This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Daily self-monitoring of diet and physical activity behaviors is a strong predictor of weight loss success. Text messaging holds promise as a viable self-monitoring modality, particularly among racial/ethnic minority populations.
This pilot study evaluated the feasibility of a text messaging intervention for weight loss among predominantly black women.
Fifty obese women were randomized to either a 6-month intervention using a fully automated system that included daily text messages for self-monitoring tailored behavioral goals (eg, 10,000 steps per day, no sugary drinks) along with brief feedback and tips (n=26) or to an education control arm (n=24). Weight was objectively measured at baseline and at 6 months. Adherence was defined as the proportion of text messages received in response to self-monitoring prompts.
The average daily text messaging adherence rate was 49% (SD 27.9) with 85% (22/26) texting self-monitored behavioral goals 2 or more days per week. Approximately 70% (16/23) strongly agreed that daily texting was easy and helpful and 76% (16/21) felt the frequency of texting was appropriate. At 6 months, the intervention arm lost a mean of 1.27 kg (SD 6.51), and the control arm gained a mean of 1.14 kg (SD 2.53; mean difference –2.41 kg, 95% CI –5.22 to 0.39;
Given the increasing penetration of mobile devices, text messaging may be a useful self-monitoring tool for weight control, particularly among populations most in need of intervention.
Clinicaltrials.gov: NCT00939081; http://clinicaltrials.gov/show/NCT00939081 (Archived by WebCite at http://www.webcitation.org/6KiIIcnk1).
Regular self-monitoring of diet and physical activity behaviors is one of the strongest predictors of weight loss success [
Text messaging shows promise as an alternative eHealth self-monitoring approach [
Text messaging has become ubiquitous [
We recruited women aged 25-50 years, with a body mass index (BMI) greater than or equal to 25 kg/m2. Other inclusion criteria were willingness to (1) come to all study assessments over 6 months, (2) use a personal cell phone to send and receive up to 5 texts per day for 6 months without compensation for the text messages, and (3) be randomized into either treatment arm. Exclusion criteria included pregnancy or planned pregnancy within the next 6 months and a history of myocardial infarction or stroke within the past 2 years.
We partnered with a nonprofit church-based community wellness organization located in Raleigh, NC to recruit participants. The wellness organization provided the location for enrollment events and aided in recruitment by advertising the study in common spaces and during community meetings and church services. Additional recruitment was conducted in the surrounding community via flyers posted in neighborhood businesses and outreach to adults in the area who had expressed interest in weight loss research trials. Recruitment took place between June and September 2010.
Interested participants visited a study website to complete initial eligibility screening that assessed self-reported age, gender, height, weight, and race (American Indian or Alaska Native, Asian, black or African American, Native Hawaiian or other Pacific Islander, white, or other). Eligible participants were then invited to an in-person enrollment event. Ineligible participants were directed to a website where they could access publicly available weight loss information. At the enrollment events, study staff obtained informed consent and collected baseline anthropometric and survey measures. Participants were then randomized to the intervention arm or the education control arm using a computer-generated algorithm. Because of the pilot nature of this study, participants were re-evaluated at 6 months, with no additional study visits. Participants received a US $35 gift card to a local store as an incentive for participation. The Duke University Institutional Review Board approved this study.
The intervention (Shape Plan) included daily tracking of tailored behavior change goals through text messaging and personalized daily and weekly feedback via text messaging and email, respectively. Participants also received information sheets about behavioral goals, a pedometer, 2 face-to-face group sessions, and a skills training DVD.
Behavior change goals were determined using the interactive obesity treatment approach (iOTA) [
At baseline, the iOTA algorithm ranked behavior change goals for intervention participants. Participants were instructed to self-monitor the 2 top goals daily for 12 weeks. All intervention participants also received a walking goal of at least 7000 steps every day. The physical activity goal increased based on participants’ performance, up to 10,000 steps per day. The survey was re-administered at 3 months and updated goals were assigned using the same algorithm.
Walk 7000/8000/10,000 steps every day
No sugary drinks
Eat 5 or more fruits and vegetables every day
No chips, cookies, or candy
Switch to low-fat dairy
No fast food
Eat breakfast every day
Watch no more than 2 hours of TV every day
No fried food
No snacks or dessert after dinner
No more than 1 alcoholic drink per day
Eat red meat no more than once a week
At the baseline enrollment event, Shape Plan participants received a group-based orientation to the intervention led by community health educators experienced in delivering information on weight control. The orientation included a review of the iOTA goals, calorie balance, a demonstration of the text messaging self-monitoring and feedback, and an action planning session. Goal setting and text messaging monitoring began the following day. At 3 months, in an effort to reduce the number of face-to-face meetings, participants received a set of videos with skills training information on topics such as healthy eating patterns, eating cues, recognizing hunger, problem solving to meet goals, goal setting, exercise tips, and safety and action planning for the upcoming Shape Plan goals. At 6 months, participants received another hour-long group face-to-face session that focused on problem solving, assessment of overall progress, and tips for maintaining behavior changes.
The text messaging protocol included 1 daily morning text message at 8:00 am, which asked participants to report performance on their goals from the previous day (eg, “How many steps did you walk yesterday?”;
Additionally, Shape Plan participants received a weekly automated email on Sundays with a summary of their progress. Participants with at least 3 days of self-monitoring data received a weekly email with personalized feedback that included a summary of goal attainment and a graph of progress over the previous week. For participants with low adherence (3 or fewer texts in 1 week), the email did not include a summary, but rather acted as a prompt to improve adherence (eg, “We only received 2 text messages from you this week. In order for you to be most successful losing weight in Shape Plan, you should track your numbers and send us a text every day”).
To control for contact and isolate the behavior change goals, self-monitoring via text messaging and feedback, participants randomized to the education control arm received (1) 2 in-person group education sessions, one at baseline and another at 6 months; (2) a set of videos at 3 months that covered topics such as healthy eating patterns, eating cues, recognizing hunger, exercise recommendations, and how to read a nutrition facts food label; (3) pedometers; and (4) a “prescription” to walk 10,000 steps per day. Control arm participants received no text messaging during the study period, but had the option to receive a 3-month version of the text messaging intervention after the study was complete.
At baseline, a variety of sociodemographic variables were collected through an online survey to characterize the sample, including age, race/ethnicity, household income, education, marital status, and employment.
A study database collected and stored text messaging self-monitoring data. Adherence was defined as the proportion of self-monitoring texts received of the number expected over the 6-month period (N=167). We examined total adherence and adherence by study week.
At baseline and 6 months, trained staff measured participant heights to the nearest 0.1 cm using a calibrated stadiometer (Seca 214). Weights were measured to the nearest 0.1 kg using an electronic scale (Seca Model 876) [
At 6 months, intervention participants completed a 23-item online questionnaire to assess intervention satisfaction. Using a 4-point Likert scale with response options ranging from strongly agree (=4) to strongly disagree (=1), participants rated whether they found daily self-monitoring through text messaging to be easy, helpful overall, helpful for increasing daily steps, and important. Similarly, participants reported whether daily text messages were the appropriate frequency and whether they were satisfied with the feedback received via text messaging.
We used chi-square tests and
Screenshot of self-monitoring via text message with feedback.
Flowchart of participant enrollment and retention in the study.
Mean age of participants was 38.3 years (SD 8.2). Participants were obese (BMI mean 35.8 kg/m2, SD 6.1), and had an average weight of 99.1 kg (SD 20.0) (
The mean daily goal attainment score over the 6-month intervention period was 6.3 (SD 2.8, IQR 4.0-8.2), indicating moderate-high adherence to behavioral goals. Similarly, the mean number of steps reported was 4994 (SD 2741, IQR 3016-6489). There was no significant correlation between average goal attainment score and text messaging adherence (
Baseline characteristics of participants by study arm.
Characteristic | Total |
Control |
Shape Plan |
|
|
|
|
|
|
|
Married or living with a partner | 25 (50) | 9 (38) | 16 (62) |
|
Divorced/separated or never married | 25 (50) | 15 (62) | 10 (38) |
|
|
|
|
|
|
Employed | 41 (82) | 19 (79) | 22 (85) |
|
Unemployed | 9 (18) | 5 (21) | 4 (15) |
|
|
|
|
|
|
4 year college degree or higher | 32 (64) | 15 (62) | 17 (65) |
|
Less than a 4 year college degree | 18 (36) | 9 (38) | 9 (35) |
|
|
|
|
|
|
Non-Hispanic black | 41 (82) | 18 (75) | 23 (88) |
|
Non-Hispanic other | 9 (18) | 6 (25) | 3 (12) |
|
|
|
|
|
|
<40,000 | 16 (32) | 9 (38) | 7 (27) |
|
40,000-69,999 | 18 (36) | 7 (29) | 11 (42) |
|
≥70,000 | 16 (32) | 8 (33) | 8 (31) |
Age (years), mean (SD) | 38.3 (8.2) | 39.0 (9.0) | 37.6 (7.4) | |
Weight (kg), mean (SD) | 99.1 (20.0) | 96.0 (23.1) | 102.0 (16.6) | |
BMI (kg/m2), mean (SD) | 35.8 (6.1) | 34.6 (5.8) | 36.9 (6.2) |
Adherence to daily self-monitoring via text message by study week (n=26).
At 6 months, most participants strongly agreed that texting was easy (70%, 16/23) and helpful (68%, 15/22), and 76% (16/21) either somewhat or strongly agreed that the text messages helped them increase the number of daily steps walked. Almost three-quarters of participants (71%, 15/21) reported that it took less than 3 minutes to reply to texts and most (79%, 15/19) responded to prompts for self-monitoring via text message in the morning. More than half (57%, 12/21) felt that receiving daily texts was very important and approximately three-quarters of participants (76%, 16/21) felt the frequency of texting was appropriate. Most (82%, 18/22) were satisfied with the feedback content they received via text messaging and 62% (13/21) felt the feedback was very important.
Change in weight and body mass index (BMI) between baseline and 6 months by study group (N=50).
Weight change variables | Control, mean (SD) |
Shape Plan, mean (SD) |
Mean difference (95% CI) |
|
Change in weight (kg) | 1.14 (2.53) | –1.27 (6.51) | –2.41 (–5.22, 0.39) | .09 |
Percent weight loss (%) | 1.32 (2.77) | –0.97 (5.35) | –2.29 (–4.70, 0.12) | .06 |
Change in BMI (kg/m2) | 0.42 (0.90) | –0.47 (2.42) | –0.89 (–1.93, 0.15 ) | .09 |
In this pilot study, we found that daily text messaging for behavioral self-monitoring is both feasible and positively perceived. Approximately half of participants were fully adherent to daily self-monitoring through text messages during the 6-month study and 84% stayed active in the intervention throughout the study period. In contrast to previous studies that included paper-based self-monitoring modes [
The eHealth and mobile health (mHealth) approaches to self-monitoring offer numerous advantages over traditional approaches. Self-monitoring via text messaging is cheaper, easier to program, and more proximal to behavior changes as compared to paper-based and Web-based self-monitoring. Both eHealth and mHealth self-monitoring strategies seem not to exhibit the same steep decline typically seen in self-monitoring adherence [
What participants are self-monitoring may be as important to adherence as how they are self-monitoring. Typical self-monitoring behaviors include detailed reports of food intake, including portion size, and calorie and fat content. Although effective, self-monitoring of this type exhibits poor adherence [
To date, our weight loss outcomes are similar to those of other text messaging trials [
Our pilot study used text messaging to collect self-monitoring data on diet and physical activity behaviors, but we did not gather data on body weight or include any type of coaching support. This may have affected the weight losses achieved because body weight self-monitoring has been shown to be effective in the absence of self-monitoring of other behaviors [
Black women have the highest prevalence of obesity compared to any other group [
Additionally, mobile phone use is high among blacks, and black individuals are more likely than white individuals are to use mobile phones to look for health or medical information [
This study has several strengths. Most of our sample (82%) were black women, which is a group typically underrepresented in weight control research. Although a strength of our study, the findings may not generalize to other populations and settings. The goal of this study was to test the feasibility of text messaging for self-monitoring through a low-intensity weight loss intervention among an understudied population. We isolated the impact of text messaging self-monitoring along with feedback with a control arm that received comparable group information sessions and the use of a pedometer. This study also has some limitations worth mentioning. A few participants (n=8) experienced barriers to participation, such as cost of text messaging and disconnected cell phone service, or they were no longer interested in participating. Although providing phones and/or text messaging plans may have enhanced our adherence rates, the use of personal cell phones provides insight into the “real world” feasibility of text messaging for weight control and improves the generalizability of the intervention.
Although comparable to other eHealth weight control interventions, higher adherence rates are needed to produce greater weight losses. This pilot study was low intensity and did not include any contact with study staff outside the assessment visits. Our main goal was to assess the feasibility of using text messaging for self-monitoring behavioral goals among a predominantly racial/ethnic minority population. To enhance adherence, future studies using this approach might also include daily tracking of weight and provide feedback on weight loss progress via text message. More frequent skills training through monthly videos may also boost adherence rates. Including elements of accountability and support (eg, monthly coaching calls with a lifestyle counselor) can also be an effective way to enhance adherence, but including these components will increase intensity and make it more difficult to ascertain the independent effects of the text messaging. Given that this was a pilot study, the small sample size limited our power to assess whether this intervention led to significantly greater weight loss as compared to an education control arm. Future studies should examine the efficacy of this approach with a larger sample size, longer duration, and multiple measures throughout the study period.
Text messaging holds promise as a self-monitoring modality for weight control, particularly among groups most at risk for obesity-related morbidities. Given that the majority of evidence indicates that greater adherence leads to better outcomes, our study suggests that mHealth-based approaches to self-monitoring may enhance engagement and reduce the burdens commonly associated with other modes. Our intervention was relatively low intensity and has greater potential for dissemination compared to higher intensity interventions. As technology penetration increases in the target population, the use of this modality will become increasingly relevant and helpful as an intervention tool for weight control.
CONSORT-EHEALTH checklist V1.6.2 [
body mass index
electronic health
interactive obesity treatment approach
mobile health
The authors would like to thank the support received by Rachel Kroll Bordogna and staff at the Duke Obesity Prevention Program, particularly Michele Lanpher, Daniel Dix, and Jade Miller. We are also grateful to the community organization, Diversified Resources for Better Living, that was pivotal in helping recruit our participants. Lastly, we would like to thank the women participating in the Shape Plan. This trial was funded by grant K22CA126992 awarded to Dr Gary Bennett.
None declared.