The first step in the development process was a needs assessment and an analysis of the problem. This step was completed through stakeholder (educators, parents, researchers, and community partners involved in the field) consultation, informal conversations with childcare center staff, and a literature review.

Through the literature review, the consumption of pulse-based foods was explored to elicit additional information on barriers and determinants. Additionally, food-based interventions conducted in childcare settings were reviewed to identify gaps and successes that would be helpful for the development of this intervention. For the stakeholder consultation process, at the beginning of the study, an eight-member multidisciplinary team of researchers, graduate students, and community partners was established to conduct problem analysis, fine-tune the PDTK intervention, and participate in recipe selection and taste testing. Two early childhood educators (ECEs) and preschool children (n=23) from a pretest childcare center in Saskatoon, Saskatchewan, Canada, who were not involved in the pilot study or roll-out study, took part in the development of the PDTK intervention. This pretest site in Saskatoon was chosen owing to its proximity to the University of Saskatchewan campus and willingness to participate in the PDTK development process. Informal discussions were carried out with the pretest childcare center staff on findings obtained from the literature review. The literature review revealed that the emphasis of nutrition interventions targeting preschool children in childcare settings has been on the consumption of fruits and vegetables [22-25]. Furthermore, studies indicated that most effective interventions were multicomponent and involved children, parents, educators, and the food environment [26,27].

Determinants of pulse consumption at individual, social, and environmental levels identified from the literature [28,29] are presented in Figure 1. Very few studies in Canada have addressed these determinants. A childcare center intervention study conducted by Froehlich Chow et al [30] focused on increasing educators’ and cooks’ knowledge of pulses that increased their knowledge and utilization of pulse-based foods in these settings.

Recipe books and websites were reviewed to create a matrix of child-friendly pulse-based recipes for lunch and snacks that could be adopted in the intervention using predefined criteria (Multimedia Appendix 1). The Kids Book of Lentils; Pulses; Cooking with Beans, Peas, Lentils and Chickpeas; Lentils for Every Season; The Best of Heart Smart Cooking; and Pulse Canada websites were among the resources reviewed for recipes [31,32]. In addition to pulse-based foods that could be prepared at childcare centers, commercially purchased products, such as puffs, lentil pasta, and biscuits, were also included in the matrix.

The criteria for recipe selection were developed by consulting different guidelines, but mainly focusing on the 2007 Eating Well with Canada’s Food Guide and Saskatchewan Ministry of Education’s Child Care Regulations [33,34]. Criteria considered during the selection of recipes were as follows: (1) capacity to offer 0.5 to 1 food guide serving, with at least two food groups for mixed dishes; (2) whether the food item could be eaten by children independently; (3) risk of choking; (4) availability of ingredients; (5) risk of allergy; (6) cost of ingredients; and (7) ease of preparation. This initial selection process generated a possible list of 30 pulse-based recipes (Multimedia Appendix 1).

The eight-member team, along with undergraduate nutrition students, preschool children, and their parents in the pretest center participated in sensory evaluation of the recipes at different stages of the PDTK development (Figure 2). Initially, the recipes were evaluated by the eight-member team for taste, appearance, and the likelihood of acceptance in a childcare center. This evaluation resulted in a list of 11 acceptable recipes that could be incorporated without any modifications into the PDTK, as well as a few recipes that would require further modification such as reducing strong flavors, increasing the proportion of pulses, and improving texture. After the recipe modification, students in an undergraduate nutrition class were invited to taste four of the recipes that needed modifications based on the team’s evaluation to see if they needed additional adjustments. Further sensory evaluation of selected recipes was then carried out with preschool children at the pretest center to determine their acceptability. A modified method previously used in kindergarten settings was employed for this assessment [35]. Groups of three to five children of similar ages were paired with one researcher. Researchers explained the procedure to the children, and samples of pulse-based foods were presented to them. They could ask for more portions if they wanted. The researchers observed and took detailed notes on whether the children ate all, some, or none of the portions or if they asked for more. The reports from each researcher were used to record how many children tasted the foods, and the results indicated that the majority of children were willing to taste the recipes and expressed their liking. In addition to gathering information on willingness to taste and liking of the pulse-based foods, this process also helped in shaping the final sensory evaluation procedure to be used for the pilot testing project. Parents and their children attending a parent-teacher consultation meeting at the pretesting site also provided an informal evaluation of two of the recipes (lentil smoothie and lentil pizza). Parents tasted the recipes and provided feedback verbally. Most parents indicated that they liked both recipes. Overall, this process served to identify 15 recipes that were incorporated into the PDTK.

Findings from the expert consultation and literature review in step 1 helped to shape the desired objectives of the PDTK intervention. The general program objective was to increase the consumption of pulse-based foods among preschool children in childcare settings. Considering the barriers and determinants related to consumption of pulses by children identified in step 1, the program objective was translated into performance objectives at individual (children), and social and environmental levels (parents, ECEs, and cooks). These are activities that the children, parents, ECEs, and cooks needed to complete to promote behavior change. Parents, ECEs, and cooks were identified as key players who could potentially influence children’s food choices [26,36] and were a focus when planning the intervention. The objectives that focused on children, parents, ECEs, and cooks to increase consumption of pulse-based foods among preschool children are presented in Textbox 1.

Studies have shown that implementing a successful nutrition education intervention requires a theoretical component [37-39]. For the PDTK, two theoretical frameworks were selected to guide intervention development. The frameworks selected for this intervention were the SCT and Piaget cognitive development theory.

The most dominant and extensively used theory for implementing nutrition education programs for children is the SCT [40]. The SCT is known for its comprehensive approach, taking into consideration environmental, personal, and behavioral factors [41], which help to define intervention components [42]. This theory is often chosen because of its “emphasis on approaches that are important to youth, such as positive reinforcement” [42]. The SCT is also considered an effective framework for program development owing to interactions between individuals, their environment, and learning capacity [43]. The constructs of behavioral capability, observational learning, positive reinforcement, and environmental changes were used as guidelines in constructing the PDTK intervention components. Even though the SCT is used as the main theoretical framework, the developmental stage of preschool-aged children makes it difficult to apply all SCT constructs, such as self-evaluation. Recognizing this, the Piaget development theory was used to complement the SCT. The Piaget cognitive development theory offers specific explanations of a child’s cognitive development.

The Piaget cognitive development theory highlights play and self-discovery and provides direction in the preparation of age-specific education content. According to Piaget, there are four developmental stages of children, which progress in a linear fashion. The preoperational period, which is most applicable to PDTK development, focuses on children aged 2 to 7 years, is classified as the second stage of a child’s cognitive development, and involves the development of symbolic thought and consideration of the world through an egocentric perspective [44]. This stage of development is the most applicable to children during the preschool years, as at this stage, a child cannot use logic or combine ideas and only learns by copying the environment, discovering, questioning, classifying, socializing, and tangibly understanding concepts [45]. This copying of the environment, discovery, and questioning result in a schemata perception that influences children’s new experiences [46]. Hence, a new schema can be formed in children’s minds about new foods, such as pulses, if they are exposed to them through educational activities.

Taken together, both theories helped to create the project’s strategy to encourage children to adopt healthier diets and improve pulse consumption. For example, creating social and environmental changes through constant exposure to pulses and pulse-based dishes was expected to increase preferences for and consumption of pulse-based dishes. Previous studies have shown that repeated taste exposure can influence liking and willingness to consume new foods [47,48]. Subsequently, a matrix of objectives matching the performance objectives defined in step 2 was developed based on the constructs of the SCT while taking children’s cognitive development into consideration (Table 1 and Table 2). In addition to the SCT and Piaget cognitive development theory, other theory-based methods were considered and implemented where applicable. These included active learning, sensory learning, and imagery. For example, children were encouraged to actively participate in creating their own pulse-based wrap, taste testing pulse-based buffets, growing pulses in a pot, naming pulses on a picture, and creating crafts using pulses.

During the fourth step, information obtained in the previous steps was organized into program components. The components of PDTK were designed to target the individual and environmental determinants identified in the previous step of IM. For example, the food service guide, which included a guide to cooking pulses in a childcare center, recipes, and shopping tips, could enhance the environmental conditions and reinforce pulse-related knowledge, thereby impacting purchase and consumption of pulses. The objectives pertaining to the behavior of parents and guardians were facilitated through the development of a newsletter as a mechanism to support their healthy eating behavior. The components of the PDTK encompassed the following: (1) a 12-week lesson plan, (2) a food service guide for incorporating pulses in the childcare center’s menu with an example of a 4-week cycle menu, (3) pulse recipes, and (4) four parent newsletters, which are described in more detail below.

This step focused on strategies to promote adoption of the PDTK and the development of an implementation plan. The adoption and implementation strategies included active involvement of primary stakeholders, such as ECEs, cooks, and directors of the childcare facilities. To ensure adoption, each lesson plan and its activities were evaluated for content, flow, and age-appropriateness during the pilot testing of the study. The pilot-testing process would ensure program component evaluation by the major stakeholders.

For implementation and pilot testing, the childcare center was provided with a PDTK manual that incorporated the program components. Additionally, the staff received training in implementing the program. Trained graduate students and ECEs provided the 12 PDTK lessons. The pulse-based foods served for lunch were mainly prepared by the cooks in the childcare center facilities. For pilot-testing, graduate students assisted in the preparation of pulse-based foods used for sensory evaluation. Overall, researchers provided minimal support during the roll-out compared with pilot testing.

In this final step, a plan to evaluate the feasibility and impact of the PDTK program was developed. A multimethod approach with both quantitative and qualitative data collection tools was designed. The quantitative approaches to evaluate feasibility and effectiveness included (1) sensory evaluation, (2) plate waste assessment, and (3) pre-post knowledge testing regarding pulses among children. For qualitative data collection, one-on-one semistructured interview guides were developed to elicit detailed information regarding the PDTK intervention from the ECEs and cooks. The ECEs in the pilot testing childcare center were asked to provide qualitative feedback through a predeveloped weekly lesson plan evaluation form. The lesson plan evaluation carried out in phase II pilot testing provided detailed feedback that was used to revise the PDTK before the roll-out in phase III.

Quantitative data were analyzed using SPSS version 24 (IBM Corp) and SAS version 9.4 (SAS Institute Inc). SPSS was used to generate descriptive statistics and for paired comparisons. Chi-square and McNemar tests were used for paired nominal data to compare the preindividual and postindividual pulse knowledge scores. Data analysis techniques were adapted from a study conducted by Sigman-Grant et al [55]. A paired sample t test was also conducted on spreads (green split pea vs red bean) to determine if there was any difference between sensory acceptance at the beginning and the end of the intervention. SAS was used to generate t test and generalized estimating equation model results for plate waste measurements. The consumption proportions were calculated for each of the intervention and control recipes using the following formula: consumption proportion = sum of the amount of all intervention (or control) foods a specific child eats / sum of the amount of intervention (or control) foods given to that specific child. Using aggregate data, an independent t test was performed to compare the mean consumption and the mean consumption proportions of both the intervention and control recipes. Results were considered statistically significant if a P value of <.05 was obtained.

For qualitative data analysis, the audio recordings of the semistructured interviews were transcribed verbatim. The transcripts were then sent to the interviewees for verification. The verified transcripts were then analyzed in three phases. The first phase involved an initial review of the transcript for common concepts and themes. The process of analysis consisted of reading each individual response to each question several times and highlighting the major ideas that emerged from the transcript. The second phase of analysis consisted of an auditing process, where concepts or themes were reviewed by an independent researcher as a means of verification, to determine if there was common consensus or disagreement with the themes identified in the initial analytical process. After the verification process, the third phase involved rearranging quotes into the selected thematic categories. Subsequently, comparisons were made between categories to describe the findings [56].

Ethical approval was obtained from the University of Saskatchewan Behavioral Research Ethics Board. Prior to data collection, consent and verbal assent were obtained from parents and children, respectively.