Who among us doesn’t spend too much time on our phone? While the impact of mobile phone use is not always positive, these devices do have huge potential in being utilised in digital health interventions for mental health problems.
Anxiety disorders are the most common mental health disorders in the world, with around 1 in 5 young adults experiencing general anxiety disorder (GAD) in the past year (Terlizzi & Villarroel, 2020). Young adulthood is a particularly vulnerable period in life due to ongoing developmental changes (e.g., brain development, life transitions); this is reflected in research, with suggestions that 75% of mental health disorders begin before the age of 24 (Fusar-Poli, 2019). Yet there is often unmet need, with only approximately 1 in 3 young adults who suffer from anxiety accessing healthcare services (Niermann et al., 2021).
Cognitive behavioural therapy (CBT) is a recommended first-line treatment for anxiety (NICE, 2020). However, conventional face-to-face CBT can be costly and the waiting lists are long. Digital interventions may offer a possible solution to this, with a recent menta-analysis demonstrating their efficacy in reducing symptoms of anxiety (read Theophanis and Andie’s Mental Elf blog here). Digital healthcare interventions have gained increased attention in recent years due to their affordability, accessibility and scalability, with text-messaging offering a familiar medium for young adults.
This blog post summarises a study by Mason and colleagues (2025), who conducted a randomised controlled trial (RCT) testing the efficacy and mechanisms of a text message delivered CBT intervention for young adults.
Traditional face-to-face CBT is often expensive and associated with lengthy wait times. Text message CBT may offer a scalable and cost-effective alternative.
Methods
This RCT recruited 102 young adults (ages 18-25 years) residing in the US through social media. Eligibility included a score of at least 10 on the Generalised Anxiety Disorder – 7 scale (GAD-7; Spitzer et al., 2006).
Participants were randomly allocated to either receive the CBT treatment via text (guided by core CBT mechanisms, such as explaining the worry process, sessions on cognitive distortions and behavioural activation) or to be in the waitlist control condition. Those in the waitlist condition were eligible to receive the treatment at the end of the 3-month study. The text intervention comprised of fully automated, tailored responses based on GAD-7 scores or participant responses. The participants received a total of 350 texts every other day over a 64-day period, averaging 10.9 texts per treatment day.
The primary outcome measure was GAD symptom severity, with secondary outcomes including behavioural activation, perseverative thinking, cognitive distortion, and intervention acceptability and engagement. Assessments were completed at baseline, as well as 1-, 2- and 3-months post-intervention.
Results
Of the 102 young adults, the mean age was 21.9 years (SD = 2.3) and the majority were female (85.3%). Most of the sample were White (37.3%), yet there was some ethnic diversity. The follow-up rate for the intervention was 99%, with all but one participant completing the 3-month assessment.
Primary outcome
At the beginning of the study, both groups reported severe levels of generalised anxiety (>15/21). While both groups experienced reductions in anxiety symptoms across timepoints, the treatment group reduced to mild symptom severity at 1-month follow-up, which continued to reduce at 2-months and was maintained at 3-months. In contrast, the waitlist group saw smaller reductions towards moderate symptom severity levels at 1-month follow-up and maintained this through to 3-months follow-up. The treatment group produced a large effect size from baseline to 3-months follow-up (d = 0.83), meaning that the group receiving text message CBT had a much larger reduction in symptoms compared to the control group. Furthermore, 25% of those in the treatment group reported no or minimal anxiety symptoms at the end of the treatment, meaning they experienced a clinically significant reduction in anxiety and a move towards high-end functioning. In comparison, only 5.5% of those in the control group reached this level.
Secondary outcomes
Mediation analyses were used to understand the mechanisms through which the CBT text treatment affected symptoms of anxiety. Findings showed that the intervention operated through an increase in behavioural activation, and reductions in perseverative thinking and cognitive distortions. The three proposed mechanisms accounted for 52% of the treatment effect of the CBT texts on the reduction of anxiety symptoms; perseverative thinking accounted for the highest percentage of the effect (25.3%), followed by behavioural activation (17.3%) and cognitive distortions (9.3%).
Finally, intervention acceptability and engagement were measured through participant reported satisfaction measures on a 5-point Likert scale (1 = strong disagreement, 5 = strong agreement). Overall, both acceptability and engagement were high. Nearly all participants (90.4%) found the texts easy to understand and complete, and the majority (84%) completed 80% of the texts. Participants found the treatment helpful (M = 3.9, SD = 0.8) and reported moderate levels of being able to implement the skills learned for their anxiety (M = 3.3, SD = 0.90).
Young adults receiving text message CBT for anxiety showed reductions in generalised anxiety symptoms from severe to mild across the 3-month study period.
Conclusions
This study by Mason et al. (2025) demonstrates promising results in the use of a text message CBT treatment for generalised anxiety symptoms in young adults. The treatment showed high levels of acceptability and engagement, with minimal drop-out. The reductions in anxiety symptoms operated through the three hypothesised mechanisms of perseverative thinking, behavioural activation and cognitive distortions, indicating that the treatment appears to work through multiple overlapping mechanisms of cognitive and behavioural change working together.
Overall, the authors provide further evidence and mechanistic results for text message CBT for young adults; to read more about their findings in the context of depression, read Becky’s blog.
This study by Mason et al. (2025) suggests that text message CBT for generalised anxiety may result in clinically significant change for some, with symptoms reducing from severe to minimal for 25% of the intervention group.
Strengths and limitations
Strengths
- The study used a strong methodological design to address a highly relevant question, with RCTs being considered the gold standard for evaluating interventions and offering more robust evidence.
- The study recruited through social media, which enabled access to a geographically spread and ethnically diverse sample of young adults across 33 US states; a group who may otherwise be difficult to engage through traditional recruitment methods such as clinic referrals.
- Increasingly, online data collection can result in fraudulent responses; the researchers managed this by implementing several fraud prevention measures which were clearly reported, increasing the reliability of the data.
- The study also reported a remarkably low drop-out rate of 1%, which is usually unheard of in research. It is far lower than the average drop-out rate of 22.2% found in a recent meta-analysis of digital treatments for anxiety disorders (Domhardt et al., 2021), indicating higher internal validity and further reinforcing the acceptability and feasibility of the intervention.
Limitations
- With only 102 participants across both arms (49 in the treatment group) this is a small trial, and results should be interpreted with appropriate caution. Small trials tend to produce inflated effect sizes, and the study is underpowered for detecting whether the treatment works differently for different groups. The non-significant individual indirect effects in the combined mediation model may partly reflect this limited sample size.
- The comparator was a waitlist control, meaning the study demonstrates that text message CBT beats receiving nothing, not that it outperforms existing treatments. Waitlist designs are known to produce larger effect sizes than active comparator designs, so the headline figure of d = 0.83 should be read in that context.
- The primary outcome measure, the GAD-7, had poor internal consistency at baseline (Cronbach’s alpha = 0.67, below the conventional threshold of 0.70), meaning the starting measurement of anxiety severity (the thing the whole study hinges on) was unreliable. This improved substantially over time, reaching 0.90 by 3 months, but it is a genuine limitation on the baseline data.
- The sample was predominantly female (85.3%), limiting what we can conclude about how the intervention works for men, who may engage with text-based CBT differently or experience GAD differently.
- The study did not control for several potentially important variables, including medication use, comorbid mental health conditions, and socioeconomic status, any of which may have influenced outcomes. Note that sex and race were checked and found to have no significant relationship with baseline anxiety scores, so their exclusion was deliberate rather than an oversight.
- Participants could access on-demand booster text messages (the “4MOOD” feature) at any time during the trial, but the frequency of use was not reported and not controlled for in the analysis. Those who used this feature more may have experienced greater improvements, making it difficult to attribute outcomes consistently across participants.
- The proposed mechanisms accounted for 52% of the treatment effect, leaving 48% unexplained. While 52% is a meaningful proportion for a mediation study, it also means that for nearly half of the observed improvement, we don’t yet know why the treatment worked.
- Finally, participants were followed up for only 3 months. Since GAD is often chronic or recurrent, longer follow-up is needed to establish whether the benefits are sustained once the texts stop.
Only 1% of participants dropped out from the study, which is practically unheard of , but was this influenced by the payment provided?
Implications for practice
While some participants in the current study reported severe symptoms of anxiety, text message interventions are unlikely to replace face-to-face therapy for those with more complex or high-risk anxiety. However, these remote interventions do offer some advantages over traditional in-person therapy for certain individuals, including young adults for whom access barriers are common. Traditional face-to-face therapy can be costly and have long waiting lists, whereas text message and digital interventions are less expensive (Catarino et al., 2023), can reach more people, don’t require travel, and can be accessed flexibly from anywhere. Further, the anonymity and privacy afforded by text message interventions may be appealing to young adults, particularly if they are experiencing feelings of shame, embarrassment, or are afraid of judgement (Conrad, 2024; Curcio & Corboy, 2020).
However, research into digital interventions for anxiety, specifically in young adults, is still evolving. As much as this study showed promising results, more studies need to be conducted to be able to draw stronger conclusions for clinical practice, including the comparison between face-to-face and text message CBT; the current study only indicated that text message CBT was better than receiving nothing, as opposed to being more effective (and cost-effective) than face-to-face treatments. Nonetheless, this study did show that there is potential in text message CBT, while also showing the mechanisms through which the intervention worked.
Future studies should first seek to replicate these findings in a larger, adequately powered trial before drawing firm conclusions for practice. Beyond replication, future work could address other limitations of the current study, such as including confounding variables, considering the possible dose effect of on-demand booster messages, and following up participants over a longer period of time to see if benefits are sustained. Further studies could also explore what other mechanisms might be at play, since 48% represents a considerable unknown and limits the ability to optimise the intervention going forward. Finally, although the current study’s text messages were automated, there were still members of staff reading every text behind the scenes as a safety measure, so it would be important to understand how this might be implemented and resourced by healthcare services, and to estimate the cost effectiveness of the intervention.
Text message CBT for generalised anxiety in young adults is a promising direction, but this single small trial is best understood as a foundation for further research rather than a basis for practice change.
While it is too early to offer concrete implications for clinical practice, text message CBT interventions offer a lot of potential and are worth looking out for in future research.
Statement of interests
Julia Anna Hernandez Duch has no conflicts of interest to declare.
Edited by
Dr Nina Higson-Sweeney.
Links
Primary paper
Michael J. Mason, J. Douglas Coatsworth, Aaron Brown, Nikola Zaharakis, Michael Russell, & Jeremy Mennis (2025). Treating young adult generalized anxiety disorder with text-message delivered cognitive behavioral therapy: A randomized clinical trial testing efficacy and clinical mechanisms. Journal of Anxiety Disorders, 116, 103089. https://doi.org/10.1016/j.janxdis.2025.103089
Other references
Appleton, B. (2024). SMS CBT TLC CYP? 🤔 Does delivering cognitive behavioural therapy via text messages help improve depression in young people?. The Mental Elf.
Catarino, A., Harper, S., Malcolm, R., Stainthorpe, A., Warren, G., Margoum, M., Hooper, J., Blackwell, A. D., & Welchman, A. E. (2023). Economic evaluation of 27,540 patients with mood and anxiety disorders and the importance of waiting time and clinical effectiveness in mental healthcare. Nature Mental Health, 1(9), 667–678. https://doi.org/10.1038/s44220-023-00106-z
Conrad, J. A. (2024). Digitization and its Discontents: The Promise and Limitations of Digital Mental Health Interventions. Journal of Contemporary Psychotherapy, 54(3), 209–215. https://doi.org/10.1007/s10879-024-09620-2
Curcio, C., & Corboy, D. (2020). Stigma and anxiety disorders: A systematic review. Stigma and Health, 5(2), 125–137. https://doi.org/10.1037/sah0000183
Domhardt, M., Nowak, H., Engler, S., Baumel, A., Grund, S., Mayer, A., Terhorst, Y., & Baumeister, H. (2021). Therapeutic processes in digital interventions for anxiety: A systematic review and meta-analytic structural equation modeling of randomized controlled trials. Clinical Psychology Review, 90, 102084. https://doi.org/10.1016/j.cpr.2021.102084
Fusar-Poli, P. (2019). Integrated Mental Health Services for the Developmental Period (0 to 25 Years): A Critical Review of the Evidence. Frontiers in Psychiatry, 10, 355. https://doi.org/10.3389/fpsyt.2019.00355
Kyriacou, T., & Ashdown, A. (2022). Anxiety disorders, should we be turning to digital interventions?. The Mental Elf.
NICE. (2020). Generalised anxiety disorder and panic disorder in adults: Management (No. CG113). NICE. https://www.nice.org.uk/guidance/cg113?utm_source=chatgpt.com
Niermann, H. C. M., Voss, C., Pieper, L., Venz, J., Ollmann, T. M., & Beesdo-Baum, K. (2021). Anxiety disorders among adolescents and young adults: Prevalence and mental health care service utilization in a regional epidemiological study in Germany. Journal of Anxiety Disorders, 83, 102453. https://doi.org/10.1016/j.janxdis.2021.102453
Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Löwe, B. (2006). A Brief Measure for Assessing Generalized Anxiety Disorder: The GAD-7. Archives of Internal Medicine, 166(10), 1092. https://doi.org/10.1001/archinte.166.10.1092
