Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Monday, March 11, 2019 at 4:00 PM to 4:30 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?

Advertisement

Citing this Article

Right click to copy or hit: ctrl+c (cmd+c on mac)

Published on 11.02.20 in Vol 8, No 2 (2020): February

Preprints (earlier versions) of this paper are available at http://preprints.jmir.org/preprint/16499, first published Oct 04, 2019.

This paper is in the following e-collection/theme issue:

    Original Paper

    Improving Calcium Knowledge and Intake in Young Adults Via Social Media and Text Messages: Randomized Controlled Trial

    The University of Sydney, School of Life and Environmental Sciences, Camperdown, Australia

    Corresponding Author:

    Anika Rouf, MND

    The University of Sydney

    School of Life and Environmental Sciences

    Level 4 East, Charles Perkins Centre

    Camperdown, 2006

    Australia

    Phone: 61 86274704

    Fax:61 86271605

    Email: arou9270@uni.sydney.edu.au


    ABSTRACT

    Background: Calcium is an important nutrient for the attainment of peak bone mass during adolescence and young adulthood. However, these life phases are characterized as hard to reach for health promotion. Social media platforms offer a promising channel as they are relatively low cost but used ubiquitously by youth.

    Objective: The aim of the CAlcium Nutrition-Dietary Opportunities (CAN-DO) study was to conduct a randomized controlled trial to test the effectiveness of Facebook alone or with text messaging as channels to deliver a theory-based program to encourage optimal calcium intake.

    Methods: The intervention was a 3-arm parallel trial. Young adults aged 18 to 25 years were recruited through university and social media for a 6-week trial. Participants were randomized to 1 of the 3 arms (ie, Facebook posts, Facebook posts plus text messages, and control group that received an electronic leaflet containing information on calcium intake). The primary outcome was change in intake of milk and other calcium-rich foods, and secondary outcomes were knowledge, self-efficacy, motivation, and habit formation concerning calcium-rich foods. Changes were assessed before and after the intervention, and the differences in change between groups were compared using multivariate regression models with multiple imputations for missing data.

    Results: A total of 211 participants (64/211, 30.3% males) participated (mean age 21.4 years, SD 2.1) in this study. At the end of the program, no increase in milk intake (odds ratio [OR] 1.51, 95% CI 0.61-3.75 Facebook; OR 1.77, 95% CI 0.74-4.24 Facebook plus text messages; P=.41) nor calcium-rich food was detected (P=.57). There was a significant improvement in knowledge in the Facebook plus text messages group (P<.001), but habit formation improved less than that in the other 2 groups (P=.01). Our results showed a moderate level of engagement with intervention content and positive qualitative feedback from participants.

    Conclusions: The CAN-DO study delivered via Facebook (with the additional support of text messages) was found to improve knowledge and was acceptable among young adults. However, further research is needed to better understand social media engagement and how to optimize the program for participants to be sufficiently motivated to increase their intake of calcium-rich foods.

    Trial Registration: Australian New Zealand Clinical Trials Registry ACTRN12620000097943; http://www.anzctr.org.au/ACTRN12620000097943.aspx

    JMIR Mhealth Uhealth 2020;8(2):e16499

    doi:10.2196/16499

    KEYWORDS



    Introduction

    Background

    As adolescents and young adults become increasingly independent, it is not uncommon for lifestyle behaviors to be adversely affected [1]. This may include decreased physical activity, increased rates of smoking and alcohol consumption, weight gain, and decreases in home-prepared meals [2-5]. Previous studies have shown that young adults are difficult to reach with traditional health promotion strategies [6,7], but it is important to support young adults through this transition to establish healthy dietary patterns for their own future health [3,8,9] and to potentially serve as role models to their children [10,11].

    Among the consequences of poor-quality diets is a low intake of calcium, which remains a global concern [12] and among Australian young adults [13]. A secondary analysis from the most recent Australian National Nutrition Survey from 2011 to 2012 shows that 69% of males and 83% of females aged 19 to 25 years failed to meet the estimated average requirements for young adults [14]. An adequate intake of calcium in adolescence and young adulthood is important for the attainment of peak bone mass and prevention of osteoporosis in later life [15,16].

    Our previous formative research has delved into the barriers and enablers to achieving adequate calcium intake for this population and revealed a gap in knowledge with respect to what amount of calcium-rich food constitutes a serve and the daily number of serves recommended [17]. Their level of motivation to improve calcium intake was low because a lack of knowledge meant more calcium seemed unnecessary, and financial factors influenced the opportunity to consume calcium-rich foods, wherein milk was seen as low cost, but sources such as nuts and fish were seen as high cost [17]. When asked about an appropriate medium to deliver an intervention program, the focus group participants preferred to learn from social media platforms, and Facebook was ranked as the preferred platform [17].

    Young adults are ubiquitous users of social media [18]. Almost 90% of young adults (aged 18-29 years) access social media platforms at least once per day [19], so it has the potential for wide reach in an intervention. To date, the small evidence base for the effectiveness of nutrition-related interventions using a commercial social media platform, such as Facebook, is inconsistent and warrants further investigation [20-22]. Our previous meta-analysis of the effectiveness of interventions to increase calcium intake demonstrated a small effect size [23] but indicates that research into an intervention to improve calcium intake of Australian young adults is warranted.

    Objective

    A previous Facebook intervention for weight loss in young adults found that the use of social media combined with text messages was effective for weight loss but not Facebook alone [20]. Previous electronic health interventions conducted in young adults found a high level of acceptability and engagement with text messages and effective dietary changes [24-26]. Therefore, the aim of this study was to determine the effectiveness of an educational and motivational program to improve calcium intake in young adults and whether the addition of text messages enhanced behavior change when compared with the Facebook arm alone.


    Methods

    Trial Design

    This was a 3-arm parallel trial with a 1:1:1 allocation ratio. The 3 groups were Facebook intervention (Facebook), Facebook intervention plus text messages (Facebook plus text), and electronic leaflet (e-leaflet) containing information on calcium intake (control). The sample size was determined using G*Power (Version 3.1.9.4, Universität Kiel), a statistical power analysis software [27]. To detect a mean difference of 125 mg calcium intake with P=.05 and 90% power, assuming a standard deviation of 259 mg, a sample size of 45 was required per arm and increased to 75 to allow for 40% dropout.

    Participants

    Young adults (males and females) aged 18 to 25 years were selected as this is the period where peak bone mass development is reached [28,29]. Inclusion criteria included owning a smartphone and a Facebook account. Exclusion criterion was having completed a nutrition course or currently undertaking a nutrition course on the basis of their high existing level of nutrition knowledge. In addition, any participants with a food allergy, known lactose intolerance, or currently taking calcium supplements (but not multivitamins) or eating disorders were excluded.

    All materials and methods of the intervention were approved by the Human Research Ethics Committee at the University of Sydney, Australia. The ethics approval number is 2018/597. Each participant was reimbursed with an Aus $10 voucher after completing the final questionnaire. This offer did not impact the voluntary nature of consent as it was provided after the intervention finished rather than at the time of consent. The reporting of outcomes was guided by the Consolidated Standards of Reporting Trials of Electronic and Mobile HEalth Applications and onLine TeleHealth checklist [30]. As neither the primary outcome nor the secondary outcomes were clinical measurements, the study was not entered into a clinical trials registry.

    Randomization and Concealment

    A randomized sequence generation was used to allocate the participants. The randomization was performed by 2 independent researchers who were not study investigators.

    Recruitment

    Recruitment strategies included social media (posts to friends and paid advertising on Facebook), posting on University website (volunteer for research study), flyers (on campus noticeboards), volunteers on a research database (previous volunteers who took part in nutrition research and agreed for contact in the future), and active face-to-face recruitment. For each of the abovementioned recruitment methods, the potential participant was made aware that participation was voluntary. Interested participants accessed the screening questionnaire for eligibility before joining the study.

    CAlcium Nutrition-Dietary Opportunities Program

    A theory-informed step-wise approach was used to develop the CAlcium Nutrition-Dietary Opportunities (CAN-DO) program using the Behavior Change Wheel system [31]. This framework posits that an individual requires capability (C), opportunity (O), and motivation (M) to perform a certain behavior (B) and includes a series of 9 intervention functions that can be mapped to the COM-B components [31,32].

    The aim was to build relevant knowledge (capability) and influence beliefs and attitudes to generate intentions for individuals to change behaviors (reflective motivation). The details of the intervention functions and relevant behavior change techniques are presented in Table 1. In brief, the behavior change techniques included goal setting (behavior), self-monitoring of behavior, social support (unspecified), instruction on how to perform a behavior, information about health consequences, behavior substitution, habit formation, credible source, and restructuring the physical environment.

    The content of the intervention was developed in 2 parts. A range of instructional videos was created to build skills in cooking calcium-rich, low-cost, and mostly plant-based meals. These were tested in focus groups for acceptability and refined based on the feedback (unpublished findings). The next step was to design text messages and Facebook posts tailored to the preferences of young adults as indicated in prior formative research [33]. The intervention content was focused on educating on calcium-containing food sources and recommended serves, tips for including more calcium, and recipe videos that provided instructions on how to incorporate calcium in main meals and snacks. Text messages were kept short (<160 characters) and designed to complement the Facebook posts. Text messages and Facebook posts also reminded participants about setting goals and tracking progress for habit formation and created social support via posts and 2-way text messaging. An infographic was created to inform participants of the recommendations and set as a cover photo on the Facebook page (Figure 1). The e-leaflet that was provided to participants in the control group is shown in Multimedia Appendix 1.

    Table 1. Details of behavior change techniques with an example.
    View this table
    Figure 1. Infographic to inform participants of the recommendations.
    View this figure

    Procedures

    Interested participants completed a screening questionnaire hosted on Research Electronic Data Capture (REDCap) [34], where they could find out more about the study by reading the participant information statement and check their eligibility. Participants who were not eligible to participate were provided with the Australian Dietary Guidelines as a resource. Eligible participants completed the consent form and proceeded to the baseline questionnaire. After completing the baseline questionnaire, each participant was randomized to Facebook, Facebook plus text, or control group and received an email with their group allocation. Participants in the Facebook and Facebook plus text groups were invited to join a closed Facebook group, where a post was made every alternate day by the researcher (AR). The site had a pinned Facebook post used to ensure that all participants were provided with background information, which included links to educational resources and an overview of the intervention. Screenshots of the posts are shown in Figure 2. The 2 Facebook groups were kept separate to avoid potential contamination between groups. In addition, participants in Facebook plus text group were sent text messages every alternate day to the post. Participants in both intervention groups were encouraged to set goals using apps available on iPhone (Productive—Habit Tracker) and Android platforms (Loop—Habit Tracker) and self-monitor their progress. The participants in the control group were emailed once with an e-leaflet on calcium and did not receive any continued support on social media. This minimal intervention was to maintain their interest in completing the study.

    Figure 2. Screenshots displaying Facebook posts that included photos, videos and the tracking apps for goal-setting and self-monitoring.
    View this figure

    Measurement of Outcomes

    Demographic information was collected from all participants, which included age, gender, educational level, postcode (for categorizing socioeconomic status), occupation, and income through a Web-based platform REDCap [34]. The postcode was used to categorize the socioeconomic status of participants using Socio-Economic Indexes For Areas (SEIFA) [35]. All outcomes were assessed at 2 time points, which were at baseline before commencing the study (T0) and at the end of intervention (T1) via a Web-based questionnaire on REDCap.

    The primary outcome (calcium intake) was estimated using a validated calcium-specific food frequency questionnaire that asks about intake over the past week [36]. Milk was measured in cups, ranging from half a cup to more than 4 cups, and the other calcium-rich foods (30 foods and beverages included) were measured by weekly frequency of intake only. The secondary outcomes measured the impact of the intervention on determinants of calcium intake, which included knowledge of calcium recommendations and serve sizes, self-efficacy, motivation for consuming calcium-rich foods, and habit formation. Knowledge was assessed by a participant’s ability to identify sources of calcium (maximum of 8), a correct serve of calcium (maximum of 8), and stating the calcium requirements for their age group (maximum of 2) using a researcher-designed questionnaire as no validated questionnaire could be found. The questions are included in Multimedia Appendix 2. A 5-item Likert scale questionnaire previously validated for other dietary behaviors was adapted to measure self-efficacy for improving calcium-rich food consumption. The maximum score possible was 25; a higher score indicated stronger self-efficacy [37]. Autonomous and controlled motivation for consumption of calcium was measured using a 4-point scale. The questions were adapted from the Self-Regulation Questionnaire [38,39], where a higher score indicated greater motivation (score out of 16). Habit formation for calcium intake was measured using the validated 4-item 7-point scale Self-Report Behavioral Automaticity Index [40]. A greater score indicated a higher automaticity to perform a certain behavior.

    Engagement and Process Evaluation

    Engagement with the platform was measured quantitatively and qualitatively as research indicates the need to do both [41]. Quantitative measures were obtained from recording Facebook analytics. After all participants had completed the intervention, the number of participants who had seen, liked, and commented on the Facebook posts was recorded. For the Facebook plus text group, the number of replies to text messages was counted for each participant, and the content was analyzed using qualitative methods (see Qualitative Analyses below).

    Feedback regarding the acceptability of the program was collected via open-response questions regarding ease of use, usefulness of program, likelihood of recommendation to others, and overall enjoyment using Likert scales (5 being highest). The other optional questions were related to intervention experience and uptake of content as well as frequency and reason for engagement. The last question provided participants with an opportunity for free text comments.

    Statistical Analysis

    To account for all participants, an intention-to-treat analysis with multiple imputations for missing values was used. This meant that all participants who were randomized at the start of the trial were retained for analyses. Owing to the large amount of missing data, 10 imputed datasets were created based on gender, SEIFA (socioeconomic index), cooking frequency per week, and baseline intake of primary (milk and calcium intake) and secondary (knowledge, self-efficacy, motivation, and habit) outcomes using Stata version 13.1 (StataCorp LP).

    The primary outcome of change in milk intake, which was categorical in number of cups, was compared between 3 groups using a logistic regression model adjusted for gender, SEIFA, cooking frequency, baseline calcium (nonmilk), baseline knowledge, baseline self-efficacy, baseline motivation, and baseline habit. The quantitative values for change in calcium intake from other dietary sources were compared using linear regression as were the variables for the secondary outcomes of knowledge score, self-efficacy for change score, and motivation and habit score, adjusted for gender, SEIFA, cooking frequency, baseline calcium intake, baseline knowledge, self-efficacy, motivation, and habit. An analysis using completers-only data was conducted and is available in the Multimedia Appendices 3-5. The distribution of missing outcome data at both time points was investigated using counts and percentages across all sociodemographic variables. Furthermore, separate general estimating equation (GEE) models for binary data were used to investigate any relationships between sociodemographic variables and missingness in each outcome, adjusted for other sociodemographic variables. An independent samples t test was used to assess differences in number of views, likes, and comments for Facebook posts between the 2 groups receiving the intervention (SPSS for Windows 22.0 software IBM Corp, released 2013). A P value of less than .05 was considered statistically significant for all tests.

    Qualitative Analyses

    The feedback from the final questionnaire was transcribed and analyzed using an inductive approach where common themes were grouped together. The NVivo 12 Plus (2018, version 12.2.0; QSR International Pty Ltd) software was used for thematic analyses.


    Results

    Participant Characteristics

    A total of 270 participants attempted the screener questionnaire. Of 270 participants, 59 were ineligible for the study or failed to continue to the baseline questionnaire. A total of 211 young adults were randomized into 3 groups. The flow of participants through the trial is shown in Figure 3. The characteristics and demographics of participants at baseline are presented in Table 2. The mean age was 21.4 years (SD 2.1), and the sample comprised 30.3% (64/211) males.

    The majority of participants (139/211, 65.9%) were enrolled in tertiary education. Nearly one-third (65/211, 30.8%) of the participants were in health care for their field of work or study. Almost two-thirds (134/211, 63.5%) of the participants were earning less than Aus $500 per week. Nearly half (94/211, 44.5%) of the sample reported themselves as being the main purchaser of household groceries. The most commonly reported cooking frequency was less than twice weekly for 37.4% (79/211) of the young adults.

    Figure 3. Participant flow diagram in the CAlcium Nutrition-Dietary Opportunities study.
    View this figure
    Table 2. Demographics of participants from the CAlcium Nutrition-Dietary Opportunities study.
    View this table

    Attrition

    Overall, 9 participants formally withdrew from the study. All participants were from the same arm (Facebook plus text) and opted out by sending a text—an option not available to other participants who could only opt out passively. The dropout time ranged from day 1 to day 29. Only 2 participants provided reasons (ie, lack of interest or time). In total, 148 (148/211, 70.1%) participants completed the final questionnaire but not necessarily every question.

    Outcomes

    Results from 209 participants (data from 2 participants could not be imputed because of incorrect postcodes) are reported in the following sections. Results using completers-only data are included in Multimedia Appendices 3-5. The percentage of data that were missing was approximately 35% for milk intake, knowledge, self-efficacy, motivation, and habit, but 75% for calcium-rich foods. This percentage was similar across all sociodemographic variables. The GEE indicated that females had lower adjusted odds ratio (OR) than males of have missing data, adjusted for all other sociodemographic variables. No other sociodemographic variables were associated with missing outcome values.

    Primary Outcomes

    Participants in the Facebook group were 1.51 times more likely to move to a higher milk category compared with those in the control group (Table 3). Similarly, those in the Facebook plus text group were 1.77 times more likely to move to a higher milk intake category. However, this was not significant (P=.41). There was no difference in the change in calcium intake from other foods between groups over the 6 weeks (P=.57; Table 4). The analysis on completers-only data demonstrated a significant increase in milk intake in the Facebook plus text messages group compared with the control group (OR 4.99, 95% CI 1.63-15.28).

    Table 3. Change in category of the amount of milk intake from baseline to the end of the intervention for all participants (n=209, using imputed dataset); overall P=.41.
    View this table
    Table 4. Change in calcium intake per day in mg (excluding milk) from baseline to the end of the intervention using logistic regression for all participants (n=209, using imputed dataset); P=.57.
    View this table
    Secondary Outcomes

    Changes in secondary outcomes are reported in Table 5. The answers to the knowledge questions were combined together as an overall knowledge score. The change in knowledge was significant between the groups (P<.001). Those in the Facebook plus text intervention arm had a greater improvement in mean score compared with those in the Facebook and control groups. No significant difference between groups was observed for motivation (P=.79) or self-efficacy (P=.31). For habit formation, a significant group effect was observed (P=.01), with Facebook plus text group having the least increase in score. The improvement in knowledge in the Facebook plus text messages group was also found with completers-only analysis (P=.04). The effect on habit formation was not shown in the completers-only analysis.

    Table 5. Change in secondary outcomes from baseline to the end of the intervention for all participants (n=209, using imputed dataset).
    View this table

    Engagement

    Facebook Posts and Text Messages

    Table 6 shows the engagement with Facebook posts. More participants in the Facebook plus text intervention than those in the Facebook intervention viewed the posts and liked them (P<.001 for both). In the Facebook group, 3 participants made comments on posts, whereas 4 participants in the Facebook plus text group commented on posts.

    For the Facebook plus text group, the mean number of replies from participants was 3.8 out of a maximum 21 (range 1-18). Of 75 participants, 12 made no reply texts (1 participant gave a wrong phone number, and texts could not be delivered). The highest number of replies was to the yes/no response as to whether they had set a goal on the app (n=22).

    Table 6. Engagement with the program on Facebook.
    View this table

    Process Evaluation

    The majority of participants (n=133) completed the process evaluation questions, and Table 7 shows that there were no differences between intervention groups as to ease of use, their liking, likelihood of recommending it to others, or usefulness of the program. Participant responses in relation to message reading and interactions are included in Multimedia Appendix 6.

    The thematic analysis with representative quotes is tabulated in Multimedia Appendix 7. The themes were grouped into ease of use, raised awareness, increased intake, feedback on recipes, reasons for reading/posting, and suggestions for improvement. Any comments that did not fit into these 5 groups were labeled as general feedback. There was a divergence of opinion on the ease of use, with some participants suggesting it was easy to follow, and others had more difficulty understanding and wanted more feedback. Successful participants shared their accomplishments in achieving their goals. The feedback on the recipes was overall positive, but some participants admitted they never prepared any of them. The majority of respondents chose not to share posts with reasons being they were uncertain they could add anything extra to the conversation or they did not feel comfortable with sharing. Some of the suggestions for improvement under general feedback included using an alternate platform that allows for active chat between members, sending text messages more frequently to check up on their progress, completing surveys weekly to track progress, and organizing meetings in person. Most participants viewed the notifications as a gentle and helpful reminder, whereas some found it intrusive.

    Table 7. Process evaluation of the CAlcium Nutrition-Dietary Opportunities study on intervention experience.
    View this table

    Discussion

    Principal Findings

    This study showed that a 6-week intervention to increase calcium intakes tailored to young adults delivered using a social media platform and text messages was successful in improving knowledge about calcium-rich foods. However, this did not result in a significant increase in calcium-rich food and beverage intakes. The Facebook intervention delivered alone failed to show knowledge improvement, but engagement with the social media was significantly less than that in the intervention arm receiving text messages and might explain the disparate finding. Other reasons for the difference might be that the additions of texts appear to provide a more personalized program, and the need to reply to some messages engenders accountability and perception of monitoring by the staff delivering the intervention.

    The findings of a successful outcome from the combined intervention arm concur with the earlier findings of a weight loss program delivered to overweight and obese college-aged students. Over 8 weeks, topics (1 per week) about weight loss were posted on Facebook, and the other intervention arm additionally received text messages with personalized feedback each week [20]. Although our text messages were generic, the participants’ names were included, and they were written in the Generation Y tone for which young adults had previously expressed a preference [33]. The texts provided additional prompts to set goals and self-monitor their own behavior with some further education and persuasion. These 2 behavior change techniques have been demonstrated to result in behavior change [42].

    Few dietary changes occur as a result of education alone, but it was indicated as a necessary antecedent to behavior change in this demographic based on our previous focus group findings. Although the Facebook plus text group improved knowledge, the overall score remained quite low, with the mean score only reaching 50% correct answers. Another barrier to improving calcium intakes seemed to stem from lack of motivation, with all groups scoring similarly at baseline (10 out of 16), with uniform small improvement at the end of the intervention. In future programs, more planning around the inclusion of other techniques to improve reflective motivation may be needed. Coercion, persuasion, and incentivization could be possible solutions [31]. Social media platforms readily offer the capacity for monitoring of an individual’s behavior by others, and social comparison could be applied to intake of calcium-rich foods in this case. The vacillation might be that members are uncomfortable with sharing information with others as seen here in the replies to the process evaluation. The lack of posts made by group members is also indicative that such an approach may not work to positively influence motivation. Further research to understand what would allow participants to be relaxed with sharing dietary information in a nutrition intervention is desirable. With regard to incentivization and rewards, an earlier qualitative study with young adults for the co-design of an intervention to improve vegetable intakes reported that self-rewards were unlikely to motivate them as it required too much self-organization, so social or material rewards may be a better choice [21].

    The validity of the food frequency instrument to measure changes in the primary outcome of calcium from milk and other foods in this population must also be questioned. Any self-report tool is always subject to participant bias [43]. In addition, this tool may not possess sufficient precision to detect small changes in intakes, as milk intakes are categorized in cups from half a serve of dairy to 4 or more serves of dairy. The calcium-specific food frequency questionnaire was selected rather than other tools as the burden of completion was low, but it does serve to rank individuals rather than assess absolute intake, and hence, the OR of increasing category of intake was used here.

    Improvements in calcium intake were not achieved, but the retention and engagement in the social media intervention were substantial for an electronically delivered intervention [44]. Overall, 70% of the sample was retained, and more than half of the participants viewed the posts. Previous studies report large attrition and declining engagement in social media interventions for improving health behaviors [45]. A strength of the CAN-DO study was the formative research conducted to inform program design and materials [33]. The components were generally well received, and the recipe videos commended.

    Among the limitations of this study was the overrepresentation of females comprising 70%, but this is not uncommon in nutrition studies even when males are equally targeted. In addition, in the case of calcium, it is females who are more likely to have inadequate intakes, so the population participating was appropriate. Some participants who did not do their own grocery shopping and cooked infrequently may have lessened the opportunity to alter their meals and snacks. The length of the program may have been too short to see the changes in knowledge translate into changes in consumption of calcium-containing foods, and intakes were only measured at 2 time points. An intervention delivered to university students that included a face-to-face session followed by text messages for 10 weeks did show increases in calcium intake [46]. In the future, a longer intervention might be appropriate. Finally, to include the largest number of participants, multiple imputation was used. This increases the variance in the estimate and a more conservative interpretation of results than completers-only analysis.

    Conclusions

    The CAN-DO study was found to be feasible to deliver in our selected target population. Our qualitative results from the process evaluation mostly indicate that participants enjoyed the program. However, the quantitative result shows that 1 (change in knowledge) out of 5 outcomes improved in the Facebook plus text messages group only. The CAN-DO study has provided valuable insights into the process of disseminating a social media intervention for young adults, and a number of changes to program design are indicated to improve motivation. The lack of interaction between the members of the groups requires research to discover how to facilitate members to post and provide social support. This is important as it appears that the social interaction between the interventionist and participants via the text messages results in better outcomes.

    Acknowledgments

    AR, MN, and MAF contributed to the conception and design of the study. AR conducted the research, analyzed the data, and drafted the manuscript. MAF contributed to interpretation of the results and editing of the manuscript. All authors have read and approved the final version of the manuscript. AR was supported by the Australian Postgraduate Award Scholarship from the Commonwealth Government of Australia. The authors would like to thank Nematullah Hayba and Alyse Davis for randomizing the participants; Dr Leah Shepherd for her guidance on the statistical analysis plan and assistance with code writing on Stata; Dr Stephanie Partridge for providing guidance on multiple imputation, and the Sydney Informatics Hub, a Core Research Facility of the University of Sydney, for their assistance with statistical analysis (Jim Matthews and Christopher Howden) and survey design on REDCap (Olya Ryjenko). This research did not receive funding from any agencies.

    Conflicts of Interest

    MAF receives funding from the National Health and Medical Research Council, New South Wales (NSW) Health, Australian Research Council, and Cancer Council NSW.

    This randomized study was only retrospectively registered. The editor granted an exception from ICMJE rules mandating prospective registration of randomized trials because the risk of bias appears low. However, readers are advised to carefully assess the validity of any potential explicit or implicit claims related to primary outcomes or effectiveness, as retrospective registration does not prevent authors from changing their outcome measures retrospectively.

    Multimedia Appendix 1

    E-leaflet provided to participants in Group C.

    PNG File , 2807 KB

    Multimedia Appendix 2

    Questionnaire used to measure change in knowledge at baseline and end of intervention.

    DOCX File , 636 KB

    Multimedia Appendix 3

    Change in the amount of milk intake from baseline to end of intervention (completers only).

    DOCX File , 15 KB

    Multimedia Appendix 4

    Change in calcium intake per day in mg (excluding milk) from baseline to end of intervention (completers only).

    DOCX File , 14 KB

    Multimedia Appendix 5

    Change in secondary outcomes from baseline to end of intervention (completers only).

    DOCX File , 15 KB

    Multimedia Appendix 6

    Process evaluation of the CAN-DO study on frequency of reading posts, messages and interaction.

    DOCX File , 14 KB

    Multimedia Appendix 7

    Quotations illustrating feedback from participants provided through text message replies and qualitative process evaluation (n=106).

    DOCX File , 18 KB

    Multimedia Appendix 8

    CONSORT‐EHEALTH checklist (V 1.6.1).

    PDF File (Adobe PDF File), 2890 KB

    References

    1. Lenz B. The transition from adolescence to young adulthood: a theoretical perspective. J Sch Nurs 2001 Dec;17(6):300-306. [CrossRef] [Medline]
    2. Butler SM, Black DR, Blue CL, Gretebeck RJ. Change in diet, physical activity, and body weight in female college freshman. Am J Health Behav 2004;28(1):24-32. [CrossRef] [Medline]
    3. Nelson MC, Story M, Larson NI, Neumark-Sztainer D, Lytle LA. Emerging adulthood and college-aged youth: an overlooked age for weight-related behavior change. Obesity (Silver Spring) 2008 Oct;16(10):2205-2211 [FREE Full text] [CrossRef] [Medline]
    4. Vadeboncoeur C, Townsend N, Foster C. A meta-analysis of weight gain in first year university students: is freshman 15 a myth? BMC Obes 2015;2:22 [FREE Full text] [CrossRef] [Medline]
    5. Australian Bureau of Statistics. 2014. Soft Drink, Burgers and Chips - the Diet of Our Youth   URL: https://tinyurl.com/yxywg7tx [accessed 2019-09-25]
    6. Poobalan AS, Aucott LS, Clarke A, Smith WC. Diet behaviour among young people in transition to adulthood (18-25 year olds): a mixed method study. Health Psychol Behav Med 2014 Jan 1;2(1):909-928 [FREE Full text] [CrossRef] [Medline]
    7. Howarth CC, Street C. New Policy Institute. 2000. Young Adults’ Access to Services   URL: https://www.npi.org.uk/files/9113/7545/0719/Sidelined.pdf [accessed 2019-11-27]
    8. Arnett JJ. Emerging adulthood. A theory of development from the late teens through the twenties. Am Psychol 2000 May;55(5):469-480. [CrossRef] [Medline]
    9. Dunn JE, Liu K, Greenland P, Hilner JE, Jacobs DR. Seven-year tracking of dietary factors in young adults: the CARDIA study. Am J Prev Med 2000 Jan;18(1):38-45. [CrossRef] [Medline]
    10. Loth KA, MacLehose RF, Larson N, Berge JM, Neumark-Sztainer D. Food availability, modeling and restriction: How are these different aspects of the family eating environment related to adolescent dietary intake? Appetite 2016 Jan 1;96:80-86 [FREE Full text] [CrossRef] [Medline]
    11. Larsen JK, Hermans RC, Sleddens EF, Engels RC, Fisher JO, Kremers SP. How parental dietary behavior and food parenting practices affect children's dietary behavior. Interacting sources of influence? Appetite 2015 Jun;89:246-257. [CrossRef] [Medline]
    12. Balk EM, Adam GP, Langberg VN, Earley A, Clark P, Ebeling PR, International Osteoporosis Foundation Calcium Steering Committee. Global dietary calcium intake among adults: a systematic review. Osteoporos Int 2017 Dec;28(12):3315-3324 [FREE Full text] [CrossRef] [Medline]
    13. Australian Bureau of Statistics. 2015. Australian Health Survey: Usual Nutrient Intakes, 2011-12   URL: http:/​/www.​abs.gov.au/​ausstats/​abs@.nsf/​Lookup/​by%20Subject/​4364.​0.​55.​008~2011-12~Main%20Features~Calcium~401 [accessed 2019-03-21]
    14. Rouf AS, Sui Z, Rangan A, Grech A, Allman-Farinelli M. Low calcium intakes among Australian adolescents and young adults are associated with higher consumption of discretionary foods and beverages. Nutrition 2018 Nov;55-56:146-153. [CrossRef] [Medline]
    15. Matkovic V. Calcium and peak bone mass. J Intern Med 1992 Feb;231(2):151-160. [CrossRef] [Medline]
    16. Cashman KD. Diet, nutrition, and bone health. J Nutr 2007 Nov;137(11 Suppl):2507S-2512S. [CrossRef] [Medline]
    17. Rouf A, Clayton S, Allman-Farinelli M. The barriers and enablers to achieving adequate calcium intake in young adults: a qualitative study using focus groups. J Hum Nutr Diet 2019 Aug;32(4):443-454. [CrossRef] [Medline]
    18. Greenwood S, Perrin A, Duggan M. Pew Research Center. 2016 Nov 11. Social Media Update 2016   URL: https://www.pewresearch.org/internet/2016/11/11/social-media-update-2016/ [accessed 2019-04-30]
    19. Sensis. multi-screen site. 2017 Jun 22. Sensis Social Media Report 2017: Chapter 1 – Australians and Social Media   URL: https://irp-cdn.multiscreensite.com/535ef142/files/uploaded/Sensis-Social-Media-Report-2017.pdf [accessed 2018-08-22]
    20. Napolitano MA, Hayes S, Bennett GG, Ives AK, Foster GD. Using Facebook and text messaging to deliver a weight loss program to college students. Obesity (Silver Spring) 2013 Jan;21(1):25-31 [FREE Full text] [CrossRef] [Medline]
    21. Nour M, Chen J, Allman-Farinelli M. Young adults' engagement with a self-monitoring app for vegetable intake and the impact of social media and gamification: feasibility study. JMIR Form Res 2019 May 10;3(2):e13324 [FREE Full text] [CrossRef] [Medline]
    22. NSW Food Authority. Evaluation of Kilojoule Menu Labelling   URL: http://www.foodauthority.nsw.gov.au/_Documents/scienceandtechnical/fastchoices_evaluation_report.pdf [accessed 2019-10-02]
    23. Rouf AS, Grech A, Allman-Farinelli M. Assessing the efficacy and external validity of interventions promoting calcium or dairy intake in young adults: A systematic review with meta-analysis. Crit Rev Food Sci Nutr 2018;58(15):2600-2616. [CrossRef] [Medline]
    24. Partridge SR, McGeechan K, Hebden L, Balestracci K, Wong AT, Denney-Wilson E, et al. Effectiveness of a mHealth Lifestyle Program With Telephone Support (TXT2BFiT) to Prevent Unhealthy Weight Gain in Young Adults: Randomized Controlled Trial. JMIR Mhealth Uhealth 2015 Jun 15;3(2):e66 [FREE Full text] [CrossRef] [Medline]
    25. Allman-Farinelli M, Partridge SR, McGeechan K, Balestracci K, Hebden L, Wong A, et al. A Mobile Health Lifestyle Program for Prevention of Weight Gain in Young Adults (TXT2BFiT): Nine-Month Outcomes of a Randomized Controlled Trial. JMIR Mhealth Uhealth 2016 Jun 22;4(2):e78 [FREE Full text] [CrossRef] [Medline]
    26. Partridge SR, Allman-Farinelli M, McGeechan K, Balestracci K, Wong AT, Hebden L, et al. Process evaluation of TXT2BFiT: a multi-component mHealth randomised controlled trial to prevent weight gain in young adults. Int J Behav Nutr Phys Act 2016 Jan 19;13:7 [FREE Full text] [CrossRef] [Medline]
    27. Erdfelder E, Faul F, Buchner A. GPOWER: a general power analysis program. Behav Res Methods Instrum Comput 1996;28(1):1-11. [CrossRef]
    28. National Institute of Health. National Institute of Health Osteoporosis and Related Bone Diseases National Resource Center. 2018. Osteoporosis: Peak Bone Mass in Women   URL: https://www.bones.nih.gov/health-info/bone/osteoporosis/bone-mass [accessed 2019-10-01]
    29. Lu J, Shin Y, Yen M, Sun SS. Peak bone mass and patterns of change in total bone mineral density and bone mineral contents from childhood into young adulthood. J Clin Densitom 2016;19(2):180-191 [FREE Full text] [CrossRef] [Medline]
    30. Eysenbach G. CONSORT-EHEALTH: implementation of a checklist for authors and editors to improve reporting of web-based and mobile randomized controlled trials. Stud Health Technol Inform 2013;192:657-661. [Medline]
    31. Michie S, Atkins L, West R. The Behaviour Change Wheel: A Guide To Designing Interventions. United Kingdom: Silverback Publishing; 2014.
    32. Michie S, West R, Campbell R, Brown J, Gainforth H. ABC of Behaviour Change Theories. London: Silverback Publishing; 2014.
    33. Rouf A, Allman-Farinelli M. Messaging for interventions aiming to improve calcium intake in young adults-a mixed methods study. Nutrients 2018 Nov 5;10(11):pii: E1673 [FREE Full text] [CrossRef] [Medline]
    34. Project REDCap. 2018.   URL: https://projectredcap.org/ [accessed 2019-02-08]
    35. Australian Bureau of Statistics. 2018. 2033.0.55.001 - Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA), Australia, 2016   URL: https://www.abs.gov.au/ausstats/abs@.nsf/mf/2033.0.55.001 [accessed 2019-06-12]
    36. Pasco JA, Sanders KM, Henry MJ, Nicholson GC, Seeman E, Kotowicz MA. Calcium intakes among Australian women: Geelong Osteoporosis Study. Aust N Z J Med 2000 Feb;30(1):21-27. [CrossRef] [Medline]
    37. Ma J, Betts NM, Horacek T, Georgiou C, White A, Nitzke S. The importance of decisional balance and self-efficacy in relation to stages of change for fruit and vegetable intakes by young adults. Am J Health Promot 2002;16(3):157-166. [CrossRef] [Medline]
    38. Coa K, Patrick H. Baseline motivation type as a predictor of dropout in a healthy eating text messaging program. JMIR Mhealth Uhealth 2016 Sep 29;4(3):e114 [FREE Full text] [CrossRef] [Medline]
    39. Levesque CS, Williams GC, Elliot D, Pickering MA, Bodenhamer B, Finley PJ. Validating the theoretical structure of the Treatment Self-Regulation Questionnaire (TSRQ) across three different health behaviors. Health Educ Res 2007 Oct;22(5):691-702. [CrossRef] [Medline]
    40. Gardner B, Abraham C, Lally P, de Bruijn G. Towards parsimony in habit measurement: testing the convergent and predictive validity of an automaticity subscale of the Self-Report Habit Index. Int J Behav Nutr Phys Act 2012 Aug 30;9:102 [FREE Full text] [CrossRef] [Medline]
    41. Merchant G, Weibel N, Patrick K, Fowler JH, Norman GJ, Gupta A, et al. Click 'like' to change your behavior: a mixed methods study of college students' exposure to and engagement with Facebook content designed for weight loss. J Med Internet Res 2014 Jun 24;16(6):e158 [FREE Full text] [CrossRef] [Medline]
    42. Michie S, Abraham C, Whittington C, McAteer J, Gupta S. Effective techniques in healthy eating and physical activity interventions: a meta-regression. Health Psychol 2009 Nov;28(6):690-701. [CrossRef] [Medline]
    43. Althubaiti A. Information bias in health research: definition, pitfalls, and adjustment methods. J Multidiscip Healthc 2016;9:211-217 [FREE Full text] [CrossRef] [Medline]
    44. Klassen KM, Douglass CH, Brennan L, Truby H, Lim MS. Social media use for nutrition outcomes in young adults: a mixed-methods systematic review. Int J Behav Nutr Phys Act 2018 Jul 24;15(1):70 [FREE Full text] [CrossRef] [Medline]
    45. Chau MM, Burgermaster M, Mamykina L. The use of social media in nutrition interventions for adolescents and young adults-A systematic review. Int J Med Inform 2018 Dec;120:77-91. [CrossRef] [Medline]
    46. Shahril MR, Dali WP, Lua PL. A 10-week multimodal nutrition education intervention improves dietary intake among university students: cluster randomised controlled trial. J Nutr Metab 2013;2013:658642 [FREE Full text] [CrossRef] [Medline]


    Abbreviations

    CAN-DO: CAlcium Nutrition-Dietary Opportunities
    e-leaflet: electronic leaflet
    GEE: general estimating equation
    NSW: New South Wales
    OR: odds ratio
    REDCap: Research Electronic Data Capture
    SEIFA: Socio-Economic Indexes For Areas


    Edited by G Eysenbach; submitted 04.10.19; peer-reviewed by G Shakerinejad, PH Lin, J Bray; comments to author 28.10.19; revised version received 04.11.19; accepted 13.11.19; published 11.02.20

    ©Anika Rouf, Monica Nour, Margaret Allman-Farinelli. Originally published in JMIR mHealth and uHealth (http://mhealth.jmir.org), 11.02.2020.

    This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mHealth and uHealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.