Monday-focused tailored rapid interactive mobile messaging for weight management 2 (MTRIMM2): results from a randomized controlled trial

Introduction

Approximately 40% of adults in the United States (1,2) and approximately 32% of Baltimore City residents are obese (3). In 2018, the U.S. Preventative Services Task Force (USPSTF) recommended that adult patients who are obese be referred to behavioral interventions that target both diet and physical activity (PA), and suggests that patients should aim to achieve a 5% weight loss (4). For patients who are overweight but do not have cardiovascular disease risk factors, the decision to offer referrals should be individualized, but referral is recommended for overweight patients who have hypertension, dyslipidemia, pre-diabetes, or diabetes (4). Based on the results of a recent systematic literature review and meta-analysis, behavioral interventions have been found to be effective for weight loss and reduced the risk of developing type 2 diabetes mellitus (5).

Because 96% of Americans own a mobile phone, weight loss interventions delivered via mobile messaging, also known as short messaging service (SMS), offer an opportunity to reach many individuals across a range of socioeconomic groups at relatively low cost (6,7). Unlike interventions delivered in-person, mobile health interventions reduce the burden on the participants to travel to appointments (8,9). In addition, SMS interventions offer the opportunity for participants to get just-in-time messages during key decision-making moments.

However, there has been limited research on the effectiveness of SMS weight loss interventions. A 2020 systematic literature review and meta-analysis on weight loss and weight maintenance interventions in which SMS was the main intervention modality, identified only six randomized controlled trials (RCTs) conducted in the United States with adult participants (7). None of the RCTs followed the participants longitudinally after the completion of the interventions.

Furthermore, while there is growing evidence that dietary and health-related behaviors fluctuate across different days of the week, particularly between weekdays and weekends (10,11), there has been limited research regarding the implications of these findings for behavioral interventions. For instance, studies of Google search volume on different days of the week, have demonstrated increased interest in healthy behaviors early in the week (12). In addition, there is initial evidence that a consistent pattern of compensatory caloric reduction early in the week versus weekend days are associated with successful weight loss outcomes (13,14). A retrospective cohort study of 7,007 consistent users of a popular weight loss smart phone application identified fluctuating patterns of caloric consumption across the week. Greater weight loss occurred among users who recorded lower average consumption on Mondays relative to other days (10). However, to our knowledge no prior RCT has been conducted to determine whether a SMS-based weight loss intervention targeting key health behavior decision-making periods during the week results in improved weight loss compared to interventions using other intervention delivery methods.

To address these gaps, we conducted a text-messaging intervention study in a diverse sample of overweight or obese adults in Baltimore, Maryland. The objectives of the present study were to (I) evaluate the efficacy of a 16-week SMS intervention that targeted weekly periodicity in health behaviors using text messaging, and (II) assess whether changes in key study outcomes were sustained six months after participants completed the intervention. We hypothesized that at the completion of the intervention, participants receiving the SMS intervention would sustain greater weight loss than the control participants receiving the same messages bundled in a printed packet, and that these improvements would be sustained 6 months post-intervention completion. We present the following article in accordance with the CONsolidated Standards Of Reporting Trials (CONSORT) reporting checklist (available at https://dx.doi.org/10.21037/mhealth-21-3).

Methods

Monday-Focused Tailored Rapid Interactive Mobile Messaging intervention for Weight Management 2 (MTRIMM2) was a two-parallel group, superiority, RCT of a 16-week text-messaging weight loss intervention, in Baltimore, Maryland, conducted from September of 2016 until September of 2018 (ClinicalTrials.gov Identifier: NCT04506996).

Eligibility

The eligibility criteria were: (I) age ≥18 and <71, (II) BMI >25, (III) willingness to receive text messages and having a phone that was capable of receiving text-messages. English proficiency was not pre-specified as an inclusion criterion. However, all participants were proficient in English. There were no exclusion criteria.

Recruitment

Participants were recruited from the East Baltimore communities surrounding Johns Hopkins Hospital in Baltimore City, Maryland. In-person recruitment events took place at a Walgreens pharmacy, through the Brancati Center for the Advancement of Community Care, and in an indoor public food market near the medical campus. Additionally, flyers were distributed to faith-based and food-related community programs, the Johns Hopkins Weight Management Center, throughout medical and educational facilities, at clinics, and at housing complexes. Recruitment ended due to funding limitations. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013), and was approved by the Johns Hopkins Bloomberg School of Public Health’s Institutional Review Board (IRB number: 00007115). All participants in the study provided written consent and received gift cards for participating.

Intervention strategy

The content of the MTRIMM2 intervention was informed by the formative research conducted for a prior successful text-messaging intervention with primarily African American participants in Baltimore, Maryland (15). The intervention was developed using concepts from the transtheoretical model, the health belief model, and self-regulation theory (16-19). A unique feature of this intervention was that each Monday served as an opportunity for participants to have a fresh start. Participants were encouraged to reach a new specific goal each week. The goals guided participants through certain processes of behavior change by: (I) by raising consciousness (by exposing participants to new ideas that support health behavior change) and providing cues to action, (II) counter conditioning (by encouraging participants to engage in simple and attainable substitutions of healthy behaviors for unhealthy behaviors), (III) stimulus control (removing cues that might result in unhealthy behaviors), and (IV) increasing participants’ self-efficacy to avoid deviating from their goals in tempting situations (16). The remaining messages served as cues for participants to create plans to achieve their goals, provided tips for overcoming barriers participants experienced in reaching their goals, offered reminders for participants to self-monitor, and provided encouragement (16,19).

Educational intervention received by all participants

After participants completed the baseline assessment, study staff reviewed educational materials with study participants regarding: (I) the benefits of increased PA and ideas for increasing PA (20), and (II) making healthier food choices, using an American Diabetes Association handout, which encouraged increasing consumption of fruits and non-starchy vegetables, choosing whole grains instead of food made with refined grains, choosing lean sources of protein, choosing low-fat dairy and milk products, and limiting intake of sugary beverages and high-calorie snack foods and desserts (21).

After participants completed their week 8 study assessment, they were provided with additional educational materials regarding increasing PA, examples of aerobic exercise, strength training, and life-style activities that could help them be more active, and were walked through common barriers to engaging in PA as well as potential solutions. Finally, participants received tips about how to pick healthier food options at fast-food restaurants (22). See https://cdn.amegroups.cn/static/public/mhealth-21-3-1.pdf for a copy of the educational materials provided to study participants.

SMS group

The SMS group received at least seven text messages each week for 16 weeks. The messages centered around eight goal topics: (I) eating only when hungry, (II) increasing lifestyle-related PA and reducing sedentary behavior, (III) exercising more, (IV) eating less sugar and reducing calories from beverages, (V) eating a lower fat diet, (VI) eating a balanced diet, (VII) controlling portion sizes, and (VIII) making healthier food choices in social situations.

Depending on the day of the week, the text messaging topics varied (see Table 1 for the weekly schedule of messages topics, and https://cdn.amegroups.cn/static/public/mhealth-21-3-2.pdf for the full list of messages participants received). For each of the eight goal topics, six goals were developed by the study team. The order that these goal topics, goals, and default goals were chosen or sent to the study participants each week was predetermined using randomization to make sure that participants did not receive the same goals multiple times. For example, everyone who selected that they would like to receive a message regarding eating a lower fat diet on week one of the intervention, received the following goal “This week, check if the food you love has more than 10% of the saturated fat you need per day. Try to find lower fat versions by comparing nutrition labels”. Everyone who selected that topic on week nine, received another goal.

Table 2 provided examples of goals messages (sent on Monday Mornings) and reminders and tip messages (sent on Wednesday Mornings). Due to an unknown error, during the first week of the text-messaging intervention, participants’ weight was recorded as 100 pounds (which is lower than any participants’ weight). Consequently, the next time participants answered how much they weighed (in response to weekly Friday evening text-messages), they were categorized as having gained weight even if they did not and received a response to not get discouraged about their progress. The subsequent responses were recorded correctly, and participants received the correct feedback.

We used Textit in as a text-messaging platform for designing the messaging architecture and housing the messaging content. We initially used Clickatell as an aggregator (a company that sends the messages to study participants). However, in April of 2017, several participants reported that they stopped receiving text-messages for approximately two weeks. We learned that Clickatell was reporting false positive text-message delivery statuses, and this may have affected up to 25 participants who were actively receiving the intervention. In response, we switched to a different aggregator, Twilio. To deliver the missed content, we extended the intervention period for the potentially affected text-messaging participants for up to two weeks, unless participants notified us that they did not experience an interruption.

Printed messages group

Participants in the printed messages group received a printed copy of the same messages that the SMS group received (see https://cdn.amegroups.cn/static/public/mhealth-21-3-2.pdf). However, the first eight weeks’ worth of messages and craving tips were given after the baseline assessment, and the rest were given after the first follow-up assessment (~8 weeks post randomization). The messages for each week were clearly laid out and labeled. Participants were given spaces to record their answers (i.e., which goal they were selecting for each week). The study staff reviewed the first week’s messages together with the study participants to get participants comfortable with the format of the printed messages.

Randomization

A statistician configured the REDCap (23,24) randomization module (a Web-based survey program) so that the allocation sequence generated in Excel using the RAND function equally balanced treatment groups by gender and age groups (18–21, 22–40, 41–70 years). The allocation sequence using a block size of four, was concealed from all other study staff. Participants were randomized after completing the baseline assessment in REDCap. The study’s principal investigator, but not the study staff, was blinded to the randomization results.

Data collection

Data for the MTRIMM2 study was collected at four times: (I) baseline, (II) eight weeks after participants were randomized (mid-way through the intervention), (III) at the completion of the intervention 16 weeks post-randomization, and (IV) 42 weeks after randomization (6 months after completion of the intervention). Study surveys were administered using REDCap. Participants had the choice of completing the study questionnaires on their own, using their own computers, prior to the in-person study visit or using a tablet or paper form during a study visit. Participants also had the option to have the data collectors read each question and record their answers.

Outcome assessment

The primary study outcomes were percent weight change from baseline at each study visit. The secondary outcomes were the absolute change in BMI (kg/m²), fruit and vegetable intake, fat intake, PA, self-efficacy, and perceived benefits of and barriers to exercise.

Weight

Participants had a choice of two different locations to have their height and weight measured by trained study staff. Participants were asked to complete their follow-up appointments at the same location as their baseline appointment, to avoid using different scales. At one location, weight was measured using a physician beam scale, which was calibrated each day using standard weights. An electronic scale was used at the second location. Participants were asked to take off their shoes, heavy clothing, and objects from their clothing and pockets that could affect their measurements.

Height

Height was measured at least twice at each study visit, with participants’ shoes off, using stadiometer (Seca 213 Portable Measuring Rod, Chino, CA). Data collectors were trained to ensure that participants’ head positions produced parallel and level Frankfort Horizontal planes. If the two height measurements differed by >0.25 inches, a third measurement was taken, and all values were averaged. Height was then averaged across all study visits.

Daily average fruit and vegetable servings consumed

Fruit and vegetable intake was measured using National Cancer Institute’s (NCI) 19-item All-Day Fruit and Vegetable Screener, which was developed based on results from the Eating at America’s Table study (25-27). The questionnaire assessed consumption frequency and portion sizes. Participants were asked about fruits and vegetables eaten alone or in combination with other foods, as well foods and beverages such as fruit juices, lettuce/salad, potatoes, dried beans, vegetable soups, and tomato sauces. Study participants were shown measuring cups to help them better estimate portion sizes. The total daily average fruit and vegetable servings consumed was calculated using NCI’s 2001 scoring protocol and procedures for handling missing data (28).

Percent energy from fat

Percent energy from fat intake was assessed using the NCI’s 16-item Percent Energy from Fat Screener (29-31). The screener assessed the consumption frequency of items such as skim milk, margarine, butter, mayonnaise, oil, salad dressings, sausages, beef, cheese, cereal, and fruits. The eight answer choices ranged from “never [in the past year]” to “2 or more times per day.” In addition, participants were asked how often they used reduced-fat margarine when they prepared food. Percent energy from fat intake was calculated using to the NCI’s 2004 scoring protocol (32).

PA

PA performed for at least 10 minutes at a time in the week prior to completing study surveys was assessed using the long version of the International Physical Activity Questionnaire (IPAQ) (33). The questionnaire focused on four domains of PA: (I) performed at work or during volunteer activities, (II) transportation-related (such as walking to do errands), (III) performed during housework and caring for family, and (IV) recreational, sport, and leisure-time PA (33). Participants were asked about the number of days, hours, and minutes they performed each activity.

This data was used to calculate the total metabolic equivalent of tasks (MET)-minutes per week, and the MET values for walking, moderate, and vigorous PA based on the IPAQ 2005 data processing and analysis guidelines, but with two modifications (34). First, instead of truncating the hours of time spent walking, and engaging in moderate and vigorous PA to 21 hours per week for each category, we first identified the MET values that corresponded to the highest intensity PA within each category (i.e., 3.3 MET for walking, 5.5 MET in the moderate PA, 8 MET for vigorous PA). Then, we calculated the total MET-minutes of engaging in those activities for 21 hours/week, and truncated data within each category based on those MET values (4,158 MET-minutes/week for walking, 7,560 MET-minutes/week for moderate PA, and 10,080 MET-minutes/week for vigorous PA).

Second, due to a lack of instructions on how to handle truncation of data for individual variables (i.e., moderate PA while doing yard work), we were not able to calculate valid totals for each of the four domains of PA (transportation PA, PA at work, leisure time PA, and domestic PA).

Self-efficacy to lose weight

Self-efficacy to lose weight was assessed using the 20-item Weight Efficacy Lifestyle Questionnaire (WEL) (35). The items assessed ability to resist eating in 20 different situations, classified in five domains: (I) when experiencing negative emotions, (II) when a variety of food is available, (III) when experiencing social pressure, (IV) when experiencing physical discomfort, and (V) when engaging in “positive” activities such as reading, watching television, and getting ready to going to bed. A 10-point response scale was used, with answers ranging from zero to 9. Higher scores represent greater confidence in being able to avoid eating.

Benefits of and barriers to exercise

The 43-item Exercise Benefits/Barriers Scale (EBBS) was used to assess the perceived benefits of exercise and barriers to exercise experienced by study participants (36). Participants were asked how much they agreed or disagreed with statements such as and “I will live longer if I exercise” (exercise benefit) and “exercise takes too much time from family relationships” (exercise barrier). For the Exercise Benefit Scale, a four-point response scale was used: “strongly disagree” =1, “disagree” =2, “agree” =3, and “strongly agree” =4, with higher score representing greater perceived benefit of exercising. Answers for the Exercise Barrier Scale were reverse scored, and higher scores corresponded to greater perceived barriers to engaging in exercise.

Engagement with the interventions

To assess participants’ engagement with the intervention, participants rated on a 5-point Likert scale, ranging from strongly agreed to strongly disagree, whether they read every message that was either sent via SMS or provided to them in a printed format. In addition, for participants in the SMS group, the response rates to text-messages that asked questions were tracked. However, participants were informed that responding to the text messages was optional.

Statistical analysis

Descriptive statistical analyses were conducted using STATA version 15.1 software (StataCorps LLC, College Station, TX, USA). Using Statistical Analysis Software version 9.4 (SAS Institute Inc., Cary, NC, USA), we conducted intention-to-treat analysis using piecewise linear mixed regression models, with random intercepts and slopes, in order to use all available data from participants who completed at least one follow-up assessment. This reduces the chance for a Type I error (incorrectly rejecting the null hypothesis) compared to repeated measures ANOVA models. Fixed effects were included for the study arm (SMS vs. printed messaging), two time periods, and interactions of study arm and time period. A random effect was included for the participants. Depending on the trajectory of the study outcomes over time, either 0–8 and 8–42 weeks, or 0–16 and 16–42 weeks two time periods were employed in the models. To assess the change over the whole study period (from 0 to 42 weeks), we also ran linear mixed regressions to estimate the overall trend in study outcomes from 0 to 42 weeks. Results were considered statistically significant if P value were ≤0.05.

Sample size calculations

Sample size calculations were based on published data from a similar study (15). We expected an effect size of 0.946, with 95% confidence interval of 0.359 to 1.533. The SMS group was expected to experience a weight loss of 3.7 kg (SD =4.28 kg), and the control group a weight loss of 0.2 kg (SD =2.64 kg). Twenty percent participant dropout was expected. Using a two-sided t-test of mean differences, to achieve a power of 80%, a sample size of 48 participants in total was needed (95% CI: 20–308). Our recruitment goal was 200 participants.

Results

Study participants

Of the 165 people who completed study screening, 156 met the inclusion criteria, and 155 enrolled and were randomized. Seventy-eight participants were randomized to receive SMS messages (SMS intervention group), and 77 to receive printed messages (printed messages group). However, two of the participants who were randomized to receive printed messages, received the SMS intervention in error. Nevertheless, we operated under the intention-to-treat model. After randomization, 146 participants completed the 8-week follow-up assessment (94.2%), 136 completed the 16-week follow-up (87.7%), and 125 completed the 42-week follow-up assessment (80.6%). See Figure 1 for more details. Participants who completed at least one follow-up assessment were included in regression analyses (n=146, printed messages group =74, SMS group =72).

Figure 1 CONSORT diagram. ¹, of the participants who were counted as having completed study assessments, some only partially completed the assessments. ², number of participants who completed the assessments either earlier than two weeks before they were due for the assessment, or more than two weeks past the due date.

Of the participants who completed the baseline assessment, 87.1% were women, 53.5% self-identified as African American or Black, 33.5% as Caucasian or White, and 93.5% as non-Hispanic or Latinx. Participants’ ages ranged from 21–68, with the average age of 44.6 (SD =13.4, median =46.1). Most participants were employed (78.7%) and had diverse levels of educational attainment (5.8% less than high school, 24.5 % high school, 7.7% associates, trade, or vocational school, 31.6% Bachelor’s degree, and 22.6 % Master’s degree, 7.1% professional or doctoral degrees, 0.6% not reported).

There were no statistically significant differences in baseline sociodemographic characteristics between the intervention and control groups. See Table 3 for details. There were also no statistically significant differences at baseline in the study outcomes between the SMS and the printed messages groups. See Table 4 for details.

Engagement with the intervention

In response to a statement “I read every M-TRIMM2 motivational message” of the 65 participants in the SMS group who answered, 67.7% strongly agreed and 16.9% agreed. In contrast, of the 67 participants in the printed messages group who answered, 22.4% strongly agreed and 28.4% agreed. This difference was statistically significant based on the Kruskal-Wallis rank test (P<0.001), and indicated that SMS group reported reading more of the messages they received than the printed messages group.

Most participants in the SMS group responded to messages asking them to select goals for the week (mean =63.3%, SD =27.3%, 80.5% at week 1, 51% at week 16), and whether they made plans to reach their goals (mean =57.6%, SD =26.0%, 76.6% at week 1, 46.8% at week 16). Fewer participants requested tips for dealing with cravings or reported their weight (55.8% at week 2, 39.0% at week 16). See Table 5 for details. The changes in the weekly response rates are shown in Figure S1.

Study outcomes

Based on results of piecewise mixed linear regression models, at the 42-week assessment, on average, both the printed messages group and SMS group experienced small reductions in weight, BMI, and percent energy from fat consumption compared to baseline. The average percent weight loss from baseline, was 1.23 for the SMS group (95% CI: 0.36 to 2.09 P=0.006) and 0.86 for the printed messages group (95% CI: 0.01 to 1.70, P=0.047) at the 42-week assessment (Figure 2). The corresponding weight loss (kg) for the SMS group was 1.19 (95% CI: 0.35 to 2.03, P=0.006), and 0.96 kg for printed messages group (95% CI: 0.14 to 1.78, P=0.022) (Figure S2). Similarly, compared to baseline, at 42-week the average BMI decreased by 0.52 for the SMS group (95% CI: 0.20 to 0.85, P=0.002), and by 0.32 for the printed messages group (95% CI: 0.02 to 0.65, P=0.035) (Figure 3). The percentage energy from fat consumption decreased by 2.14 (95% CI: 0.89 to 3.39, P<0.001) for the SMS group and by 1.43 for the printed messages group (95% CI: 0.22 to 2.63, P=0.021) (Figure 4).

Figure 2 Results of piecewise linear mixed model: average change in percent weight loss from baseline in the SMS and printed messaging groups with 95% confidence intervals. Percent weight loss was calculated by subtracting baseline weight from weight at 8, 16, and 42 weeks and multiplying results by 100. SMS, short messaging service.

Figure 3 Results of piecewise linear mixed model: average change in BMI from baseline in the SMS and printed messaging groups with 95% confidence intervals. SMS, short messaging service; BMI, body mass index.

Figure 4 Results of piecewise linear mixed model: average differences in percent of energy from fat from baseline in the SMS and printed messaging groups with 95% confidence intervals. SMS, short messaging service.

No statistically significant differences were detected between the SMS group and the printed messages group over time for any of the study outcomes. See Table 6, Figures 2-4, and Figures S3-S10 for results of piecewise mixed linear regression models. Results of regressions examining the overall trends over the 42-week study period, without the piecewise component, are presented in Table S1. For the SMS group, response rates to text-messages were not statistically significantly associated with study outcomes (Table S2). No adverse events were reported by the study participants.

Sensitivity analysis

We conducted an as treated analysis (the two participants who received text messages by mistake as being in the SMS group) to examine whether regression results differed for the intention to treat analysis (the two participants were treated as being in the printed messages group to which they were randomized). Results remained qualitatively the same, see Table S3.

Discussion

The published literature indicates evidence of weekly fluctuating patterns of health behaviors and caloric consumption (10-14). This study sought to develop and evaluate an SMS-based weight loss periodic messaging intervention that that prompts participants to start fresh each week, embark on a new goal each Monday, and committee to engaging in healthy behaviors even on the weekends. Weight loss interventions delivered via SMS provide an opportunity to reach a large number of participants at a relatively low cost and can also allow participants to receive timely feedback without the inconvenience of in-person appointments. However, the results of our randomized control trial of the MTRIMM2 16-week weight loss intervention did not identify statistically significant differences between the SMS and printed messages groups for any of the study outcomes [percent weight loss, weight loss (kg)], BMI, perceived exercise benefits and barriers, scores on the Weight Efficacy Lifestyle questionnaire, and PA measured as METs/week [total PA, walking PA, moderate PA, vigorous PA)] at the end of the 16-week intervention or 6 months after the completion of the intervention. On average, both the printed messaging and SMS groups experienced small, statistically significant reductions in weight, BMI, and percent energy from fat consumption at the 42-week assessment compared to baseline. This suggests that both interventions have low efficacy.

A recent systematic review and meta-analysis of SMS weight loss interventions found that the effect of the intervention was generally low (pooled average weight change of −2.28 kg after an active intervention period (95% CI: −3.18 to −1.39 kg); and −0.68 kg during the maintenance phase (95% CI: −1.31 to −0.05 kg) (7). However, interventions of longer duration, and those that included in-person components were more effective (7). Over a 42-week study period, participants in MTRIMM2 SMS group lost an average of 1.19 kg and participants in the printed messages group lost 0.96 kg. We expected that a 16-week intervention was sufficient to result in behavior, because a prior, SMS-enhanced weight loss study conducted in a similar study population in Baltimore, Maryland, found that differences between intervention groups were observed in as little as three months (15). Recent literature reviews suggest that for weight loss interventions, longer intervention duration and greater tailoring of the intervention as well as the addition of nutritional counseling by a registered dietician and exercise counseling may have produced better outcomes (4,7,37-40). Greater intervention tailoring could have been achieved by setting individualized caloric intake targets, developing custom meal plans, allowing participants to set their own goals (instead of picking from three options each week), providing more robust weight loss feedback, and incorporating the use of lifestyle self-monitoring applications into the intervention (40-42).

This study had several limitations. First, we registered the study protocol on clinicaltrials.gov after the data collection was completed. However, the registration occurred before data analysis commenced, and an independent statistician who was not involved in the study design or implementation conducted the regression analyses and verified the interpretation of results. Second, participants self-selected to enroll in the intervention, and therefore the participation was not random, which limits generalizability. Third, the percent weight loss and change in BMI outcomes do not capture changes in body composition and did not allow us to assess whether participants had a reduction in adipose tissue as a result of the intervention. Fourth, the IPAQ questionnaire only asks participants to report activities that they have engaged in for at least 10 minutes at a time. However, the most recent PA guidelines for Americans state that even activities that last fewer than 10 minutes can produce health benefits (43). Furthermore, retrospective measures of PA and diet are subject to recall bias. Fifth, while data collectors were blinded to allocation, they were not blinded to the participants’ intervention group assignment after the baseline assessment was completed and participants were randomized. However, we do not expect that this influenced objectively measured outcomes such as weight. Another limitation is that up to 25 participants in the SMS group experienced a two-week interruption in receiving text-messages due to an issue with the Clickatell SMS aggregator. We extended the intervention period for those participants by two weeks, to deliver the missed content. The interruption may have slowed down the momentum participants experienced to engage in healthier behaviors and reduced the intervention’s efficacy. Finally, the response rate to text-messages was moderate for the SMS group participants (an overall average of 51.2%), and greater engagement might have resulted in better outcomes. However, response rate to text-messages might not be a good indicator of engagement because participants were told that responding to the text-messages was optional, and even if they did not pick their weekly goal, a choice would be made for them automatically. In fact, 84.6% of participants indicated an affirmative response when asked whether they had read most of the messages they received. This percentage was much higher than for the participants in the printed messages group (50.8%).

This study had several strengths. To our knowledge, this was the first RCT conducted to determine whether a SMS-based weight loss intervention targeting key health behavior decision-making periods during the week results in improved weight loss compared to receiving the same content in paper documents at two follow-up assessments. However, weight loss for the SMS and printed messages group was small, and further research is needed to examine whether more intensive interventions targeting weekly periodicity in health behaviors, with components such as individualized dietary counseling, are more efficacious than interventions that do not take into account weekly periodicity of health behaviors. A second study strength is that because this was an RCT, participants in the intervention and control group were similar in observed, and likely unobserved, characteristics, which reduces the potential for confounding. Third, the study had a relatively long duration and tracked participants during the 16-week intervention and 6 months post-intervention completion. This allowed us to assess whether changes in health outcomes that could have resulted from the intervention were sustained after the intervention was completed. In addition, we had good participant retention at 8-week (94.2%), 16-week (87.7%), and 42-week assessments (80.6%). Finally, the study sample was diverse with regard to age, race, and educational attainment of the study participants. However, future studies should examine the effectiveness of similar SMS interventions in Latinx and in Spanish speaking adults.

In conclusion, in a RCT setting, a text-messaging intervention for the purpose of promoting weight loss among overweight and obese adults was not statistically superior in efficacy to paper-based messaging. Larger studies may be needed to demonstrate superiority of the SMS messaging over paper delivery methods. Further studies are also needed to establish methods that might render the text-messaging intervention more efficacious than traditional paper educational messaging. A promising avenue for future research is to provide messaging which has been more specifically tailored to the characteristics and needs of each participant (38,42). Whether there are differences based on the population characteristics with respect to the efficacy of text-based weight control interventions also needs to be studied.

Acknowledgments

Funding: This study was funded by the Lerner Center for Public Health Promotion. The Lerner Center for Public Health Promotion was not involved in analysis or interpretation of the study results. AYK received doctoral training funding from the Johns Hopkins CLF-Lerner Fellowship, and a postdoctoral training grant from the National Institutes of Health (5T32HL129953-04). LH received NIH funding (MD012811, NR018916) and funding from the New York Regional Center for Diabetes Translation Research. The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnote

Reporting Checklist: The authors have completed the CONSORT reporting checklist. Available at https://dx.doi.org/10.21037/mhealth-21-3

Trial Protocol: Available at https://dx.doi.org/10.21037/mhealth-21-3

Data Sharing Statement: Available at https://dx.doi.org/10.21037/mhealth-21-3

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://dx.doi.org/10.21037/mhealth-21-3). This study was funded by Lerner Center for Public Health Promotion, though which AYK, LF, TK, and KW salary support. The Lerner Center for Public Health Promotion was not involved in analysis or interpretation of the study results. AYK received doctoral training funding from the Johns Hopkins CLF-Lerner Fellowship, and a postdoctoral training grant from the National Institutes of Health (5T32HL129953-04). LH received NIH funding (MD012811, NR018916) and from the New York Regional Center for Diabetes Translation Research. The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The trial was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by Johns Hopkins Bloomberg School of Public Health’s Institutional Review Board (Federalwide Assurance #0000287, IRB number: 00007115) and informed consent was taken from all individual participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Hales CM, Fryar CD, Carroll MD, et al. Differences in obesity prevalence by demographic characteristics and urbanization level among adults in the United States, 2013-2016. JAMA 2018;319:2419-29. [Crossref] [PubMed]
2. Hales CM, Carroll MD, Fryar CD, et al. Prevalence of obesity among adults and youth: United States, 2015-2016. NCHS Data Brief 2017;1-8. [PubMed]
3. Baltimore City Department of Health. Baltimore Community Health Survey 2014 report. Baltimore City, MD, 2016.
4. U. S. Preventive Services Task Force. Behavioral weight loss interventions to prevent obesity-related morbidity and mortality in adults: US Preventive Services Task Force Recommendation statement. JAMA 2018;320:1163-71. [Crossref] [PubMed]
5. LeBlanc ES, Patnode CD, Webber EM, et al. Behavioral and pharmacotherapy weight loss interventions to prevent obesity-related morbidity and mortality in adults: updated evidence report and systematic review for the US Preventive Services Task Force. JAMA 2018;320:1172-91. [Crossref] [PubMed]
6. Pew Research Center. Mobile Fact Sheet. 2019. Available online: https://www.pewresearch.org/internet/fact-sheet/mobile/. Accessed 08/26/2020.
7. Skinner R, Gonet V, Currie S, et al. A systematic review with meta-analyses of text message-delivered behaviour change interventions for weight loss and weight loss maintenance. Obes Rev 2020;21:e12999. [Crossref] [PubMed]
8. Rathbone AL, Prescott J. The use of mobile apps and SMS messaging as physical and mental health interventions: systematic review. J Med Internet Res 2017;19:e295. [Crossref] [PubMed]
9. Rogers E, Aidasani SR, Friedes R, et al. Barriers and facilitators to the implementation of a mobile insulin titration intervention for patients with uncontrolled diabetes: a qualitative analysis. JMIR Mhealth Uhealth 2019;7:e13906. [Crossref] [PubMed]
10. Hill C, Weir BW, Fuentes LW, et al. Relationship between weekly patterns of caloric intake and reported weight loss outcomes: retrospective cohort study. JMIR Mhealth Uhealth 2018;6:e83. [Crossref] [PubMed]
11. Yang PH, Black JL, Barr SI, et al. Examining differences in nutrient intake and dietary quality on weekdays versus weekend days in Canada. Appl Physiol Nutr Metab 2014;39:1413-7. [Crossref] [PubMed]
12. Ayers JW, Althouse BM, Johnson M, et al. What's the healthiest day? Circaseptan (weekly) rhythms in healthy considerations. Am J Prev Med 2014;47:73-6. [Crossref] [PubMed]
13. Orsama AL, Mattila E, Ermes M, et al. Weight rhythms: weight increases during weekends and decreases during weekdays. Obes Facts 2014;7:36-47. [Crossref] [PubMed]
14. Mattila E, Lappalainen R, Parkka J, et al. Use of a mobile phone diary for observing weight management and related behaviours. J Telemed Telecare 2010;16:260-4. [Crossref] [PubMed]
15. Lin M, Mahmooth Z, Dedhia N, et al. Tailored, interactive text messages for enhancing weight loss among African American adults: the TRIMM randomized controlled trial. Am J Med 2015;128:896-904. [Crossref] [PubMed]
16. Prochaska JO, Velicer WF. The transtheoretical model of health behavior change. American Journal of Health Promotion 1997;12:38-48. [Crossref] [PubMed]
17. Janz NK, Becker MH. The Health Belief Model: a decade later. Health Education Quarterly 1984;11:1-47. [Crossref] [PubMed]
18. Baumeister RF, Vohs KD. Strength Model of Self-Regulation as limited resource: assessment, controversies, update. In: Olson JM, Zanna MP, editors. Advances in Experimental Social Psychology, Vol 54. Advances in Experimental Social Psychology. San Diego: Elsevier Academic Press Inc; 2016:67-127.
19. Rosenstock IM, Strecher VJ, Becker MH. Social learning theory and the Health Belief Model. Health Educ Q 1988;15:175-83. [Crossref] [PubMed]
20. Centers for Disease Control and Prevention. Physical activity and health: a report of the Surgeon General. Atlanta, Georgia. 1996. Available online: https://www.cdc.gov/nccdphp/sgr/pdf/sgraag.pdf. Accessed 12/09/2020.
21. American Diabetes Association. The Best Food Choices. Arlington, VA, 2015.
22. American Diabetes Association. What can I eat? Arlington, VA, 2015.
23. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377-81. [Crossref] [PubMed]
24. Harris PA, Taylor R, Minor BLThe REDCap consortium, et al. building an international community of software platform partners. J Biomed Inform 2019;95:103208. [Crossref] [PubMed]
25. National Institutes of Health. National Institutes of Health Eating at America's Table Study Quick Food Scan. 2000. Available online: https://epi.grants.cancer.gov/diet/shortreg/instruments/eats_all-day.pdf. Accessed 08/06/2020.
26. Subar AF, Thompson FE, Kipnis V, et al. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires : the Eating at America's Table Study. Am J Epidemiol 2001;154:1089-99. [Crossref] [PubMed]
27. Thompson FE, Subar AF, Smith AF, et al. Fruit and vegetable assessment: performance of 2 new short instruments and a food frequency questionnaire. J Am Diet Assoc 2002;102:1764-72. [Crossref] [PubMed]
28. National Cancer Institute. Scoring the All-Day Screener. 2005. Available online: https://epi.grants.cancer.gov/diet/screeners/fruitveg/scoring/allday.html#how. Accessed 08/06/2020.
29. National Cancer Institute. National Cancer Institute Quick Food Scan: Percent Energy from Fat Screener. 2000. Available online: https://epi.grants.cancer.gov/diet/shortreg/instruments/percent-energy-from-fat-screener.pdf. Accessed 08/06/2020.
30. Thompson FE, Midthune D, Subar AF, et al. Development and evaluation of a short instrument to estimate usual dietary intake of percentage energy from fat. J Am Diet Assoc 2007;107:760-7. [Crossref] [PubMed]
31. Thompson FE, Midthune D, Williams GC, et al. Evaluation of a short dietary assessment instrument for percentage energy from fat in an intervention study. J Nutr 2008;138:193S-9S. [Crossref] [PubMed]
32. National Cancer Institute. Percentage Energy from Fat Screener: scoring procedures. 2004. Available online: https://epi.grants.cancer.gov/diet/screeners/fat/scoring.html. Accessed 08/06/2020 2020.
33. Booth M. Assessment of physical activity: an international perspective. Res Q Exerc Sport 2000;71:114-20. [Crossref] [PubMed]
34. IPAQ Research Committee. Guidelines for the data processing and analysis of the International Physical Activity Questionnaire. 2005. Available online: https://sites.google.com/site/theipaq/scoring-protocol. Accessed 08/06/2020.
35. Clark MM, Abrams DB, Niaura RS, et al. Self-efficacy in weight management. J Consult Clin Psychol 1991;59:739-44. [Crossref] [PubMed]
36. Sechrist KR, Walker SN, Pender NJ. Development and psychometric evaluation of the exercise benefits/barriers scale. Res Nurs Health 1987;10:357-65. [Crossref] [PubMed]
37. Singh N, Stewart RAH, Benatar JR. Intensity and duration of lifestyle interventions for long-term weight loss and association with mortality: a meta-analysis of randomised trials. BMJ Open 2019;9:e029966. [Crossref] [PubMed]
38. Ryan K, Dockray S, Linehan C. A systematic review of tailored eHealth interventions for weight loss. Digit Health 2019;5:2055207619826685. [Crossref] [PubMed]
39. Ahern AL, Wheeler GM, Aveyard P, et al. Extended and standard duration weight-loss programme referrals for adults in primary care (WRAP): a randomised controlled trial. Lancet 2017;389:2214-25. [Crossref] [PubMed]
40. Sherrington A, Newham JJ, Bell R, et al. Systematic review and meta-analysis of internet-delivered interventions providing personalized feedback for weight loss in overweight and obese adults. Obes Rev 2016;17:541-51. [Crossref] [PubMed]
41. Cavero-Redondo I, Martinez-Vizcaino V, Fernandez-Rodriguez R, et al. Effect of behavioral weight management interventions using lifestyle mHealth self-monitoring on weight loss: a systematic review and meta-analysis. Nutrients 2020;12:1977. [Crossref] [PubMed]
42. De Leon E, Fuentes LW, Cohen JE. Characterizing periodic messaging interventions across health behaviors and media: systematic review. J Med Internet Res 2014;16:e93. [Crossref] [PubMed]
43. U.S. Department of Health and Human Services. Physical Activity Guidelines for Americans, 2nd edition. Washington, DC: U.S., 2018.