing • Wash and handle salad greens and other vegetables • Culinary methods and skills—measure, whisk, emulsify, peel, and grate
Lesson 2: Food today			Perfect stove-top popcorn
	• Explore how and why food has changed in 100 years • Identify and describe whole and highly processed foods • Build awareness of the processes that transform whole foods	• Recipe reading • Teamwork • Learn and apply new culinary terms, methods and skills—stove usage and safety; stove-top popcorn • Map and explore ingredients	• Bombay behl popcorn • Cajun popcorn • Cinnamon-sugar popcorn
Lesson 3: Informed choice			Dips made with legumes
	• Understand nutrient and calorie density • Analyze food packages • Discover the economic and nutritional impact of processing on food	• Recipe reading • Teamwork • Learn and apply new culinary terms, methods and skills—pureeing, folding, grate peel and garnish • Map and explore ingredients	• French white bean dip • Mexican black bean dip • Hummus • Thai peanut dip
Lesson 4: Plant power			Whole plant salad—students assemble a chopped salad using all the parts of a plant
	• Describe a plant-based diet and understand the health benefits • Identify different parts of the plants we eat	• Learn and practice knife skills—chop and rock chop • Learn cuts for various vegetables
Lesson 5: Whole plant foods and your health
	• Recall one’s diet • Assess one’s diet using a modified My Plate • Calculate ratios and percentages of dietary components • Strategize ways to increase intake of whole, plant foods	• Recipe reading • Teamwork • Learn and apply new culinary terms, methods and skills—soup preparation, sauté, boil, simmer • Reinforce and expand knife skills—slice and mince • Map and explore ingredients	• Korean vegetable soup • Guatemalan vegetable soup • Spanish gazpacho • Italian minestrone
Lesson 6: Seeds—grains and India			Indian cuisine
	• Learn what a whole grain is • Learn about the food system comparing a local food cooperative to a national supermarket chain • Explore and identify types of whole grains, preparation processed grains and refined grains	• Recipe reading • Team work • Learn and apply culinary terms, methods and skills—tempering spices; pilaf and grain dish preparation • Reinforce knife skills and practice new cuts • Map and explore ingredients	• Vegetable pulao • Red lentil dal
Supplemental materials	• Student food journals: organize all relevant food and cooking information—record outcomes of recipes, reflections on foods and lessons, define new culinary and food terms, keep recipes, handouts, homework assignments, etc. • Parent newsletters: inform parents of lessons, recipes prepared in class, and give homework assignments. Also provides parents with useful culinary and nutrition information as well as give recommendations for exploring the local food system with their child. • Field trip report: document local food system field trips that students can take with their families or as a class • Recipe reports: document student experiences when they cook at home for or with their families • Videos: students watch and review proper knife skills and cooking techniques that will be used in the recipe • Exploration stations: used during the cooking lesson, they provide students a meaningful, hands-on way to reinforce and expand on concepts and skills introduced during the food and nutrition lesson while students are between cooking activities

One of 2 instructors led the entire class in the weekly session and supervised the classroom support team. Both instructors were members of the study staff. One was a registered dietitian with extensive experience in school-based culinary interventions and the developer of the YCA curriculum. She trained, observed, and coached the second instructor. The classroom teacher and 2 dietetic interns completed a 2-hour training and provided classroom support. Table 2S presents a summary of key personnel and financial resources needed for YCA implementation.

Instrumentation

Students completed a survey measuring the primary outcomes and other variables. The survey was compiled from previously published measures, adapted measures, and measures created for this study. Cognitive interviews were conducted with several youth to ensure comprehensibility of survey items and acceptability of length.

The primary outcome measures were times per day F/V, VG, and WG consumption. Other measures included psychosocial, knowledge, and engagement variables potentially related to the primary outcomes.

Primary outcomes

Fruit and vegetable consumption was measured using validated, self-report items from the Centers for Disease Control and Prevention (CDC)’s Youth Risk Behavior Surveillance System (YRBSS).^44,45 Seven items (2 fruits, 5 vegetables) assessed F/V and VG consumption over the past week with responses ranging from 0 to 4 or more times per day using a 7-point scale. Eight items assessing WG consumption were adapted from the Power of 3: Get Healthy with Whole Grains Foods program adult survey (University of Minnesota Extension). The recall frame for WG consumption and the response scale were changed to align with the YRBSS questions. Responses were reported in times per day. Internal validity of the modified scale in the study population was moderately strong (Cronbach’s α =.72).

Other measures: Readiness to increase F/V consumption

The survey included 2 questions from the University of California Fruit and Vegetable Inventory to assess readiness to increase F/V consumption that were validated in adult women.⁴⁶ The 5 response choices were based on the stages of change model and ranged from “I am not thinking about eating more F/V (precontemplation),” to “I am already eating 3 or more servings of F/V a day (action).”

Self-efficacy

Self-efficacy measures asked questions in 3 domains: confidence in tasting new foods (1 question), confidence in personal cooking skills (3 questions), and confidence in ability to consume 5 servings of F/V a day (2 questions). The self-efficacy questions on tasting new foods and cooking skills were constructed and validated in adults for the Cookwell⁴⁷ community-based cooking skills intervention and used a 7-point Likert scale ranging from “extremely confident” to “not at all confident.” Internal validity in this study population was modest (Cronbach’s α =.6). By adapting the Cookwell format, YCA investigators created the items to assess confidence in ability to consume 5 servings of F/V a day.

Nutrition knowledge

Nutrition knowledge was assessed using 17 survey items compiled by study investigators. One question assessed understanding of general health benefits of F/V consumption.⁴⁸Five items from the Power of 3 survey asked youth to identify WG items from paired food choices. One question regarding WG food identification was used from the University of Georgia Whole Grain Intake survey for adults.⁴⁹ Another question on WG food identification was adapted from that survey.⁴⁹ One question from the Hawthorne Unified School District asked students to identify a healthy snack from a list of options (unpublished). Study investigators created 8 items to assess knowledge about food processing topics specific to the YCA curriculum.

WG preference

Students completed a measure from the Power of 3 Program youth survey to indicate preference for 7 paired WG and non-WG choices.

Vegetable preference

Vegetable preference was measured using a 26-item tool adapted from Cook-shop’s model for kindergarten-sixth grade youth (5- to 12-year-olds).²⁸ Study investigators modified the list of foods to those relevant to YCA and adapted the wording of response choices. Internal validity in the study population was high (Cronbach’s α =.81).

Student engagement

Student engagement, defined as behavioral and emotional participation in academic activities in the classroom, was measured by a 20-item tool developed by Skinner et al⁵⁰ (Cronbach’s α =.85). Response options were on a 4-point Likert scale ranging from “very true” to “not true at all.”

Procedure

The survey was paper-and-pencil, closed-ended, and took approximately 30minutes to complete. To enhance comprehension, students had the survey in front of them while a study staff member read the survey aloud to the whole class. To minimize reporting bias, this study staff member was not present in the classroom for any of the YCA sessions. The survey was administered approximately 1 week prior to the first YCA session (baseline) and approximately 7 weeks later, after the sixth YCA session (post-intervention).

Data Analysis

The primary outcomes were changes in times per day of (1) F/V consumption, (2) VG consumption, and (3) WG consumption after the 6-week program. Secondary outcomes included vegetable and WG preferences; self-efficacy for cooking, tasting new foods, and eating 5 servings of F/V per day; student engagement; readiness to increase F/V consumption; and nutrition knowledge.

Two-sample, t-statistic-based confidence intervals (CIs) were used to assess the unadjusted differences between pre-post changes in primary outcomes for intervention and control groups.

Missing data was handled using multiple imputation of item missingness (when full surveys were missing, no imputation was attempted). This was accomplished using the mi package in the R software program.

Multilevel models with school effects were used to estimate associations between the primary and secondary outcome variables and the intervention after adjusting for clustering by classroom. Matching variables (student ethnicity, grade level, mobility, and free/reduced-price lunch) were included as random effects at the school level to adjust for any inconsistencies between matched schools. Other confounders included at the individual level were attendance rate, sex, and baseline values of all primary and secondary outcomes. Attendance rate was the percentage of days in the semester of YCA implementation that a student was in school. These data came from school district records.

The cutoff p=.05 was used to establish statistical significance in all analyses.

RESULTS

Final sample size was 248 students (125 intervention, 123 control). Fifty-nine of these students missed the survey at baseline and/or posttest. This was likely due to absences or other random mechanisms. After imputation of skipped survey items, there were 195 students included in analyses (100 intervention, 95 control). There is little evidence of differences in baseline characteristics between those who did and did not complete the posttest (full results available from authors); only the stage of change for vegetable consumption showed a significant difference, with those not completing the posttest scoring an average of 0.5 points lower (95% CI: −0.88, −0.14).

Classroom size ranged from 30 to 34 students. Classes had high rates of participation in the free/reduced-price lunch program and were racially/ethnically diverse (Table 2). The average rate of school change per year (school mobility) for the intervention and control groups were 12.4% and 7.9%, respectively, though this is not a statistically significant difference. Intervention and control groups varied significantly with respect to race/ethnicity, with control classrooms having a higher proportion of white and other race/ethnicity students. One method for evaluating the potential impacts of baseline imbalance on outcomes is a sensitivity analysis. In a separate analysis, consumption of F/V and VG were shown to be quite robust to possible unmeasured confounding (paper forthcoming).

Table 2.

Classroom-Level Characteristics of Youth Participating in the Youth Chef Academy Pilot Study, by Intervention Condition (N = 8 Classrooms)

Classroom Characteristic	Intervention Classrooms (N = 4)	Control Classrooms (N = 4)
Students receiving free/reduced-price lunch	88.9%(74.1, 100.0)*	79.6%(53.1, 92.0)	p = .35†
Race/ethnicity			p < .0001‡
African American	47.3%(0, 95.2)	22.0%(4.0, 56.2)
Hispanic	37.2%(3.4, 100)	29.6%(6.2, 72.0)
White	14.6%(0, 44.4)	33.5%(6.2, 54.5)
Other	0.9%(0, 3.7)	15.0%(4.0, 31.2)
School mobility	12.4%(0, 33.3)	7.9%(0, 15.6)	p= .57†
Mean attendance (percent days in school during semester of YCA)	93.1%(91.2, 97.1)	95.1%(94.7, 95.4)	p= .18†

^*Values given are mean (min, max). The range is used rather than a conventional confidence interval due to the small number of observations and high variability in most characteristics.

^†Differences between groups assessed using 2-sample t tests.

^‡Differences between groups were assessed using chi-square tests of independence.

Three process indicators for this study were classroom attendance, instructor effect, and intervention fidelity. Mean attendance rate was high at 93.1% of school days in the semester of YCA implementation in intervention classrooms and 95.1% in control classrooms (Table 2). Attendance rate was not a significant predictor in any of the models. There was no evidence of differences in student outcomes by YCA instructor (p=.25 for WG, p=.37 for F/V, p=.94 for VG). To ensure intervention fidelity, all 6 sessions in the 4 intervention classrooms were fully implemented by 1 of 2 members of the research team.

Baseline values of primary and secondary outcome measures were comparable between intervention and control groups (Table 3). A borderline significant difference was observed for self-efficacy for consuming 5 servings of F/V per day. The average percent of school days attended in the semester of YCA implementation was statistically significantly lower in intervention classrooms; the difference corresponds to only about 3.5 days of school missed. Of note, both groups reported high F/V, VG, and WG consumption at baseline.

Table 3.

Baseline Characteristics and Measures for Youth Participating in the Youth Chef Academy Pilot Study, by Intervention Condition (N = 195)

Baseline Characteristic	Intervention (N = 100)	Control (N = 95)
Percent female	51.4(50.2)*	54.3(50.1)
Percent school days attended	93.0(0.08)	95.1(0.04)†
Fruit and vegetable consumption (times per day)	4.3(3.1)	4.5(3.6)
Whole grain consumption (times per day)	4.3(4.4)	3.4(3.1)
Vegetable preferences (0–78)†	37.4(10.0)	39.6(10.1)
Whole grain preferences (0–7)	2.9(1.7)	2.9(1.7)
Self-efficacy for cooking (0–18)	16.5(4.5)	15.4(4.6)
Self-efficacy for tasting new foods (0–6)	3.2(2.1)	3.3(2.3)
Self-efficacy for eating 5 servings of fruit and vegetables per day (0–12)	9.4(3.7)	10.4(3.3)‡
Student engagement (0–60)	38.7(8.8)	40.8(8.6)
Readiness to increase fruit consumption (0–3)	2.0(0.8)	2.1(0.9)
Readiness to increase vegetable consumption (0–3)	1.7(0.9)	1.7(1.1)
Nutrition knowledge (0–17)	9.6(2.8)	10.0(2.7)

^*Cell values are mean (SD) unless otherwise noted.

^†Scales include range of possible values as Scale Name (min-max).

^‡Difference significant at p < .05 using a 2-sample t test.

Raw estimates of the primary outcome yielded a change of −0.61 times/day of F/V consumption in the control group, compared to 0.21 times/day in the intervention group (difference 0.82, 95% CI [−0.14, 1.78], p=.096). A difference in the direction of change was also observed for the change in VG consumption: −0.53 times/day in the control group and 0.05 in the intervention group (difference 0.58, 95% CI [−0.15, 1.31], p=.12). For WG consumption, we observe a change of −0.7 times/day in the control group compared to −1.0 times per day in the intervention group (difference −0.3, 95% CI [−1.3, 0.7], p=.56).

For the primary analyses, we excluded observations where the change in consumption was more than 10 times/day over 6 weeks; these values were both unrealistic as observations and destabilized model estimation. Three cases were excluded for analyses of WG consumption and 2 for analyses of F/V consumption (no reported changes in VG consumption exceeded this threshold). Attendance rate was transformed (using the arcsin-square root transformation) to reduce skewness and linearize the relationship with changes in consumption.

After adjustment for matching variables and individual-level confounders, we observe a significant increase in F/V and VG consumption, and a marginally nonsignificant increase in WG consumption (approximately 1.2, 0.9, and 0.9 times/day, respectively) (Table 4). Baseline consumption of F/V was a significant predictor of change in F/V and VG consumption, while stage of change for fruit consumption had a significant association with change in vegetable consumption. Baseline WG consumption, stage of change for fruit consumption, and baseline self-efficacy for cooking were significant predictors of change in WG consumption. Consumption at baseline was negatively associated with the change in consumption in all cases, as expected. Children who consume F/V or WG frequently at baseline are unlikely to increase their consumption much further due to constraints on possible intake of food in any given day.

Table 4.

Multilevel Model Results for Primary Outcomes of Change in Fruit and Vegetable (F/V), Vegetable, and Whole Grain Consumption From Baseline to Posttest Among Youth Participating in the Youth Chef Academy Pilot Study (N = 195)

Variable	F/V Model Coefficient (N = 113)	Vegetable Model Coefficient (N = 116)	Whole Grain Model Coefficient (N = 115)
Intervention	1.2*	0.9*	0.9†
F/V consumption	−0.6***	−0.4***	−0.07
Whole grain consumption	0.1	0.003	−0.7***
Readiness to increase fruit consumption	0.4	0.5**	0.6*
Readiness to increase vegetable consumption	−0.2	−0.2	−0.3
Attendance rate (transformed)	−0.04	0.1	−0.8
Male	0.2	0.5	−0.03
Self-efficacy for cooking	−0.1	−0.04	−0.1**
Self-efficacy for tasting newfoods	−0.07	−0.3	−0.1
Self-efficacy for consuming 5 F/V servings per day	0.04	1.0	0.002
Student engagement	−0.02	−0.02	−0.01
Nutrition knowledge	0.05	0.06	0.1
Vegetable preferences	−0.01	−0.02	−0.02
Whole grain preferences	−0.07	−0.1	0.1

Significance:

^†p < .10

^*p < .05

^**p < .01

^***p < .001.

Variables used for matching schools (race/ethnicity, grade level, mobility, and free/reduced-price lunch) are included as school-level random effects. All variables are as measured at baseline.

In the secondary analyses (Table 5), vegetable preference, student engagement, and nutrition knowledge showed significant intervention effects. Notably, the effect on vegetable preference changed from a statistically significant increase to a statistically significant decrease after imputation; suggesting that the result is more uncertain than the p-value would imply. As such, vegetable preference was not included as a significant finding.

Table 5.

Estimated Effect of Intervention on Change in Secondary Outcomes From Baseline to Posttest Among Youth Participating in the Youth Chef Academy Pilot Study (N = 195)

Variable	Estimate (SE)	p-Value
Vegetable preferences	−4.4(1.7)	.002
Whole grain preferences	0.6(0.3)	.10
Self-efficacy for cooking	0.9(0.8)	.28
Self-efficacy for tasting newfoods	0.5(0.3)	.14
Self-efficacy for eating 5 servings of fruit and vegetable per day	−0.3(0.5)	.48
Student engagement	2.3(1.0)	.03
Readiness to increase fruit consumption	0.1(0.3)	.56
Readiness to increase vegetable consumption	−0.05(0.2)	.80
Nutrition knowledge	1.7(0.4)	<.001

Model includes variables used for matching schools (race/ethnicity, grade level, mobility, and free/reduced-price lunch) as school-level random effects and all variables used in primary outcome models (see Table 4).

DISCUSSION

These data support our primary hypothesis that, compared to the control classrooms, students exposed to the YCA curriculum would demonstrate greater improvement in F/V and VG consumption over the 6-week intervention period. Our hypothesis about greater improvement in WG consumption among students exposed to YCA was not fully supported with a positive increase (0.9 times per day) approaching but not reaching significance. In comparing our findings (reported in times per day consumption) to the literature, it should be noted that some studies report consumption in servings per day; these may not be equivalent measures. With that consideration, Evans et al⁵¹ found an increase of 0.25 portions per day of F/V in a 2012 meta-analyses of 21 school-based nutrition programs. Examining cooking programs specifically, 2 studies reported statistically significant change in F/V consumption and showed increases of 1³² and 2³³ servings per day. A cooking program reporting significant positive change in WG consumption could not be found. The Wang et al’s³² and Brown and Hermann’s³³ programs show success in changing F/V consumption. With a constellation of features not fully present in these 2 programs, YCA adds to this evidence as an in-school, in-depth culinary program for middle-school-aged youth targeting F/V and WG consumption and linked to academic standards. Future research should examine whether treatment effects persist over time as well as whether increases in times per day consumption translate into increases in consumption volume.

Significant changes were found in nutrition knowledge in the intervention group compared to the control group. Nutrition knowledge⁴¹ has been shown to be a determinant of dietary intake; indicating that this variable may be a pathway through which an intervention can influence dietary change. Similar to YCA, studies that specifically used cooking as a means of changing knowledge related to eating F/V generally reported positive results.^{24,28–30,32–35,38,52} Differing from previous studies,⁴¹ YCA did not produce significant changes in any of the measured self-efficacy variables. It is possible that the self-efficacy measures in this study were not robust enough. This finding suggests that YCA produced change in a known determinant of dietary behavior change, helping focus culinary-based nutrition education programs on key elements for change and possibly increasing feasibility by reducing measurement burden.

Relevant to school administrators and teachers, student engagement (behavioral and emotional participation in learning activities across the school day) improved significantly. This finding suggests that an experiential program like YCA can increase general interest in the classroom beyond the time spent in the program itself, which could be an important incentive for school administrators to add experiential nutrition education to their curricula.

Public health advocates have noted that school-based cooking education could improve long-term dietary consumption in youth.^24,53 Linking culinary curricula to academic standards increases the feasibility of adding programs like YCA to an already-stretched school day. Other feasibility considerations remain for optimal implementation effectiveness. For example, homework assignments, including home cooking, hone skills and encourage changes in the home eating environment. However, monitoring homework completion can be challenging. The homework completion tool used in this study was not consistently completed by the students even though anecdotal reports revealed that most students cooked at least one, if not several, of the YCA recipes at home. Use of incentives or other monitoring tools, such as social media, should be researched to improve reporting of these data, as should strategies to increase parent involvement in supporting cooking at home as noted by Fordyce-Voordam.⁵⁴ Additionally, while several studies measured confidence and perceived self-efficacy in cooking skills or attitudes toward cooking,^55–58 no method for measuring actual cooking skills in youth has been validated other than through direct observation, which may be unfeasible depending on sample size and constraints of the classroom setting. Development of such a method would constitute a significant contribution to the literature in this domain due to limitations of using measures of perceived ability as a proxy for actual ability.

Limitations

The nonequivalent control group design can result in the intervention and control groups having different characteristics at baseline. Statistically significant differences in the outcome variables between groups at baseline were not found. There were significant differences in race/ethnicity between groups at baseline, which could impact generalizability. This design is common when groups are naturally assembled, like classrooms, and is considered quite strong.⁵⁹ Another consideration is that the approach used for classroom selection introduced the potential for selection bias, though our sensitivity analysis showed our outcomes were quite robust to unmeasured confounding. Considering measurement bias, the complete survey and some of the individual measures were not previously tested in the population of interest, though we calculated internal validity where possible. The YRBSS questions are best suited to population-level monitoring rather than assessment of individual change. Additionally, assessment of the outcomes was based on self-reported data, which can be subject to bias related to accurate reporting of food consumption.

Conclusion

Nutrition education incorporating hands-on cooking skills, food system knowledge, and links to academic standards has potential to improve dietary patterns among youth. Practitioners can consider whether incorporating hands-on cooking lessons into existing nutrition education programming could improve behavioral outcomes. Future research should examine experiential nutrition education in the school setting to determine if a higher frequency or intensity of this type of curriculum impacts the program outcomes and longer-term maintenance of postintervention behavior changes.

IMPLICATIONS FOR SCHOOL HEALTH

The results of this study demonstrate that culinary-based nutrition education can successfully be linked to academic standards in math, literacy, and social science while developing tangible skills and changing dietary behavior in as few as 6 lessons. Urban schools serving low-income students struggling with low attendance and test scores may want to consider including hands-on cooking-based education as this project provides evidence that an active learning model can increase student engagement in classroom activities beyond the nutrition curriculum itself.

Since this study, demand for the YCA program has increased greatly with teachers and administrators citing student enrichment as the key attribute that makes this program a desirable addition to their classroom. Enrichment education programs, including culinary arts, are frequently the first programs eliminated when schools suffer budget cuts. The district where YCA was tested is no exception, and enrichment programs have been greatly decreased over the years. Interdisciplinary curricula like YCA, which promote health while teaching academics through fun, hands-on learning, provide schools with a cost- and time-effective solution to improve students’ academic performance while still providing these enrichment educational experiences.

Supplementary Material

Supplementary Table 1

Table S1: Social Cognitive Theory in the Youth Chef Academy Program

Supplementary Table 2

Table S2: Summary of Personnel and Financial Resources Needed for YCA Implementation