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.
Mobile technology to promote exercise is effective; however, most evidence is from studies of younger groups in high-income countries. Investigating if short message service (SMS) texting can affect exercise participation in older adults from an upper-middle-income country is important considering the proliferation of mobile phones in developing regions and the increased interest of older adults in using mobile phones.
The main objective was to examine the short- and long-term effects of SMS text messaging on exercise frequency in older adults. Secondary objectives were to investigate how SMS text messages impact study participants’ exercise frequency and the effects of the intervention on secondary outcomes.
The Malaysian Physical Activity for Health Study (myPAtHS) was a 24-week, 2-arm, parallel randomized controlled trial conducted in urban Malaysia. Participants were recruited via health talks in resident associations and religious facilities. Older Malaysians (aged 55-70 years) who used mobile phones and did not exercise regularly were eligible to participate in the study. Participants randomly allocated to the SMS texting arm received an exercise booklet and 5 weekly SMS text messages over 12 weeks. The content of the SMS text messages was derived from effective behavior change techniques. The non-SMS texting arm participants received only the exercise booklet. Home visits were conducted to collect outcome data: (1) exercise frequency at 12 and 24 weeks, (2) secondary outcome data (exercise self-efficacy, physical activity–related energy expenditure, sitting time, body mass index, grip and leg strength) at baseline and at 12 and 24 weeks. Intention-to-treat procedures were applied for data analysis. Semistructured interviews focusing primarily on the SMS text messages and their impact on exercise frequency were conducted at weeks 12 and 24.
In total, 43 participants were randomized into the SMS texting arm (n=22) and the non-SMS texting arm (n=21). Study-unrelated injuries forced 4 participants to discontinue after a few weeks (they were not included in any analyses). Overall retention was 86% (37/43). After 12 weeks, SMS texting arm participants exercised significantly more than non-SMS texting arm participants (mean difference 1.21 times, bias-corrected and accelerated bootstrap [BCa] 95% CI 0.18-2.24). Interview analysis revealed that the SMS text messages positively influenced SMS texting arm participants who experienced exercise barriers. They described the SMS text messages as being encouraging, a push, and a reminder. After 24 weeks, there was no significant difference between the research arms (mean difference 0.74, BCa 95% CI –0.30 to 1.76). There were no significant effects for secondary outcomes.
This study provides evidence that SMS text messaging is effective in promoting exercise in older adults from an upper-middle-income country. Although the effects were not maintained when SMS text messaging ceased, the results are promising and warrant more research on behavioral mobile health interventions in other regions.
Clinicaltrials.gov NCT02123342; http://clinicaltrials.gov/ct2/show/NCT02123342 (Archived by WebCite at http://www.webcitation.org/6eGSsu2EI).
Population aging is a global phenomenon that is projected to continue [
The global rise in the older adult population is linked to an increase in the prevalence of noncommunicable diseases (NCDs), disability, and other health problems which, in turn, increase the burden on public and private health care systems [
Behavioral health interventions focusing on PA and/or exercise are increasingly delivered via mHealth approaches [
Text messaging has been shown to be successful in promoting PA and/or exercise in young adults [
Only one study reported on the effects of SMS text messaging in older adults [
The Malaysian Physical Activity for Health Study (myPAtHS) is a randomized controlled trial (RCT) for older Malaysians who do not follow a regular exercise routine. We chose this age group because population aging is a great challenge for Malaysia [
Our study aimed to (1) determine if older Malaysians receiving an exercise booklet and weekly SMS text messages exercise more than participants who only receive an exercise booklet, (2) examine if the effects of the SMS text messages are maintained when the SMS text messages are removed, (3), investigate how the SMS text messages support participants to exercise, and (4) investigate the effects of the SMS text messages on secondary outcomes (eg, exercise self-efficacy, weekly PA-related energy expenditure, daily sitting time, body mass index [BMI], grip strength, and lower body strength).
The myPAtHS is a RCT that uses a parallel study design. All participants were introduced to a set of exercises and received an exercise booklet. Participants randomized into the SMS text-messaging arm (SMS texting arm) received an additional 60 text messages over 12 weeks. Participants randomized into the other arm did not receive SMS text messages (non-SMS texting arm). After enrollment, the primary study outcome was assessed at weeks 12 and 24. The study design and protocol were approved by the Faculty of Medicine Ethics Committee, University of Malaya, and was registered (Clinicaltrials.gov NCT02123342). This trial is reported according to the CONSORT statement [
The study took place in urban Malaysia (Kuala Lumpur and Petaling Jaya) from June 2014 to January 2015. In Malaysia, 73% of the population lives in urban areas and Kuala Lumpur and Petaling Jaya are the most densely populated cities [
Eligible participants were English-speaking community-dwelling Malaysians aged between 55 and 70 years, who were not exercising regularly (no structured exercise more than once weekly), had no health conditions that would restrict moderate exercise, used a mobile phone with SMS text-messaging function, and were interested in health-promoting exercise.
Participants were recruited from local resident associations and religious facilities in April and May 2014. With the support of representatives from the respective organizations, one study team member conducted health talks for older adults within the recruitment area. The study was briefly introduced as an exercise for health program and eligibility criteria were described (SMS text messaging was not mentioned). Those who were interested in taking part were given an information sheet and asked to provide contact details so that a study team member could call them later. Approximately one week after the health talks, potential participants were called. During this call, eligibility criteria were checked, initial oral consent was obtained, and a baseline home visit was scheduled. Home visits were conducted because some participants did not have personal transportation and the public transportation system is not easily accessible. During the home visits, final eligibility checks were conducted, study procedures were explained (eg, time lines, potential risks), informed consent was obtained, and enrollment finalized.
The overall sample was stratified into participants enrolling with their spouse and participants enrolling without a spouse. There is evidence that older adults enrolling in an exercise intervention with a spouse exercise significantly more than those who do not [
During the baseline home visit, all participants were introduced to a set of exercises and received an exercise booklet (myPAtHS booklet) developed by one of the study team members, an exercise physiologist with experience in training older adults. This booklet contained information on the benefits of exercise, some safety instructions, and descriptions of 12 age-appropriate strengthening exercises that could be executed without any specific equipment. Brief warm-up and cool-down sections were included as well. The exercises targeted major muscle groups of the arms/shoulders, upper trunk/neck, and legs. They were described using pictures, explanations of key movements, and hints where the exercises should be felt. One practical exercise session was conducted during the initial home visit to ensure correct execution. Participants were advised to exercise as often as possible each week to increase health benefits, but no other formal recommendations were provided. To ensure that participants’ mobile phones were operational and participants were competent using the SMS text message function, they were asked to confirm receipt of a text message sent before the baseline home visit.
During the 12 weeks following the baseline home visit, 60 SMS text messages were sent to SMS texting arm participants (during weekdays). Text messages were sent automatically via an online tool specifically developed for this study. This tool allowed the research team to schedule the SMS text messages for every participant and it was also used to confirm delivery of the SMS text messages (
In this study, a mixed methods approach was applied to collect outcome data. Quantitative data was supplemented by qualitative data from semistructured interviews.
The primary study outcome was weekly exercise frequency (exercise sessions using the exercise booklet). It was assessed immediately after the 12-week SMS texting intervention period and after 24 weeks. This outcome was measured with an exercise log appended to the exercise booklet. Participants were asked to record dates, times, and duration of exercise sessions. During the baseline home visit, participants were shown how to record their exercise routine and one trial was conducted to ensure correct data entry. Additionally, one example of a correct entry was provided (in the booklet). Completed logs were exchanged for new ones on subsequent follow-up home visits.
A number of secondary outcomes were assessed at baseline and at weeks 12 and 24. Exercise self-efficacy is strongly associated with exercise participation in older adults [
Physical activity-related energy expenditure, in weekly Metabolic Equivalent of Task (MET) minutes (MET-minutes), and daily time spent sitting (in hours) were measured using the short form of the International Physical Activity Questionnaire (IPAQ). The validity and reliability of this instrument are well established and it is widely used [
In addition, BMI (in kg/m2) was calculated from body height and weight using the Seca Clara 803 Digital Personal Scale (Seca GmbH & Co KG, Hamburg, Germany). We assessed maximum grip strength (in kg force) of the dominant hand using the North Coast Hydraulic Hand Dynamometer (North Coast Medical Inc, Morgan Hill, CA, USA). This device has been used in previous studies with different groups of older adults and provided valid and reliable data [
During the follow-up home visits, we also conducted semistructured interviews with all research participants to complement the quantitative data. The interviews lasted approximately 20 minutes. We were primarily interested in how the participants in the SMS texting arm perceived the SMS text messages. Questions about the impact of the SMS text messages, their content, and what was done with them were discussed (
A total of 36 participants (18 per arm) was estimated to provide 80% power at α=.05 to detect a difference of one weekly exercise session between the arms at week 12, assuming a standard deviation of 1.1 session. We anticipated a dropout rate of 15%; hence, we aimed to include 42 participants (21 per arm) [
Statistical analysis using SPSS version 21.0 included descriptive statistics of age, sex, education, employment status, health status, and marital status. The intention-to-treat principle framed the analyses. However, intervention-unrelated injuries resulted in dropouts at week 12 and no primary outcome data was collected from these participants (n=4), consequently invalidating the use of imputation procedures. None of these participants were included in any of the analyses. For all other analyses, we used the last observation carried forward procedure for missing data. We also conducted a per protocol analysis for those participants with complete outcome data using the same procedures as in the intention-to-treat analysis (
Weekly exercise frequency at week 12 (SMS texting period) was compared between study arms using an independent
For the secondary outcomes, data were converted into 2 change variables: one between baseline and week 12 and one between baseline and week 24. For each variable, an ANCOVA comparing the change scores between the arms at each time point was conducted with the baseline scores entered as a covariate. For each arm, we estimated model-adjusted means, 95% confidence intervals, and
The first author transcribed (not verbatim) the interviews and categorized responses into broad predefined themes (eg, exercise program, SMS text message content, effects of the SMS text messages). Themes were further divided into subthemes that were partly derived from the responses of the participants (eg, exercise benefits, perception of how the SMS text messages affected exercise). Direct quotations from participants were extracted to exemplify the results derived from the interviews. Finally, a coauthor checked the interview analysis results for accuracy and discrepancies were resolved via discussion.
Participants were recruited in April and May 2014, and follow-up data were collected until January 2015.
Demographic characteristics of the research participants.
Characteristic | SMS texting |
Non-SMS texting |
Total |
|
Age (years), mean (SD) | 63.64 (4.58) | 62.90 (4.48) | 63.28 (4.50) | |
|
|
|
|
|
|
Male | 6 (27) | 5 (24) | 11 (26) |
|
Female | 16 (73) | 16 (76) | 32 (74) |
|
|
|
|
|
|
Secondary | 6 (27) | 4 (19) | 10 (23) |
|
Postsecondary | 1 (5) | 3 (14) | 4 (9) |
|
College/university | 15 (68) | 14 (67) | 29 (68) |
|
|
|
|
|
|
Working | 6 (27) | 4 (19) | 10 (23) |
|
Not working | 16 (73) | 17 (81) | 33 (77) |
|
|
|
|
|
|
Fair | 4 (18) | 3 (14) | 7 (16) |
|
Good | 16 (73) | 13 (62) | 29 (67) |
|
Very good or Excellent | 2 (9) | 5 (24) | 7 (16) |
|
|
|
|
|
|
Single/Separated/Widowed | 4 (18) | 4 (19) | 8 (19) |
|
Married | 18 (82) | 17 (81) | 35 (81) |
Study participant flow.
Over the 12-week intervention period, participants in the SMS texting arm exercised more frequently per week (mean 3.74, SD 1.34) compared to participants in the non-SMS texting arm (mean 2.52, SD 1.85). This difference (mean difference 1.21, bias-corrected and accelerated [BCa] 95% CI 0.18-2.24) was significant (
Weekly exercise frequency decreased by 0.43 sessions (95% CI 0.12-0.74) from week 12 to week 24 in the overall sample (
Treatment effects on secondary outcomes.
Outcome | SMS texting, mean (SD) |
Non-SMS texting, mean (SD) |
Week 12 | Week 24 | ||||||
|
Baseline | Change to week 12a | Change to week 24a | Baseline | Change to week 12a | Change to week 24a | Adj differencea (95% CI) |
|
Adj differencea (95% CI) | P |
Exercise self-efficacy score | 81.94 (18.74) | –6.47 (19.67) | –3.02 (25.52) | 81.55 (17.53) | –9.87 (19.65) | –14.32 (25.51) | –3.39 (–16.21, 9.43) | .60 | –11.31 (–27.95, 5.34) | .18 |
PA-related energy expenditure (weekly MET-minutes) | 662.29 (497.29) | 383.43 (843.42) | 434.69 (728.69) | 968.71 (1479.10) | 377.43 (842.56) | 171.97 (728.22) | –5.99 (–558.14, 546.15) | .98 | –262.72 (–739.77, 214.33) | .27 |
Daily sitting time (hours) | 7.28 (3.39) | –0.36 (2.09) | –0.10 (1.90) | 8.52 (2.20) | –1.17 (2.09) | –0.77 (1.89) | –0.81 (–2.19, 0.58) | .25 | –0.67 (–1.92, 0.58) | .29 |
BMI (kg/m2) | 23.50 (3.47) | 0.23 (0.51) | 0.15 (0.68) | 22.39 (2.81) | 0.32 (0.51) | 0.28 (0.68) | 0.09 (–0.25, 0.42) | .60 | 0.13 (–0.32, 0.58) | .56 |
Grip strength (kg) | 25.93 (8.70) | 0.32 (2.01) | 1.44 (2.55) | 25.51 (6.34) | 0.31 (2.01) | 1.09 (2.55) | 0.02 (–1.32, 1.29) | .98 | –0.34 (–2.01, 1.32) | .68 |
Lower body strength (repetitions in 30-sec chair-stand test) | 13.44 (3.42) | 2.21 (3.40) | 3.55 (2.80) | 14.90 (3.81) | 2.77 (3.39) | 3.43 (2.79) | 0.56 (–1.67, 2.80) | .61 | –0.12 (–1.96, 1.72) | .90 |
a Adjusted for baseline.
Over the 24-week study period, a total of 4 adverse events occurred, all in the SMS texting arm (slipped disk: n=2; shoulder injury: n=1; hospitalization: n=1), none of which resulted directly from the study.
The semistructured interviews at week 12 revealed that the participants from both arms were satisfied with the exercise program and faced few or no problems performing the exercises. They also reported that they improved their fitness, their health, and experienced elevated mood.
In the SMS texting arm, 9 participants reported few or no barriers to exercising (50%, 9/18). These participants indicated that the SMS text messages had limited impact and that they would have performed similarly without them. In contrast, 9 SMS texting arm participants (50%, 9/18), experienced a number of personal barriers (eg, laziness/tiredness, lack of motivation) to exercising. Despite these barriers, none of these participants discontinued exercise. They affirmed the value of the SMS text messages, which they described as very important and encouraging. They used words such as “cheering,” “hopeful,” and “inspiring” to express how they perceived the SMS text messages. For example, one participant said, “The text messages gave hope that I can do it.” Four participants perceived the messages as an important push for them to exercise when they felt lazy. Interestingly, 4 participants reported feeling guilty when they received the SMS text messages on those days when they had no intention to exercise. They explained that the SMS text messages made them aware of their commitment and then they scheduled their exercise. Finally, 2 participants also reported that the SMS text messages served as a reminder on busy days.
In the non-SMS texting arm, there were also participants who experienced barriers to exercising (52%, 11/21). Six of these participants did not discontinue exercise, but suggested that an encouraging prompt would have been helpful. The remaining 5 participants exercised very infrequently before they discontinued (after 3 to 8 weeks). Three of them thought that reminders or prompts would have been important to help them continue exercising.
The SMS texting arm participants who experienced few or no barriers to exercising during weeks 1 to 12 reported that they continued to exercise regularly, although some reported less exercise due to various reasons including traveling and busy schedules. Because the SMS text messages were not important to them from the beginning, they did not miss them.
Participants who experienced barriers to exercising during weeks 1 to 12 noted a decrease in barriers and 3 participants reported that the SMS text messages were no longer necessary. However, 2 participants said that it was very difficult for them to continue exercising without the SMS text messages. They reduced their exercise frequency: “I slowed down a little bit without it [text messages] because I did not get reminders.” One participant reported that, without the SMS text messages, she exercised much less because she “did not feel the push and pressure from the text messages.” In the non-SMS texting arm, participants reported few or no changes during weeks 13 to 24 versus weeks 1 to 12.
All SMS text messages were delivered to the SMS texting arm participants as scheduled, without technical problems. Participants who experienced barriers to exercising in the SMS texting arm (50%, 9/18) read all 60 SMS text messages and one participant saved them as well. In comparison, most participants who experienced few or no barriers to exercising ignored the SMS text messages after some time (78%, 7/9). The content of the SMS text messages was perceived as positive. Participants liked the encouragement the SMS text messages provided. One participant said that he “felt that his efforts were appreciated.” Thirteen participants thought that the SMS text messaging frequency was too high (5 text messages per week), whereas 5 participants were in favor of the frequency. Some participants suggested sending more SMS text messages during the initial weeks and reducing the SMS text message frequency over time.
This is the first RCT investigating a mHealth approach to promote exercise in older adults from a non-HIC. From the results, participants who received 60 encouraging SMS text messages over 12 weeks exercised significantly more than participants who did not receive such SMS text messages (mean difference 1.2 times per week). The SMS text messages were perceived as positive encouragement, especially for participants who experienced a number of barriers to exercising. Exercise frequency decreased significantly in the SMS texting arm when SMS text messages ceased. These findings suggest that SMS text messages have a strong impact on exercise participation in older adults, but the effect does not seem to be sustainable once they are removed.
In accordance with previous studies, we found that our SMS text messages had a marked short-term effect on exercise [
Additionally, the call for designing behavioral mHealth interventions around effective BCTs has recently increased [
Finally, our interview analysis revealed that the SMS text messages had a particularly strong impact on participants who experienced a number of barriers to exercise. This is an interesting finding that might explain why the SMS texting arm participants, on average, exercised more compared to the non-SMS texting arm participants. In each research arm, an equal proportion of participants experienced exercise barriers. However, participants who received SMS text messages continued to exercise, whereas a number of participants who did not receive SMS text messages discontinued exercise after some time. Researchers should examine the impact of SMS text messages in older adults who face exercise barriers.
Examining the long-term effects of a behavioral health intervention beyond its conclusion is important for research translation [
It is possible that the SMS text messages were not sent long enough to stabilize the acquired exercise routine [
Finally, Fjeldsoe et al [
We did not observe any significant changes on secondary outcomes throughout the course of the study. That exercise self-efficacy did not change in the SMS texting arm compared to the non-SMS texting arm was particularly surprising considering that the BCTs incorporated in the SMS text message content were supposed to promote exercise self-efficacy [
This study was limited by a lack of statistical power and the small sample size. Although our sample size calculation was based on the available literature [
A major strength of the current study was the investigation of behavioral change maintenance in older adults after the SMS text messages were removed; thereby, we filled an important gap in the evidence [
One of the great potentials of mHealth is that it can reach those most in need of health interventions, including people in non-HICs [
Consort-EHealth checklist V1.6.1.
Screenshots of the online tool used to send SMS text messages to SMS texting arm participants.
Questions from the semistructured interviews (focus on the SMS text messages).
Per protocol analysis.
bias-corrected and accelerated bootstrap
behavior change technique
Exercise Self-Efficacy Scale
high-income country
International Physical Activity Questionnaire
Metabolic Equivalent of Task
Malaysian Physical Activity for Health Study
noncommunicable disease
physical activity
randomized controlled trial
We wish to thank Mandy Janich for designing the myPAtHS exercise booklet and Pang Boon Yuen (Wolftek Technology Sdn Bhd) for the development of the online SMS texting tool. We also wish to express our appreciation to the study participants and the representatives of the residents’ associations and religious facilities for their support. This work was supported by the University of Malaya/Ministry of Higher Education (UM/MOHE) High Impact Research Grant (UM.C/625/1/HIR/MOHE/ASH/02). The grant giver had no role in designing the experiment, collecting the data, and preparing the manuscript.
AMM and SK conceived the study, participated in its design and coordination, and drafted the manuscript. TM helped in designing the study, analyzing the data, and drafting the manuscript. All authors read and approved the final manuscript.
None declared.