The prevalence of obesity in children and adolescents has increased in recent decades, and this rising trend has surged in parts of Asia [1]. In addition to short-term health consequences, overweight and obese children have a greater risk of adverse psychosocial health conditions and earlier onset of noncommunicable diseases [2]. Understanding factors contributing to obesity risk in children, such as physical inactivity and poor diet, is critical to prevent the accelerating burden of chronic diseases in later life.

However, there is a lack of good quality and timely data on schoolchildren, particularly those aged 10-14 years [3]. Parents may help to report the health behaviors of their children at home but often are not fully aware of their activities out of home, especially when children are increasingly independent in their food and activity choices at this age [4]. Although questionnaires completed by children are valuable tools to collect health behavioral data, they have the following limitations: (1) they often focus on a single health behavior; (2) they typically involve long-term recall, which is cognitively demanding for children (eg, examine activities during the past 7 days or last month) [5,6]; and (3) they rarely capture emerging health behaviors, such as nonscreen time sedentary behaviors, light-intensity physical activities, outdoor time, and tasks done concurrently (eg, screen time and eating) [7,8]. Although objective methods of measuring health behaviors are increasingly being used, they do not provide information on the type of activity or food consumed [9,10]. To address these limitations, a growing number of studies have used the time-use approach in children to collect information on health behaviors [11].

Time-use diaries prompt users to provide an account of a series of activities in a 24-hour day through previous day recalls [11]. Children are not cognitively developed to recall well before 8 years of age, but by the age of 10 years, they have the independent ability to provide accurate reports [12,13]. In addition, they allow the simultaneous capture of multiple energy balance–related behaviors. Most existing studies rely on paper-based diaries [11], whereas recent studies have developed software-based diaries [14-17], which have been demonstrated to engage children more effectively. However, to our knowledge, these studies focus mostly on capturing physical activities and sedentary behaviors but less comprehensively on dietary behaviors; for example, types of food consumed [14,17] and food portions [15,16] were not collected. Furthermore, there is a lack of application containing databases of Asian foods and common activities of children within the Asian context, such as tuition, also known as shadow education, which is prominent in East Asian countries [18]. In addition, the location of activities has been less studied, and no study to date has collected data on simultaneous activities (eg, eating while studying) in children aged 10-14 years, which is valuable for a better understanding of the activity contexts in children [19].

To bridge these gaps, we developed a self-administered, web-based time-use application to assess diet and movement behaviors in schoolchildren: My E-Diary for Activities and Lifestyle (MEDAL). This study aims to describe the development of MEDAL and evaluate its usability among schoolchildren in Singapore.

Several data sources were reviewed to identify the elements and requirements for developing a self-administered, web-based application for assessing diet and movement behaviors in schoolchildren. Both electronic and hand-searched strategies were adopted to identify existing software-based diet and physical activity assessment tools for schoolchildren. MyDailyMoves in the Netherlands [17], WebCAAFE in Brazil [16], Synchronised Nutrition and Activity Program in England [15], and Multimedia Activity Recall for Children and Adolescents in Australia [14] were evaluated. We also held interviews and meetings with experts from various knowledge domains, including custom software developers, experienced educators, and dieticians. The key points emanating were (1) application should be simple, easy to navigate, and take no more than 20 minutes per day to complete; (2) use of appealing visuals to engage children, and they prefer images to text and clicking instead of typing; and (3) enable children to report activities sequentially from wake-to-bed time (analogous to a time-use diary), so as to enhance recall—this method of behavior chaining has been shown to help children remember their activities [15]—and (4) complete the application at the end of each day to minimize omissions and intrusions [14,20]; children’s accuracy in reporting has been shown to improve with shorter retention intervals (ie, time between event occurrence and reporting) [21].

On the basis of the formative research in stage 1, we adopted an iterative stepwise refinement approach to design, develop, and test features of the application [22]. User testing was conducted with children using methods such as paper-based simulations, audio clips, and mock-ups (ie, static wireframes) to understand the children’s preferences for the interface and their comprehension of the task. At the end of user testing, user feedback was evaluated, and necessary improvements and amendments were conveyed to the software developers to inform the next development cycle. Multimedia Appendix 1 summarizes the key feedback and changes made during each development phase. Small groups of schoolchildren (aged 7-13 years), ranging from n=5 to n=15, were involved in all development stages to ensure that design decisions were tailored to their needs until the full functional system was launched. The sample sizes for each stage of user testing were adequate, as studies have shown that 80% of the usability problems were found with the first 5 users [23], and the diminishing return of uncovered problems was apparent when the sample size increased [24]. The key takeaways from the development phase were as follows: (1) recording should be limited to 4 days (similar to other diet and activity monitoring tools), as children found it tedious to record for 7 days; (2) prompts should be included to aid recall; (3) child-appealing visuals and intuitive designs should be used to ease navigation and motivate completion; and (4) younger children (<9 years) may need more assistance to complete MEDAL independently.

Of the 228 participants, 204 logged on to MEDAL and completed the usability questionnaire and were included in the analysis (Figure 2). The cohort comprised 57.8% (118/204) boys, with a median age of 10.9 (25th percentile: 10.6 and 75th percentile: 11.2) years, and 15.4% (31/201) were overweight (Table 1). The median time taken to complete MEDAL was 14.2 minutes per day, and the proportion who completed 3-4 days of recording, 1-2 days of recording, and <1 day of recording were 57.8% (118/204), 20.1% (41/204), and 22.1% (45/204), respectively. Compared with boys, girls spent more time completing MEDAL (median 15 min per day vs 13 min per day; P=.03) and were more likely to complete 3-4 days of recording (P=.01).

When assessing the usability of MEDAL, the majority (193/203, 95.1%) of participants agreed that instructions were clear, that MEDAL was easy to use (173/203, 85.2%), that they (172/202, 85.1%) liked the application, and that they preferred recording their activities on the web than on paper (167/202, 82.7%; Table 2). Recording for 4 days and ability to recall events were reported to be the least desirable among other factors (146/203, 71.9% and 154/203, 75.9%, respectively). Overall, the percentage agreement for the 10 questions ranged from 71.9% (146/203) to 95.1% (193/203).

Compared with boys, girls were more motivated by the gamification to complete recording on MEDAL (75/85, 88.2% vs 84/117, 71.8%; P=.005) and agreed that the time spent on completing 1-day entry was appropriate (71/85, 83.5% vs 78/112, 69.6%; P=.03; Table 2). Compared with boys, a higher proportion of girls reported being able to recall all the activities they did (84/118, 71.2% vs 70/85, 82.3%; P=.07) and food and drinks they consumed (90/117, 76.9% vs 75/86, 87.2%; P=.07).

A total of 23 participants provided open-ended suggestions. Participants (n=8) who recorded MEDAL for ≤2 days preferred recording on paper rather than on the web (Textbox 1). It was also reported that there was insufficient time to complete MEDAL. For participants (n=15) who completed 3 or 4 days of recording, there were suggestions to include more food and activity options and to have the ability to edit timings of activities (n=3 girls). Participants also proposed to allow recording on MEDAL for more than 4 days (n=3 boys) and modifying the game design, such as inclusion of more avatars and games and wanting a reward that is different from a plant.

In this study, we described the development and evaluated the usability of MEDAL—a 4-day web-based time-use application to assess diet and movement behavior in schoolchildren in Singapore. Approximately 60% (118/204) of the participants completed more than 3 days of recording, whereas the rest recorded for 2 days or less. The majority of the participants reported having a positive experience using MEDAL. Compared with boys, girls were reported to be more motivated by gamification, had better reported recall ability, and agreed that the time spent on completing 1-day entry was appropriate.

During the development process of MEDAL, we observed that the completion rate of MEDAL was significantly lower in younger children (aged 8-9 years) than in older children (aged 11-12 years). Younger children aged <9 years may need more assistance to complete MEDAL independently; thus, MEDAL is more suitable for children aged ≥10 years.

Response rates for multiday diaries have generally been an issue among children [46]. It is perceived as tedious to report for ≥2 days, especially when activity patterns were more or less similar on school days. Respondents may also report fewer activities on subsequent days as a means of decreasing the level of burden [47]. Indeed, in our study, we found similar issues with response rates, such that only 57.8% (118/204) of the participants completed 3-4 days of recording. Recording for 4 days was also the least satisfactory component reported in the usability assessment. To strike a balance between the reliability of data and minimizing respondent burden, the collection of 2 diary days for a large study (ie, 1 weekday and 1 weekend day) or 3 days for smaller-scale investigations (ie, 1 day from Monday to Thursday, Friday, and 1 weekend day) has been deemed adequate to obtain representative time-use data in children [14,46]. Large sample sizes may level out individual behavioral misclassifications that result from the short assessment period and would consequently give accurate population-level estimates [48]. For the next version of MEDAL, we plan to provide the flexibility to record between 1 and 7 days and optimize the application for mobile devices on both Android and iOS platforms to cater to the needs of different projects, and users can access the website on the go.

We observed that participants who recorded for <2 days were less likely to have access to a computer and the internet at home (P=.02, results not shown), and qualitative data showed that they preferred recording on paper rather than on the web. Participants (20/190,10.5%) with no access to the computer and internet at home may be less tech-savvy, thus exhibiting lower usability performance [16] and greater preference for print than digital. In the next version of MEDAL, we plan to integrate a frequently asked questions section to assist participants, especially those with lower computer knowledge and skills, to troubleshoot their problems and enable them to navigate the site with ease.

Existing studies have shown that recall ability is a common difficulty reported by children [14,16], although girls tend to outperform boys in episodic memory tasks (eg, everyday events) [49]. This is in line with our findings, where we observed trends toward greater reported recall ability in girls than in boys. Participants suggested including more food and drink and activity options and pictures (eg, household chores and shopping). This feedback will be incorporated into the next version of MEDAL, as images will be useful to serve as visual memory prompts to enhance recall in children.

Studies have shown that boys have lower attentional performance and poorer inhibitory control but perform faster than girls [50]. Consistent with these findings, we found that girls spent more time completing MEDAL (median 15 min per day vs 13 min per day), but they were more agreeable to the time spent on MEDAL compared with boys. Generally, the actual average time spent on MEDAL was comparable with similar questionnaires completed by children [15,16], which is sufficiently short to hold the users’ attention [20,51].

With regard to gamification, existing studies did not include this element [14-17]. In our study, girls appeared to be more motivated by the virtual gardening game to complete MEDAL recording than boys. Some boys have suggested integrating more games and that they wish to do more so that they can receive more rewards. Introducing more games may distract users and steer them away from recording their daily activities attentively [52]. However, as most of the participants reported a positive evaluation for the in-built game, we plan to revise and accelerate the game progression rewards in the next version of MEDAL so that users are able to receive rewards more frequently. Users will be rewarded with gardening tools as they record their activities, and a mystery fruit will be revealed after each completed day of recording. On the basis of the feedback from the participants, we will also include more avatars and backgrounds for users to choose and personalize in the next version of MEDAL.

MEDAL is a novel and engaging application for schoolchildren aged ≥10 years. It includes a range of Asian food options that cater to the key ethnic groups in Asia and captures data on food portions, which allow us to have a detailed assessment of energy and nutrient intakes. This is useful, as evidence suggests that portion sizes among children are increasingly poorly regulated [53]. In addition to assessing dietary intake and movement behaviors, MEDAL also examines the location of activities and tasks performed concurrently to allow a better understanding of the context of the activities and enable interventions to be more targeted. The usability evaluation of MEDAL, including user testing during the development phase, involves several methodological strategies (eg, direct observation, interview, Likert scale questions, and open-ended feedback), which allowed us to obtain rigorous and detailed qualitative and quantitative data.

However, there are some limitations to this study. First, we did not have information on children’s academic ability. However, our usability questionnaire showed that 95.1% (193/203) of the participants understood the instructions on MEDAL, which was not unexpected, considering that English is the main language of instruction in Singapore and primary education is compulsory. However, we recognize that the cognitive abilities to recall events, capability to concentrate, and comprehension of questions may have an impact on usability testing [16] and acknowledge that children with poorer cognitive ability may experience some challenges in using MEDAL. Currently, users are only able to edit types of activities but not the timing of the activities. In addition, the current application prevents users from recording beyond 12 midnight. As late bedtimes have become pervasive in children [54], the ability to record activities beyond 12 midnight, along with the function to modify timing, will be made available in the next version of MEDAL. In addition, to cater to both observational and interventional study designs, we plan to incorporate the flexibility to hide the lifestyle report so that it will not induce behavioral change in participants of observational studies.

In conclusion, we developed MEDAL, which appears to be a feasible web-based time-use application for children aged 10-12 years and potentially useful for children above this age group. We observed some gender differences in usability performance, but the majority of the participants had a positive experience using MEDAL. The validity and reliability of MEDAL will be examined next, and the results of this study will be used to inform the next version of MEDAL to customize MEDAL for children in other parts of Asia, which includes translation to their native language and adaptation of application content to their local context.