We developed 3 key behavioral messages to reduce salt intake in 7- to 10-year-olds: (1) stop using the salt shaker during cooking and at the table, (2) switch to lower-salt foods by checking food labels (focus on bread, breakfast cereal, and cheese), and (3) swap processed salty foods (eg, processed meats, take-out pizzas and burgers, savory sauces, and snack foods) with healthier alternatives.

We assessed the potential effectiveness of these strategies to meet the primary outcome of a 20% reduction in salt intake via dietary modelling and deemed them as suitable (Multimedia Appendix 1 [27-31]). The rationale for the development of these messages is described below.

The selection of food groups to target within the intervention was informed by our previous work, which identified the main dietary sources of salt among Australian children [32]. These are bread (15%); processed meat (9%); savory sauces and condiments (6%); mixed cereal dishes (7%), which include pizza, sandwiches, and hamburgers; cheese (5%); breakfast cereals (4%); pastries (4%), which include meat pies and sausage rolls; and snack foods (4%). As sodium is widespread within the food supply [33], the main sources of dietary salt are diverse in their nutritional composition and include discretionary options, which should be limited in the diet (eg, processed meats and pastries), as well as core foods, which should form the basis of a healthy dietary pattern (eg, bread and cheese) [34]. We have previously reported that core foods provided more than half of all salt consumed (55%) by schoolchildren, with the remaining 45% provided by discretionary foods (unpublished data, 2017). Hence, to reduce salt intake, there is a need for strategies that incorporate messages related to overall healthy eating principles, such as limiting the consumption of discretionary foods, as well as messages related to the consumption of core foods that have lower levels of salt. The large variability of salt content within core foods [33,35] makes finding lower-salt options feasible. On this background, we used the Australian Guide to Healthy Eating (AGTHE) as an evidence-based healthy eating tool to embed salt-specific messages in the intervention. The AGTHE is a food selection guide that groups foods as core or discretionary choices and provides a visual representation of the proportion of the 5 food groups recommended for consumption each day [34].

We also considered the contribution of discretionary salt use to intake. While no data for Australian children are available, in Western countries it is estimated that approximately 6% of sodium consumed comes from salt added at the table and 5% is added during cooking [30,36]. Previously, we have shown in Victorian schoolchildren that reported table salt use was related to overall higher salt intake, as measured by 24-hour urinary sodium excretion [37]. In addition, we have previously reported that the use of table and cooking salt is relatively common among Victorian schoolchildren and their parents (eg, 40% of children reported using table salt and 66% of parents reported using cooking salt) [31]. These findings demonstrate the importance of targeting discretionary salt use.

Finally, to aid in the development of tailored strategies, we considered the contribution of salt from different meals as consumed by Australian children participating in the 2007 Australian National Children’s Nutrition and Physical Activity Survey [27]. We found that foods consumed at dinner contributed the most to daily salt intake (35%) (Multimedia Appendix 2 [27]). Although foods consumed during lunch provided less salt overall (25% of daily intake), these foods had the highest sodium density, indicating the consumption of particularly salty foods at lunch. These findings indicate the importance of targeting lunchtime and dinnertime food intake to help reduce daily salt intake.

Social cognitive theory is frequently used in dietary interventions targeting children [24], and it informed the development of this intervention. Social cognitive theory stipulates that behavior is determined by the reciprocal interaction of personal cognitive factors (eg, self-efficacy, outcome expectations, and knowledge), socioenvironmental factors (eg, observational learning, normative beliefs, social support, barriers, and opportunities), and behavioral factors (eg, behavioral capability, intentions, and reinforcement) [38].Developing strategies that address the constructs within social cognitive theory increases the likelihood of influencing behavior change [38]. In particular, we focused on those constructs previously shown to be related to dietary intake in children. These include self-efficacy [39], intentions (ie, goal setting) [40,41], reinforcements [42], and knowledge [43]. We developed strategies to address the intervention content with reference to Michie and colleagues’ [44] taxonomy of behavior change techniques. We selected behavior change techniques for inclusion in the intervention considering what was appropriate given the intervention mode of delivery (ie, Web based, with no face-to-face contact) and what behavior change techniques have previously been used in effective interventions targeting children’s eating behaviors [24,45]. In addition, we considered previous reports of what behavioral-based strategies were effective in reducing salt intakes. For example, in adults and children, effective strategies to reduce salt intake include providing education on reading sodium information included on food labels, cooking recipes with spices and herbs, information on selecting lower-sodium foods when eating out, and goal setting [18,19,46,47]. Multimedia Appendix 3 shows a list of strategies included in the intervention, mapped to behavior change techniques and social cognitive theory constructs.

Storytelling can be used as a means to communicate health messages and facilitate learning outcomes in children [48,49]. In conjunction with a creative writer and illustrator (Ben Pearmain Illustration, Melbourne, Australia) we created a narrative within which to embed the intervention content. To help engage children, we selected a detective theme as a basis of the story, with a personified dog as the protagonist [49,50]. Comic strips will deliver the story at the beginning of each Web session (Multimedia Appendix 4), which will set the scene for the weekly learning activities for the children to complete.

Following the collection of baseline data, we will mail a starter pack to the family home. The pack will include a 1-page overview of the program; a fridge magnet outlining the 3 key behavioral objectives of the program; AGTHE educational resources (including fridge magnet and pamphlets); a cheat sheet shopping card outlining sodium content targets for bread, breakfast cereal, and cheese; a detective logbook for the child to record weekly goals and stick badges in; and a “stop using the salt shaker” sticker and stickers of badges for completing case files and meeting goals.

Parents will receive a weekly email with access to the Web-based session; alternatively, children will be able to access each session via the study website. We developed the Web sessions using the e-learning software Articulate Storyline 2 (Articulate Global, Inc). To facilitate engagement and learning, the sessions will be packaged as a series of detective case files for the child to solve. Each case file targets key learning objectives related to the behavioral messages of the intervention (outlined in detail in Multimedia Appendix 3). Table 1 provides an overview of each session. Sessions we developed are interactive and include comic strips, characters to introduce key concepts (Figure 1), activities and games, video content, and sound effects. Multimedia Appendix 4 presents the basic format of the sessions. If the child completes the session and solves the case file, a badge is awarded (Figure 2).

At the end of each session (weeks 2, 3, and 4), the child will be asked to set a goal that is related to the key behavioral message for that week. Given the ages of the children, we created predefined options for goals from which the child could select [51]. We will provide information on possible barriers to meeting the goal and solutions to overcome. Children will be encouraged to record their goal in the detective logbook and discuss this with their parent. If the child reports meeting their goal in the following week, they will be awarded a bonus badge (Figure 2).

The first 2 Web sessions that were developed were pilot tested in children aged 8 to 12 years. We recruited children via advertisements and email invitations that we distributed among networks at our workplace (ie, Deakin University). Parent and child consent was provided and ethical approval was granted by the Deakin University Health Human Ethics Advisory Group (Project No. HEAG-H 142_2014). Parents were emailed hyperlinks for their child to access the 2 Web-based sessions. At the end of each Web session, a hyperlink was provided to access an evaluation survey to determine the acceptability and appeal of each session. We invited a subsample of children to participate in a face-to-face interview to provide more detailed feedback on each session. The results section below presents findings from this evaluation.

Parents will receive weekly emails with hyperlinks to access the online newsletter. The newsletters include educational materials relevant to the week of the program and information on the child’s weekly goal (Figure 3). Additional hyperlinks will be embedded within newsletters directing parents to extra resources, such as a video for reading food labels, supermarket cheat sheets for top picks for foods with less salt, and an herbs and spice resource for cooking without salt. The content of newsletters has been reviewed by a dietitian and tested with 2 mothers, and where necessary changes were made to the language, layout, and graphics.

A password-protected study website will be accessible during the intervention and will provide resources for children and parents. We will update the website weekly with relevant resources. The children’s section will include access to the detective case files (ie, Web-based sessions), downloadable versions of the comics, weekly case badges, bonus badges, and handouts describing potential barriers to meeting weekly goals with suggested solutions. The parent’s section will contain access to online newsletters, also available to download in PDF format, as well as other key resources (Figure 4). In relation to recipe resources, we used existing recipes from health agencies, such as the National Heart Foundation of Australia and World Action on Salt & Health, and where appropriate we provide information for modifying the recipe to reduce salt.

We will send 2 to 3 reminder SMS text messages to the parent during each week of the intervention. The messages will provide information related to weekly key messages and goal setting, and will include hyperlinks to access intervention resources, such as online newsletters and materials posted on the website.

The primary outcome of the DELISH program is the change in the child’s salt intake assessed via a 24-hour urine collection. To engage parents, we will provide them with the option of completing a 24-hour urine collection; however, this will not be mandatory. We will follow our previous protocol for collecting 24-hour urine samples in schoolchildren [52]. Briefly, the first morning urine will be voided, and this will be recorded as the 24-hour urine start time. Following this, all urine will be collected, finishing with a final collection at the corresponding 24-hour finish time. Participants will be able to commence the 24-hour urine collection either at school or at home over the weekend. We will instruct participants to complete the postintervention 24-hour urine collection on the same type of day (ie, school or nonschool day) as the baseline assessment. Children will be provided with a 2.5-L widemouthed, rimmed polypropylene bottle to collect urine. To assist with urine collection, we will provide an additional 500-mL plastic handled jug. Verbal and written instructions will be provided. Children will be asked to report any missed collections or spillages on a urine collection slip, which will be returned with the 24-hour urine sample. We will standardize urine volume and electrolytes to a 24-hour period; for example, [24 hours/urine duration (hours)] × urinary measure. We will consider 24-hour urine samples to be complete if no more than 1 collection is reported as missing (provided that we find no differences in urinary sodium by sex between those who report 1 missed collection and those who report no missed collections), duration of collection is within 18-34 hours, total volume is 300 mL or more, and urinary creatinine excretion (mmol/kg body weight/day) is greater than the fifth percentile cutoff point obtained from previous reference data among Victorian schoolchildren (ie, 7-8 years, ≥0.1 mmol/kg body weight/day; and 9-10 years, ≥0.1 mmol/kg body weight/day [52]).

The first morning void of the 24-hour urine collection will be collected in a separate container, before the start of the 24-hour collection, and will be assessed for sodium concentration [53]. Research staff will visit the school on the day preceding the scheduled testing to provide those children completing a school-day 24-hour urine collection with a 500-mL plastic collection container and carry bag, along with verbal and written instructions, so the child can collect the overnight urine sample on waking on the day before starting the 24-hour urine sample at school. Children who complete weekend-day 24-hour urine collections will be provided with the collection materials and instructions for both the 24-hour urine collection and overnight collection on the day of testing at the school. Children will be asked to report onto a urine collection slip any spillages while collecting the overnight urine.

Urine samples will be transported to a commercial pathology company for urinalysis. Total volume of both urine samples will be measured, and each will be assessed for sodium and potassium concentrations using indirect ion-selective electrodes and for urinary creatinine concentration using the Jaffe reaction [54] on the Siemens Advia 2400 analyzer (Siemens Healthcare GmbH, Erlangen, Germany). Per participant, 2×10-mL aliquots will be taken from the 24-hour urine collection for storage and transferred to –80°C conditions. We will use the molecular weights of sodium (23 g/mol), sodium chloride (58.5 g/mol), and potassium (39.1 g/mol) to convert laboratory values of millimoles to milligrams [55].

Trained research staff will collect anthropometric measurements at the school site on the day of testing. Height will be measured to the nearest 0.1 cm using a calibrated portable stadiometer, seca model 220 (Hamburg, Germany). Weight will be measured to the nearest 0.1 kg using a calibrated UC-321 portable electronic scale (A&D Medical, San Jose, CA, USA). Waist circumference will be measured to the nearest 0.1 cm using a Lufkin Executive Thinline W606PM pocket tape (Sparks, MD, USA). Waist circumference will be measured at the end of a normal expiration at the narrowest point between the lower costal border and the top of the iliac crest. For all anthropometric measures, 2 measurements will be taken. If these 2 measures differ by more than 5 mm for height, 0.1 kg for weight, or 10 mm for waist circumference, a third measurement will be taken. Where 2 measurements are taken we will use the mean in the analysis, and where 3 measurements are taken we will use the median value. We will calculate body mass index as body weight (kg) divided by the square of body height (m²). Age- and sex-adjusted body mass index z scores will be calculated using the LMS method [56] with the 2000 US Centers for Disease Control and Prevention growth charts acting as the reference population [57]. Participants will be grouped into weight categories (very underweight, underweight, healthy weight, overweight, obese) using the age- and sex-specific International Obesity Task Force body mass index reference cutoffs for children [58,59].

To assess the dietary intake of children at baseline, the parent will complete an online version of the validated Australian Child and Adolescent Eating Survey (ACAES) [60]. The ACAES is a 120-item, semiquantitative food frequency questionnaire that includes a comprehensive list of foods, enabling the estimation of all macronutrients and key micronutrients. We will compute the nutrient data from the ACAES using FoodWorks (version 3.02.581; Xyris Software [Australia] Pty Ltd). Parents will be instructed to report on their child’s dietary intake for the previous month. The purpose of collecting information from the ACAES is to provide background information on overall dietary intake within the group.

We developed a 29-item online survey assessing the constructs of knowledge (20 questions), attitudes (2 questions), behaviors (4 questions), and self-efficacy (3 questions) related to dietary salt. To simplify the survey and help with readability, we incorporated images into response options. The survey instrument was pilot tested with 3 schoolchildren aged 7-9 years, 2 mothers, 1 primary school teacher, and 2 dietitians to check for readability, age appropriateness, and content. Following this testing, we made small modifications to the survey. An online readability tool graded the survey instrument as “easy to read” and appropriate for 8- to 9-year-olds [61]. We will assess test-retest reliability of the questionnaire in a separate sample of 45 schoolchildren.

The knowledge questions include declarative (eg, food sources of salt, relationship between salt) and procedural knowledge (eg, reading food labels to pick foods with less salt). Response scales include multiple choice options, along with “true,” “false,” and “not sure” responses, with the latter included to discourage guessing. Correct answers will be scored 1 point and incorrect answers, including “not sure,” will be scored as 0, with a total maximum knowledge score of 20. Attitude questions assess the importance of salt in food to make it tasty (eg, “Salt makes food tasty”), and these questions were modelled from a previous survey conducted in adults [62] and were simplified for use with children (eg, the response scale was modified to a 5-point Likert smiley-face scale). Responses will be scored as “agree a lot”=3 points, “agree”=2 points, “I’m not sure”=0 points, “disagree”=0 points, and “disagree a lot”=0 points; hence, higher scores will reflect stronger beliefs about the importance of salt in food for taste. Behavior questions are related to discretionary salt use, previously used in children [52], and specific behaviors targeted within the intervention, such as talking to parents about salt use at home. Behavior question response items will be scored from 0 to 3, or 0 to 1 in the case of 2-item responses, with a higher score indicating better adherence to targeted salt-related behaviors. Self-efficacy questions were based on those used in previous studies in children [63] and examined key behaviors targeted within the intervention. The 3 response item will be scored as 0 to 2 (“not sure at all,” “a little sure I can,” and “very sure I can”) indicating low, medium, and high self-efficacy, respectively.

The study parent will complete an online questionnaire, containing 34 questions (baseline) and 21 questions (postintervention) assessing sociodemographic characteristics such as age, sex, and education level (13 questions) and knowledge, attitudes, and behaviors related to salt intake. The survey was based on a previously validated salt knowledge questionnaire [62,64], while other questions were modelled on those used in past surveys [65-67], and was tested for readability in 5 parents from varying demographic backgrounds. The pre- and postintervention surveys for parents were identical, with the exception of excluding demographic information on the follow-up survey.

We included 14 knowledge questions assessing declarative (ie, relationship between salt and sodium, current recommendation, main sources, health conditions related to excessive salt intake) and procedural knowledge (interpreting sodium information on food labels). These question responses are in the form of multiple choice and 5-point Likert scale (“strongly disagree” to “strongly agree”). All correct responses will be scored as 1, while incorrect responses, including “don’t know” and “not sure,” will be scored as 0. Responses to Likert scale questions will be scored as a 2 for “certainly true,” 1 for “probably true,” and 0 for incorrect answers, including “not sure” and “don’t know” responses. Negative statements will be reversed prior to scoring. The salt knowledge questions will be summed to generate scores for each question, and a total salt knowledge score will be derived by summing all the knowledge questions. The minimum and maximum salt knowledge scores are 0 and 32, respectively. We added some additional questions assessing knowledge specific to salt consumption in children, as the original validated questionnaire was not directed to parents.

We assessed 3 attitude questions (ie, parents’ personal attitude toward their own salt intakes, salt as a flavor enhancer, taste of low-salt foods, and importance of reducing salt) using 5-point Likert scales (“strongly disagree” to “strongly agree”). Scores will be assigned from 1 (“strongly disagree”) to 5 (“strongly agree”). The scores will be summed to derive a total beliefs score. The minimum and maximum attitude scores will be 0 and 6, respectively, with a score of 3 assigned for “strongly agree,” 2 for “agree,” 1 for “neither agree nor disagree,” 0 for “disagree” or “strongly disagree.” Higher scores will indicate stronger beliefs about the importance of the salt in food for taste. Some questions regarding attitudes related to salt intake will be presented as percentages.

We assessed 4 behavior questions regarding discretionary salt use (ie, adding salt to food during cooking [parent] or at the table [child and parent], and placing a salt shaker on the table [child and parent]) using 5-point Likert scales. Scores will be assigned from 0 for “always” to 4 for “never.” Parents will report the frequency of current actions taken to reduce their child’s salt intake (and actions taken in the past 1 month in the follow-up survey) using a 7-point Likert scale. The scores will be summed to derive a total behaviors score according to frequency of engaging in behavior (6 for “never,” 4 for “2-3 times/week,” etc). The minimum and maximum behavior score will be 0-53, respectively. Higher scores will indicate a higher frequency in engaging in positive salt related behaviors.

Tastings will be conducted with 2 commercially available food products, 1 snack food (potato chips) and 1 staple food (cornflakes), both of which are important sources of dietary salt for children [32]. We used the sodium content listed on packaged nutrient information panels to select 3 products of varying salt content for each food (Table 3). Nutrient profiles, particularly sugar and fat, were matched as closely as possible between samples of the same food type, as was visual appearance. Children will complete 2 taste tests for each food type during the testing session completed at schools (1) preference tests and (2) ability to rank samples according to salt content.

All samples of chips and cornflakes will be presented in plain containers in a predetermined randomized sequence. Children will be asked to have a drink of water between tasting each food sample. On completion of tasting the 3 samples, children will be asked which one they liked the most (ie, preference). The most liked sample will be removed from sight and the question repeated for the remaining 2 food samples. This forced-choice rank-order method means that each child will rank the 3 samples from most to least preferred. Following the same procedure, the child will then be asked which sample they think tastes the saltiest (ie, ability to rank samples). This rank-order method is based on previous methods [68] and has been successfully used in a similar sample of children ranking sour intensities and preference [69].