Culinary Education Programs for Children in Low-Income Households: A Scoping Review

Child obesity in the United States is at an all-time high, particularly among underserved populations. Home-cooked meals are associated with lower rates of obesity. Helping children develop culinary skills has been associated with improved nutrition. The purpose of this study is to report results from a scoping review of culinary education interventions with children from low-income families. Three databases and hand searches of relevant articles were examined. Retained articles met inclusionary criteria. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed, as appropriate. A data extraction template was developed. Data were independently extracted and verified. Only nine out of 370 articles met the inclusionary criteria and were included in the review. Most interventions were school-based, used a quasi-experimental design, and recruited minority children. Children-only was the primary intervention focus. Primary outcomes were mostly psychosocial from child self-report. Most interventions focused on children only and were guided by Social Cognitive Theory. Most reported stakeholder involvement; however, type and degree varied. All had an in-person component; only one used technology. Few reported training program leaders. Culinary education programs for children from low-income families could benefit from a broader theoretical grounding, program leader training, and greater parental involvement.


Introduction
Child obesity in the United States is at an all-time high. Among 2-19-year-olds, 35.1% are overweight and, of these, 18.5% are obese [1]. However, the risk is not equally distributed, with alarming disparities observed based on race/ethnicity [2] and household income [3]. Finding effective ways to overcome these disparities in obesity risk is a national health priority [4].
Although not a prerequisite for a healthy diet [5], consuming home-cooked meals is associated with lower rates of obesity [6] and better diet quality in both adults and children [7][8][9][10][11]. Alternatively, meals prepared outside the home are associated with poorer food choices [12][13][14], greater energy intake [15], and higher body mass index [16]. Time spent on home food preparation has decreased [17], with fewer families preparing and consuming home-cooked meals [18]. People are purchasing foods, such as fast foods, and consuming them at home [15]. Home food delivery is popular [19], with reasons ranging from not wanting to cook to saving time [20].
Inclusionary criteria included intervention studies that conducted cooking and culinary skills classes, parents and/or school-age children (5-18-year-olds) and reported psychosocial and/or behavioral outcomes. Exclusionary criteria included conference abstracts, review articles, programs for solely college students or adults (that did not pertain to eating habits of children), noninterventional studies (descriptive, qualitative, or cross-sectional), and studies that did not have a focus on a low-income population.
A total of 370 articles were screened using the inclusionary and exclusionary criteria outlined above. After screening articles by title and assessing them by reading the articles, 35 full-text articles were identified that met initial inclusionary criteria. A deeper review of articles reporting interventions conducted with children and/or their families living in low-income areas were further examined to identify stakeholder involvement, adaptations/tailoring for low-income families, and program characteristics (i.e., leader training). Articles not clearly identifying the audience was primarily low-income (defined as >50%), not published as a full journal report, not peer-reviewed (i.e., theses, dissertations), conducted outside the United States, and/or did not exclusively focus on school-age children and/or their families were excluded from further review. Nine studies met the review criteria and were included in the focused review reported here ( Figure 1).

Data Extraction
Data from the initial search were independently extracted by two authors; results were compared, and differences reconciled (PPL, DT). The articles meeting the second set of inclusionary criteria were further screened by two independent extractors (GM, DT) to identify characteristics of

Data Extraction
Data from the initial search were independently extracted by two authors; results were compared, and differences reconciled (PPL, DT). The articles meeting the second set of inclusionary criteria were further screened by two independent extractors (GM, DT) to identify characteristics of the studies (Table 1), designs (Table 2), and interventions (Table 3). A third extractor (JD) reviewed and confirmed tables.

Results
Nine studies met the criteria. Most used a quasi-experimental design and collected data at baseline and post-assessment; only two studies were randomized controlled trials. Among all studies, one study had multiple assessment points and one conducted post-intervention focus groups. All studies recruited children; however, two also recruited parents. Sample size ranged from 89-1204 participants. Eight studies recruited participants exclusively from schools, including after-school programs. Only one study recruited participants from subsidized housing complexes, churches, and community centers in addition to schools. Of the 9 studies, 7 recruited from schools with a majority of students eligible to receive free/reduced priced lunches; of the remaining two, one recruited from a school located in a low-income school district, and in the other study, families had to qualify for public assistance to be eligible to participate in the study. Six studies provided family-level socioeconomic status (SES) data; in these studies, nearly all participants qualified for free/reduced price lunch, and one reported that most families who participated had low or very low food security. In all but one study, most participants were of from an ethnic minority group (Black/African American, Hispanic/Latino). All studies were conducted in the United States: four were in Western region of the country, one in the Southwestern region, and two each in the Midwestern and Northeastern regions of the country. All studies collected data from children; three also collected data from parents. The primary method of data collection was self-report survey; however, one study conducted visual plate waste inspections, and one collected anthropometric data and offered an optional fasting blood sample. One conducted post-intervention focus groups with parents. All studies reported positive outcomes in psychosocial variables (e.g., preference, self-efficacy, etc.) ( Table 1).
Five studies reported using a theoretical framework to design the study. The most common theoretical framework was Social Cognitive Theory (SCT); two studies also used Self Determination Theory (SDT) in addition to SCT. Six studies reported involving stakeholders at varying levels during intervention design. All studies reported adapting the intervention for low-income families, although the type of adaptation varied greatly (Table 2).
Interventions included a variety of components. All studies involved an in-person activity such as cooking demonstrations, food preparation, nutrition lessons, tasting sessions, and gardening activities. One study also included a virtual gardening game played on a tablet as part of the intervention. Another study provided a grocery store tour. Support materials (e.g., toolkit) or food were provided to families in three studies. The intervention focus was the child in seven studies and both child and parent in two studies. Parent involvement ranged from none to substantial. Session frequency and duration were variable, ranging from a single 20-min session per month to an immersive school-wide program lasting a school year. A variety of individuals, including classroom teachers, nutrition or food educators, chefs, and volunteers, led the programs. Only four studies mentioned training individuals to lead the intervention. Most studies were conducted at school (e.g., classroom, cafeteria, school garden, after-school program); others were conducted at host sites in the community (Table 3).

Discussion
This review identified nine studies designed to enhance culinary skills in children and/or their families living in low-income households within underserved communities. All but two of the studies were quasi-experimental, suggesting the results should be viewed with caution because of concerns related to internal validity, such as the potential for confounding and regression to the mean [45]. Given that most of the studies were conducted in a school setting where it would be difficult to randomize students to condition, future research is needed to examine ways in which to enhance the robustness of studies using a quasi-experimental design [46].
Although the focus of this review was on children and/or their parents within low-income households, a key finding was that most participants were Hispanic or Black/African American. This finding is not surprising, given the well-documented racial and ethnic disparities in income seen in the United States [47]. However, this suggests that culinary education interventions for low-income children should also consider race/ethnicity when designing the intervention. Interventions that reflect a deep cultural sensitivity and awareness of cultural norms and values in an effort to increase perceived personal relevance, usefulness, and intervention uptake is vital [48,49]. The studies included in this review reported some degree of cultural adaptation; however, the descriptions were relatively sparse. Future research should be more explicit in the steps taken to ensure cultural relevance.
Behavioral theory guided five of the identified interventions. The most commonly cited theory was SCT [50], a theory often used to guide interventions focused on dietary change [51,52]. SDT [53], a theory focused on enhancing autonomous (i.e., self-directed) motivation was also used by two of the intervention studies. Given that motivation is an important component of sustained behavior change [53] and its success at explaining behaviors related to diet and obesity such as physical activity [54], future research should investigate additional ways to design culinary education programs guided by SDT. Four of the interventions did not identify a theoretical framework. This is concerning because theory codifies what is known about a particular behavior and provides a framework for predicting and explaining behavior [55]. Therefore, it is a necessary ingredient of behavior change interventions [55]. Of the studies reporting a theoretical grounding, few described how theory guided intervention development and/or used it to explain the intervention results. This is not uncommon in behavioral research, and there have been calls to more explicitly describe how theory was applied in the design of an intervention [52,56]. Future research should investigate which theory or combination of theories is most effective at promoting culinary skills to low-income children.
Most of the studies reported that stakeholders were involved in intervention development; however, the type and degree of involvement, and who was defined as a stakeholder, varied greatly. Stakeholder involvement (i.e., the individuals, groups, or organizations affected by the research [57]) is an important aspect of intervention development [58] with promising implications for the design of effective interventions [59]. Future research should investigate ways in which to systematically engage stakeholders throughout the design process, and evaluate the association between stakeholder involvement (i.e., type, extent) and intervention effectiveness. This will contribute to the design of more effective interventions.
The child alone was the primary intervention focus in most of the interventions. Because parents are gatekeepers of the home environment [60], it would be advantageous to include parents in culinary education interventions for children. Therefore, future research should investigate ways to design culinary education programs that include both children and parents.
A variety of components was included in the interventions. All studies involved in-person activities, which is a common delivery mode for dietary interventions [52]. Given the popularity of videogames [61] and the broad ownership of devices on which games can be played [62,63], it is interesting to note that only one of the interventions included a digital component. Videogames have been found to be effective at modifying the dietary intake of children [64]. Technology-based interventions may be particularly salient in school-based culinary education programs; students report using mobile technology for schoolwork, and some schools provide students access to tablets and/or computers in the school environment [65]. Therefore, future research should identify ways in which to use technology to develop culinary education programs for low-income children and/or their families.
Program leaders varied from teachers to registered dietitians and chefs. However, it was somewhat surprising that few interventions mentioned training program leaders to deliver the intervention. Training is likely linked to fidelity, or the degree to which an intervention is delivered as intended [66]. Fidelity has been identified as a determinant of intervention efficacy [52]; thus, identifying ways to enhance fidelity is an important aspect of intervention delivery. It is possible that the type of program leader (e.g., registered dietitian vs volunteer) will influence the form and degree of training needed. Future research should investigate this issue as well as the relationships between program leader training, fidelity, and program effectiveness.
Dose is an important concept in behavioral interventions and represents the "amount" of an intervention intended and received [66]. Intervention dose in the studies included in this review varied from several sessions to an entire school year. Although there have been attempts to identify the ideal dose for behavioral interventions targeting children, no consensus has been reached [67]. Future research is necessary to improve general understanding of dose in culinary education programs, designed for low-income children.
Finally, all studies reported positive outcomes. However, most used self-report measures; only two reported objectively assessed outcomes (e.g., visual plate waste inspections; measured anthropometrics; optional blood work). This is a concern, given the known reporting bias often associated with self-report [68]. Furthermore, only one study included post-intervention focus groups. This is a missed opportunity to understand what it was like to participate in the intervention from the participant's perspective and obtain suggestions for needed modifications. Finally, most studies assessed psychosocial outcomes rather than changes in behavior (e.g., home cooking frequency, nutritional intake). Although psychosocial outcomes are thought to be mediators of behavior [69], it would have been preferable to report intervention effects on behavior. Future research could make important contributions to the literature by reporting behavioral outcomes using objective measures when possible. Post-intervention qualitative research is needed to understand the "experience" of participating in the intervention from the perspective of families, which could ultimately guide the design of more effective and sustainable interventions that reflect the needs and interests of families [70].
As with most research, there are limitations. We limited the review to papers in peer-reviewed journals that were published in English. There may have been unpublished studies or studies conducted in other countries or reported in other languages that examined culinary interventions for low-income children and/or their families. Furthermore, the review exclusively examined studies conducted in the United States, thus limiting its generalizability. Finally, most of the studies included in the review were quasi-experimental, limiting reasonable conclusions regarding causality.

Conclusions
Culinary education for children may provide an optimal avenue for enhancing frequency of home-cooked meals and overall quality of foods consumed during childhood, and potentially in adulthood. Developing these skills may also lead to improved and sustained dietary behaviors and patterns and reduced risk of diet-related chronic disease, including obesity. Additional research is needed to enhance the design of effective interventions that achieve goals of culinary education for children and their families, especially those faced with challenges such as lower income.

Implications for Research and Practice
These findings suggest that greater emphasis needs to be placed on finding effective ways to promote culinary skills to children from low-income families in appealing, culturally appropriate ways. Greater emphasis needs to be placed on developing programs for parent-child dyads, involving stakeholders in program development, using theory to guide intervention content and development, and training program leaders to ensure programs are delivered as intended.