Children randomized to the experimental group were given access to a 6-chapter interactive nutrition comic titled “Game On” containing health messages focused on F/V and water consumption (see Figure 3). One chapter was made available at the beginning of each week for 6 weeks. Comic content was tailored (to include more information related to either F/V or water) based on responses to initial screening questions (asked at baseline) related to child F/V and water intake, child self-efficacy to increase F/V and water intake, and parent self-efficacy to support child in increasing F/V and water intake.

To ensure acceptability and relatability to our intended population, the comic storyline and its characters were informed by extensive formative research (highlighted in Table 1). Additional details related to how the formative research guided the design and development of the comic are described in another manuscript (currently under review). The comic depicts the story of a seemingly average high school student who discovers he has been tasked with the responsibility of saving the people of his world by battling an evil empire that has restricted access to F/V and water. With help from his friends, he must prepare for his epic battle by eating healthy foods and avoiding energy-dense, sugary foods/drinks. Personality traits, physical features of, and the communication/language used by characters in the comic were designed to be age-appropriate and culturally acceptable to both Latino and black urban children. Furthermore, scenarios in which the characters find themselves (eg, eating snacks in a bodega, racing to catch a subway for school) were intentionally incorporated into the story to model typical daily experiences of our intended population.

Health-related messaging was delivered through multiple mechanisms within the comic. Although the comic was primarily in black and white (the style typically observed in manga comics), health-promoting images related to F/V and water were featured in color for emphasis. In addition, characters modeled both healthy and unhealthy food behaviors, and the resulting benefits or consequences were depicted (in both image and text). Interactive features were also embedded into the comic (see Figure 3, bottom), such as orange-colored tap/click icons that either opened pop-up windows (highlighting health facts, food-related fun facts, or character information, which included favorite healthy foods/beverages) or prompted audio/visual effects to enhance engagement of the tool.

The comic was housed on a home page (see Figure 3, top), which comprised other sections including character profiles (highlighting demographic information for each character and personal fun facts) and trivia questions (combination of story-related and health-promoting questions, released as one per chapter). If all 6 trivia questions were answered correctly, the child received a prize of a downloadable screensaver image.

At the end of each comic chapter, one of the comic characters was shown encouraging the child to select a goal to work on for the week (see Figure 4, top left). Clicking on the link in the message opened a new page where the child was able to choose from 8 tailored goals related to either encouraging increased F/V or water intake (see Figure 4, top center). Goals included ones that were child-focused (eg, “I will eat fruits I like [such as grapes or bananas] as a snack” or “I will drink a glass of water when I wake up each morning”) and ones that incorporated the parent (eg, “I will ask my parent to prepare vegetables I like [such as zucchini and tomatoes] for dinner” or “I will ask my parent to drink water with me during meals”). Once selected, a tailored tip was provided to assist the child in achieving that specific goal throughout the week (see Figure 4, top right).

Before selecting a new goal (at the end of the following week’s comic chapter), the child was prompted to self-assess how they did on the goal they chose and focused on for the past week by answering the following question: “How often did you do this in the last week?” (Never, Sometimes, Most of the Time, All the Time), and tailored feedback/encouragement was provided depending on the response chosen (ie, “Congrats! Keep up the good work!” or “Things take time - don’t give up!”). If the child selected “Most of the Time” or “All of the Time” in response to the goal assessment question, the child was rewarded with bonus comic content (eg, the backstory for a specific character). The child was then prompted to select another goal to work on for the following week. A total of 5 goals could be selected (no option to select goal at end of the last chapter) and worked on during the intervention period.

A total of 4 messages were delivered to experimental group children each week (total of 23 messages throughout intervention). Messages included announcing the release of a new chapter by a comic character (see Figure 5, left), a reminder to read the comic, and a reminder to select a goal for each week. Messages were delivered via text and/or email based on participant preference identified during the baseline visit.

Parents/guardians in the experimental group received 6 Web-based newsletters with similar health messages as their child (see Figure 6, left). Newsletter content comprised healthy recipes, healthy feedings tips for the family, links to coupons to support healthy eating, and links to fun community events (eg, fall and winter festivals). In addition, they were given access to the “Game On” comic and character profiles. Newsletters were translated into Spanish and provided to those parents who expressed a preference for Spanish-language materials.

A total of 2 messages were delivered to experimental group parents/guardians (total of 11 messages throughout intervention) each week. Parents/guardians received messages announcing the release of a new newsletter. In addition, they received reminders to encourage their child to read the comic. Messages were delivered via text and/or email based on participant preference identified during baseline. Messages were translated into Spanish and provided to those parents who expressed a preference for Spanish-language materials.

Children randomized to the comparison group received 6 Web-based newsletters with F/V or water information (see Figure 3, right). Similar to the comic, newsletter content was tailored (to include more information related to either F/V or water) based on responses to initial screening questions at baseline related to child F/V and water intake, child self-efficacy to increase F/V and water intake, and parent self-efficacy to support the child in increasing F/V and water intake. Newsletters were housed on a home page, and one was made available at the beginning of each week for 6 weeks. Health-related content for the comparison group was similar to that for the experimental group but was presented in a newsletter (didactic) format. The content comprised healthy eating tips, healthy recipes, diet-related knowledge/facts, health-promoting online games, and a link to selected weekly goals. Of note, comparison group participants were provided access to the “Game On” comic on completion of all data collection in April 2018.

Similar to children in the experimental group, children in the comparison group also had a goal-setting component (see Figure 4, bottom), wherein each week, the child could click the weekly goal link in the newsletter (see Figure 4, bottom left); select from 8 tailored goals to work on, related to either encouraging increased F/V or water intake (see Figure 4, bottom center); receive a tailored tip to provide assistance in achieving that specific goal throughout the week (see Figure 4, bottom right); and self-assess at the end of the week on how they did on the goal over the past week and receive feedback/encouragement. Unlike the goal-setting component in the experimental group, the format in which the goal setting was conducted in the comparison group was non-narrative.

A total of 3 messages were delivered to comparison group children each week (total of 17 messages throughout intervention). Messages included announcing the release of a new newsletter, a reminder to read the newsletter, and a reminder to select a goal for each week (see Figure 5, right). Messages were delivered via text and/or email based on participant preference identified during the baseline visit.

Parents/guardians of the comparison group received the same 6 Web-based newsletters (in English or Spanish) as that of the parents/guardians in the experimental group. However, parents/guardians in the comparison group were not given access to the “Game On” comic and character profiles (see Figure 6, right).

One message (in English or Spanish) was delivered to comparison group parents/guardians (6 messages throughout intervention) each week announcing the release of a new newsletter. Messages were delivered via text and/or email based on participant preference identified during baseline.

A custom-built platform was created to automatically log child and parent user details (created at baseline) and usage details (over 6 weeks of the intervention). User details included username, user type (parent or child), study group (experimental or comparison), type of tailored content (F/V or water), user language (English or Spanish), user email address, user mobile phone number, and start date/time (when username was created at baseline). Usage details included week number of intervention, link clicked, click time, platform used (eg, Mac OS, iPhone OS, Windows 7), and browser used and version (eg, Chrome 38.0, Safari 11.0). For child participants, additional usage data related to goal setting included weekly goal selected, date/time the goal was selected, evaluation at the end of the week (ie, “How often did you do this in the last week?”—Never, Sometimes, Most of the Time, All the Time), and date/time the goal was evaluated.

Questionnaires were administered to the child and parent at several time points throughout the study. Usability was assessed with an adapted version of the System Usability Scale (SUS) [62]; Usefulness, Satisfaction, and Ease-of-use questionnaire (USE) [63]; and a 26-item acceptability/usability measure by Ben-Zeev et al [64] to assess 5 usability domains: usability, usefulness, ease of use, ease of learning, and satisfaction. The child’s questionnaire comprised 30 usability questions—10 questions from SUS, 15 from USE, and 5 from the acceptability/usability measure. On the basis of pilot testing with Latino and black children, modifications were made to tailor the questionnaire according to the literacy levels of our intended population. For example, the item “I found the system very cumbersome to use” was replaced with “I found it awkward to use” and “I would imagine most people would learn to use this system very quickly” was changed to “I think most people my age would learn to use it very quickly.” The parent questionnaire comprised 9 usability questions—2 questions from SUS, 3 questions from USE, and 4 questions from the acceptability/usability measure. On the basis of pilot testing, 1 item was modified. Similar to the child questionnaire, the item “I found the system very cumbersome to use” was replaced with “I found it awkward to use.” As the content of the parent component was presented in a more didactic format compared with that of the child component, fewer usability questions were relevant to include in the parent questionnaire. In addition, usability data in the form of qualitative interviews with child participants (at T2, T3, and T4) and parent participants (at T3 and T4) were collected to supplement quantitative usage and usability questionnaire data.

Process data collected throughout the study assessed the feasibility of implementing the study [65,66]. These data included quantitative/qualitative measures of recruitment and retention (ie, enrollment rate, restrictiveness of eligibility criteria, attrition rate), assessment of resource capacity (ie, staff hours needed for recruitment, participant communication approaches), data collection (ie, length of time to complete online questionnaires), and data reliability (ie, study staff adherence to protocol). Satisfaction with study participation was assessed through questionnaire items with child and parent participants at T4 (ie, frequency and format of communication and study visit scheduling with study staff) as well as qualitative observations of child participants at T1 and T4 while completing questionnaires (ie, verbal and nonverbal expressions indicating frustration, boredom, and confusion).

Child participants completed a questionnaire related to knowledge, outcome expectations, self-efficacy, behavioral intention, attitudes and preferences regarding behaviors associated with F/V, water, junk food, and sugary drinks. A total of 6 questions addressed knowledge (ie, I should eat 1 cup of fruit each day for good health), 20 questions addressed outcome expectancies (ie, Eating vegetables every day will keep me from getting sick), 9 questions addressed self-efficacy (ie, If I decide to not eat junk food every day, I can do it), 10 questions addressed intention (ie, If my parent offers me water, I will drink it), 15 questions addressed attitudes (ie, I think sugary drinks are cool), and 24 questions addressed preferences (ie, Which of the following fruits do you like or dislike?). A total of 84 questions were informed by and modified from the validated ProChildren questionnaire [67] and the validated Reynolds questionnaire [68]. Modifications to wording were made to ensure questions were appropriate for this study and to adjust for literacy levels of our intended population. A recent study conducted among Latino children also adapted the ProChildren questionnaire to measure children’s self-efficacy for eating fruits and vegetables and for consuming water [73].

Child participants completed a questionnaire that assessed the frequency of consumption of F/V, water, junk food, and sugary drinks during the past 7 days. This 17-item questionnaire comprised 8 questions (6 directly from and 2 informed by) from the validated 2017 Youth Risk Behavior Surveillance System (YRBSS) questionnaire [69], which was created to monitor obesity prevalence and related behaviors, among other priority adolescent health issues. Nationally representative samples of students along with selected large urban school districts are engaged in the data collection process. Other studies conducted specifically with black and Latino children exploring dietary and physical activity behaviors have also utilized items from the YRBSS questionnaire [74,75]. Moreover, 6 items were informed by the validated Beverage and Snack Questionnaire, which was tested with a diverse group of children as approximately 45% were from minority populations [70], and it was used in a more recent study that evaluated the efficacy of a serious game on low-income urban public school children’s dietary behaviors [76]. A total of 3 items were internally created (related to assessing the intake of different types of water).

Height and weight of child participants were measured using standardized methods [77]. Height was measured to the nearest 1/8 inch using a digital stadiometer (SECA 264), with the participant fully erect, without shoes, feet together, head in the Frankfort plane, and at the end of a deep inhalation. Weight and body composition were measured using a body composition monitor (Tanita MC-780U) wearing lightweight clothes and without socks and shoes. Weight was measured to the nearest 0.2 lb. Height and weight were measured in duplicate and recorded. A third measurement was taken if there was any uncertainty on the accuracy of height or weight measurements (eg, if height measurements differed more than 0.5 inch). An average of the measurements was used for the BMI calculations. The Centers for Disease Control and Prevention BMI percentile calculator was used to determine BMI percentage [78]. If needed, the average height was rounded down to the nearest eighth of an inch, and the average weight was rounded down to the nearest quarter pound, to accommodate the calculator’s units of measurement.

Parent participants completed a questionnaire that asked about multiple parental feeding practices, specifically including 6 questions related to environment (ie, I offer a second helping of vegetables to my child during meals at home), 4 related to involvement in purchasing/preparing food (ie, I allow my child to help prepare fruit and vegetable dishes for family meals), 7 related to encouragement (ie, I encourage my child to drink water drinks [unsweetened] before sugary beverages), 8 related to modeling (ie, I model drinking water for my child even if it is not my favorite), and 2 related to teaching about healthy food practices (ie, I discuss with my child why it’s important to eat fruits and vegetables.). The 27-item questionnaire was informed by the validated Comprehensive Feeding Practices Questionnaire [71], which has been used in previous studies with low-income Hispanic parents and African American fathers [79,80].

Parent participants completed a questionnaire related to the availability of fruits, vegetables, and water in their home and how often they store fruits, vegetables, and water in a place easily seen by their child. The questionnaire comprised 6 questions and was informed by the validated Home Environment survey [72].