The World Health Organization (WHO) declared a global obesity epidemic in 1998, but to date, progress in reversing or even halting increases in prevalence has failed [1]. There is some evidence that the rise in childhood obesity has plateaued in countries such as the United States and Australia [2]. However, young adulthood is an important population group that has been largely neglected and their steep trajectory of weight gain is mostly unrecognized [3,4]. This is of concern as incident obesity at a younger age carries increased risk of mortality from and morbidities of cardiovascular disease, type 2 diabetes, some cancers, and osteoarthritis among others [5-8]. Failure to address the weight gain of young adults may limit the success of childhood prevention programs.

Systematic literature reviews in recent years have drawn attention to the limited evidence base for successful interventions in 18- to 35-year-olds [9]. Many studies have had small numbers of subjects but have shown no potential for translation and scale-up to the community [10]. Rather, the use of young adults as subjects has been coincidental as college-based researchers find it easy to recruit on campus [11].

The life cycle phase termed “emerging adulthood” (18-24 years) signifies the transition from adolescence to leaving school, going to college or finding a job, and increasing independence [12]. It has been identified that this might be a window of opportunity to improve health behaviors as they become more receptive as the rebellion of teenage years is left behind [13]. Young adults need healthy lifestyles to both avoid obesity-related diseases in middle age and to protect their future progeny [14-16].

In 2010, the United States acknowledged young adults as a group requiring intervention for the prevention of weight gain with the award of research funding for the seven EARLY studies targeting 18- to 35-year-olds. [17]. Decreases in physical activity and continued high consumption of sugar-sweetened beverages (SSBs) and food prepared outside home are common lifestyle behaviors in young adults across many western nations [18]. With almost universal ownership of mobiles phones, (91% of young adults in the United States and 95% in Australia), this communication channel could be exploited for intervention delivery, referred to as mHealth [19]. Mobile phones have many features that can be used to provide education and counseling, such as text messaging, apps, and Internet access, in addition to the traditional voice call function.

Here we describe the effectiveness of a 9-month randomized controlled trial (RCT) of an mHealth program for 18- to 35-year-olds conducted in Australia, which aimed to improve lifestyle behaviors [20]. We hypothesized that those overweight young adults who received our 12-week “TXT2BFiT” program followed by a 6-month low-dose maintenance phase would gain less weight compared with those who received minimal intervention.

Figure 1 displays the flow of participants through the trial. In all, 1181 attempts of the screener survey were recorded with 547 of these failing to complete screening. An additional 244 failed to meet the inclusion criteria and 138 eligible participants failed to complete the visit to the primary care physician and were not randomized. Two participants were randomized after census date.

One hundred and twenty-five participants were allocated to each arm. After the 12-week program, 110 intervention participants and 104 control participants completed the Web-based survey for assessment of outcomes. At completion of the maintenance phase, 97 intervention participants and 105 control participants completed the final online survey.

Table 1 summarizes the demographic characteristics and lifestyle behaviors of the population at baseline. Participants were mostly in the overweight BMI range, of higher SES, from English-speaking backgrounds, and tertiary educated. Approximately 3 in 5 participants were females and mean WHO-5 was just below middle of the scale of 25, indicating a tendency of poor well-being. Most participants failed to meet the criteria for fruits (2 serves), vegetables (3 serves), and take-out meals (more than 1 per week). Most participants consumed less than 500 mL of SSBs weekly and the mean total daily physical activity was approximately 55 minutes of moderate physical activity on each of the past 7 days; 51.6% met the national recommendation for 30 minutes of physical activity per day (48.8% intervention, 54.4% control). A comparison of those participants retained in the study versus those lost to follow-up after 9 months showed no significant differences by demographic characteristics neither for intervention nor for the control group.

Table 2 presents body weight, BMI, and MET-minutes of physical activity per week for intervention and control groups. After the 12-week program the intervention participants weighed 3.7 kg (95% CI −6.1 to −1.3, P=.003) less than controls and after maintenance, 9 months from baseline, the intervention group weighed 4.3 kg (95% CI −6.9 to −1.8) less than controls (P=.001). The changes in BMI equated to a difference of 0.56 kg/m² (95% CI −1.22 to 0.09, P=.093) after the 12-week program and 0.78 kg/m² (95% CI −1.53 to −0.02, P=.044) at end of maintenance stage. Sensitivity analyses outlined above generated consistent findings for the primary outcome, with the addition of up to 2.4 kg at 3 months and 7.2 kg at 9 months to the imputed values. Although the intervention improved their moderate physical activity by 12 minutes per day more than the controls at 12 weeks, the differences were not significant and disappeared by 9 months. The WHO-5 score showed improvement in both groups with no significant differences between them. The mean increase in both groups was clinically meaningful, with the mean score above the cut point of 13 indicating improved well-being.

Odds ratios comparing the odds for improving intake of fruit, vegetables, SSBs, and take-out meals for intervention and control groups are reported in Table 3. After the 12-week program, the odds that the intervention group improved intake compared to the control group were significantly greater for vegetables (P=.006), SSBs (P=.024), and take-out foods (P=.013). At the conclusion of maintenance stage, the intervention group had greater odds of maintaining improvements in all 4 diet variables. At the end of the program and the end of maintenance, the intervention group had greater odds of meeting suggested intakes for all dietary variables (Table 4).

Monitoring of intervention delivery showed that 92% of coaching calls were completed during the 12-week program and 81% during 6-month maintenance period. All emails were delivered and for the text messages, 98% were delivered during the first 12 weeks and 96% during the 6-month maintenance. During the 12-week program, 53% of participants replied to more than half the text messages but only 40% replied during the maintenance phase. All participants reported using the text messages at 12 weeks and 60% at 9 months. Fifty-two percent of participants downloaded the physical activity app, 43% downloaded the fruit and vegetable app, and 31% the beverages app [30]. The data on downloads of the take-out app could not be determined. The company hosting the website made changes that lead to a temporary loss of data from this app but it did not affect the other three.

This is one of the first trials of a multicomponent mHealth program for delivery to young adults with demonstrated maintenance of weight management and nutrition-related behavior change after the 12-week program. As hypothesized, the TXT2BFiT program prevented weight gain [25] leading to weight loss that can be maintained after 6 months follow-up with minimal support. The use of mHealth to deliver effective lifestyle health promotion shows promise to halt the rising incidence of obesity during young adulthood—a group recognized as difficult to reach. The ubiquitous use of mobile phones by this age group allows for a number of communication channels to be used to deliver multicomponent programs.

The 12-week TXT2BFiT program led to additional positive health benefits through maintenance of improved diet behaviors at 9 months from baseline. This is the likely explanation for the 6-month maintenance of weight loss. In addition, increasing fruit and vegetable intake has benefits beyond weight management, in protection against cardiovascular disease and stroke and certain cancers [31-33]. In the past decade, consumption of SSBs has been associated with not only weight gain but also cardiovascular disease and type 2 diabetes [34]. The age group targeted here is the largest consumer of fast food meals among adults [35]. Such food is inevitably high in deleterious nutrients such as saturated fat and sodium, and higher frequency of intake is associated with type 2 diabetes in Australian young adults [36]. The intervention failed to produce increases in physical activity compared with control. Both groups appeared to have higher amounts of physical activity at 3 and 9 months but with wide variation. In previous research in a similar group of younger adults, 75% reported at least 150 minutes of moderate physical activity per week but most wanted to increase their physical activity [37]. This is higher than that in the current trial, with 51.6% of participants meeting national physical activity recommendations. The controls were sufficiently motivated to enroll in the study and likely had the capability and opportunity to increase their physical activity without intervention input. It is also of note that the controls did not gain weight during the 12-week program or at 9 months.

A number of other trials using new technologies have been conducted in young adults in the United States [17]. The CITY study compared treatment using 6 face-to-face group sessions followed by monthly phone calls for 24 months (PC), with both an intervention delivered via a mobile phone app and a control group [38]. Weight measurements in the 18- to 35-year-olds indicated a significant loss by the PC group compared with control and app group at 6 months but no difference between app and control groups. By 12 months, all differences disappeared [38].

This study was different from other studies in that the current intervention included 5 short coaching calls in the 3-month intensive phase as a component of the mHealth intervention. In our former pilot RCT to assess the feasibility of delivering the intervention in 50 young adults, we found no difference in weight loss between groups at 12 weeks because both reduced their weight [39]. The extra communication component in this study was the addition of short coaching calls. This allowed more personalized feedback for goal setting and review whereby their performance against a recommended behavior such as 2 daily serves of fruit was used to set goals to work toward the achievement of the target intake. It adds an additional cost to the program that amounted to approximately AU $45 per participant and cost-effectiveness comparisons with totally electronically delivered and traditional face-to-face intervention warrant further research. However, as discussed below, few electronic or entirely app-based interventions demonstrate effectiveness.

Other studies in young adults have used Web-based or email programs. Kattelmann et al delivered a 10-week Web-based intervention to 1639 US college students addressing healthy eating, physical activity, stress, and weight management in an RCT, and assessed post-intervention effects and 12-month maintenance. While both diet and physical activity behaviors improved, no changes in weight occurred and the effects were not maintained 12 months later [40]. Schweitzer et al delivered an adaptation of A Lifestyle Intervention via Email, the ALIVE program, for 24 weeks in a pilot RCT to 148 college students aged 18-20 years. While no differences in body weight were found, the intervention participants increased their intake of fruits as a snack and marginally decreased the percentage energy from saturated fat [41]. Park et al conducted a Web-based RCT in 160 US students aged 18-24 years comparing tailored advice based on the transtheoretical model of behavior change with non-tailored advice and found no differences in fruit and vegetable consumption [42]. This could suggest that using multiple components in mHealth such as text messaging and coaching calls provide a more personalized approach than Web-based techniques that young adults find helpful for changing their behaviors. The college-based RCT by Gow et al with four treatments including no intervention, 6 weeks of a Web-based intervention, 6 weeks of feedback on weight and calorie intake, and a combination of Web-based intervention with weight and calorie feedback found that the combined group attained the lowest BMI [43]. This further illustrates the advantages of feedback communication in addition to Web resources. Bertz et al studied 167 first-year US college students providing Wi-Fi scales and emailed graphs of weight to show changes compared with no feedback control group. It was found that regular weighing with feedback prevented weight gain [44]. Thus, the importance of building education and counseling with feedback along with ongoing monitoring and self-monitoring into a mHealth program is apparent. The inclusion of a range of demonstrated theory-based behavior change techniques, as in this study, also should be central to any mHealth program [45].

The most successful Web-based intervention to date was conducted in Scotland. Nikolaou et al randomized 20,975 university students to a 40-week three-arm RCT: control and two treatments [46]. While the control group gained weight (mean 2.0 kg, 95% CI 1.5-2.3), both treatment groups lost weight: treatment one −1.0 kg (95% CI −1.3 to −0.5) and treatment two −1.35 kg (95% CI −1.4 to −0.7). Both interventions were novel in their approach. The first treatment was based on the “rational” model that individuals when presented with information will make the best use of it. The messages overtly addressed the problem of weight gain and obesity. The second intervention was by “stealth” and raised discussion around social and political movements associated with food and health and obesity. For 19 weeks, participants would log into the Web-based modules to be completed weekly. The advantage of this study is that it was embedded within the university learning environment and was advertised as a new course being tested. Many undergraduate students participated after invitation, unlike the other studies that either recruited volunteers from within the college environment or the population at large. While this intervention could likely be adapted for other college students, whether it would be successful when participants must be recruited from the general community of young adults is uncertain.

One of the strengths of this study is that the age range of recruited participants was well distributed between 18 and 35 years. The program reached a greater proportion of males (40%) than usual in weight management studies [10] and included 30% participants who were not born to an English speaking family. However, our results may not be generalizable to all young adults because the study group tended to be of a higher SES and messaging might be country specific.