Efficacy of applet-based ACT focused on repetitive negative thinking for adults with comorbid anxiety and depression: A randomized controlled trial with mediation analysis

TITLE

Efficacy of applet-based acceptance and commitment therapy focused on repetitive negative thinking for adults with comorbid anxiety and depression: A randomized controlled trial with mediation analysis

COMPLETE REFERENCE

Cao, Q., Du, X., Ruiz, F. J., Sierra, M. A., Yu, X., & Ren, Z. (2025). Efficacy of applet-based acceptance and commitment therapy focused on repetitive negative thinking for adults with comorbid anxiety and depression: A randomized controlled trial with mediation analysis. General Hospital Psychiatry, 97, 56-63. https://doi.org/10.1016/j.genhosppsych.2025.09.006

STUDY TYPE

A 14-day randomized controlled trial (RCT) with a two-arm design: intervention group (N = 49) vs. wait-list control group (N = 45). Mediation analysis with assessments at baseline, mid-intervention, post-intervention, and one-month follow-up.

CONTEXT AND OBJECTIVES

Background

Anxiety and depression are the most prevalent mental disorders worldwide, affecting approximately 3.6% and 4.4% of the global population, respectively. Research indicates that these mental disorders frequently co-occur, with over 70% of individuals with depression also experiencing anxiety. This comorbidity complicates symptomatology, increases treatment challenges, and is strongly linked to higher risks of suicide, functional impairment, and diminished quality of life.

Repetitive Negative Thinking (RNT) is recognized as a transdiagnostic factor in both anxiety and depression. It is characterized by persistent and recurring patterns of negative cognition that focus on negative past events or future worries, thereby exacerbating emotional distress and increasing susceptibility to the vicious cycle of anxiety and depression. RNT serves as a core mechanism influencing mood disorders, and understanding its role can facilitate the development of more effective intervention strategies.

Acceptance and Commitment Therapy (ACT) is widely regarded as a transdiagnostic intervention due to its emphasis on enhancing psychological flexibility, a critical process across numerous psychological disorders. ACT utilizes core processes such as acceptance, cognitive defusion, present-moment awareness, self-as-context, values clarification, and committed action to foster psychological flexibility and decrease psychological inflexibility. Its transdiagnostic application focuses not only on symptom reduction but also on enhancing overall mental health and quality of life.

Ruiz et al. (2016) developed a therapeutic framework called RNT-focused ACT, which suggests that RNT usually becomes the predominant form of experiential avoidance. A series of studies have demonstrated its efficacy in treating comorbid generalized anxiety disorder (GAD) and depression. Sierra and Ruiz (2022) developed an online psychological intervention called "Disentangled, Aware, and Committed." However, several limitations remain: the mediating role of RNT in the intervention process remains insufficiently explored, traditional in-person and web-based interventions face accessibility issues, and the applicability of RNT-ACT in different cultural contexts has not been fully validated.

Research Objectives

This study developed a mobile intervention applet based on RNT-focused ACT, aiming to: (1) evaluate its effectiveness for individuals with the co-occurrence of elevated anxiety and depressive symptoms through a randomized controlled trial (N = 94; intervention group n = 49, wait-list group n = 45), and (2) examine the mediating role of RNT in the applet's impact on improving symptoms of anxiety and depression.

The hypotheses were:

H1: The RNT-ACT applet will effectively reduce levels of RNT in individuals with the co-occurrence of elevated anxiety and depressive symptoms.
H2: The RNT-ACT applet will effectively reduce symptoms of anxiety and depression in individuals with the co-occurrence of elevated anxiety and depressive symptoms.
H3: The RNT-ACT applet will reduce symptoms of anxiety and depression by lowering levels of RNT in individuals with the co-occurrence of elevated anxiety and depressive symptoms.

METHOD

Study Design

A two-arm randomized controlled trial with a 2 (Group: Intervention, Wait-list Control) × 3 (Time: Pre-intervention, Mid-intervention, Post-intervention) mixed design. The study protocol was preregistered on the Open Science Framework and received approval from the Ethics Committee of Life Sciences at Central China Normal University (Approval Number: CCNU-IRB-20231203b). The study was conducted and reported following the CONSORT checklists.

A rigorous longitudinal approach for mediation analysis was adopted, with assessments of outcome variables (RNT, depressive symptoms, anxiety symptoms) at pre-treatment, mid-treatment, post-treatment, and one-month follow-up (for the intervention group only). The wait-list group began the intervention after completing the post-treatment assessments.

Participants

Inclusion Criteria:

Age 18 years or older
Fluent in Chinese and able to use a smartphone without difficulty
Currently experiencing the co-occurrence of elevated anxiety and depressive symptoms, with GAD-7 ≥ 10 and PHQ-9 scores ≥ 10

This criterion was selected to focus on individuals with comorbid symptoms, which is a common clinical presentation and aligns with the transdiagnostic nature of the intervention.

Exclusion Criteria:

Unwillingness to participate
Self-reported suicide attempts or frequent suicidal ideation (PHQ-9 item 9 score ≥ 2)
Currently receiving psychological counseling
Participating in other intervention studies or clinical trials
Clinical diagnoses that affect program use, such as schizophrenia, schizoaffective disorder, bipolar disorder, and substance addiction

Sample Demographic Characteristics (N = 94; Completers = 89):

Intervention group (n = 49): M age = 23.27 years (SD = 5.55); 44.9% male, 55.1% female
Wait-list control group (n = 45): M age = 24.42 years (SD = 6.36); 46.7% male, 53.3% female
No significant differences in gender (χ²(1) = 0.03, p = 0.863)
Significant baseline differences in PTQ (F = 5.181, p = 0.023), PTQ-C (F = 5.525, p = 0.019), and PTQ-M (F = 7.345, p = 0.007)
Marginal differences in GAD-7 (F = 3.823, p = 0.051)

Intervention

Development: The RNT-ACT applet was developed and expanded from Ruiz et al. (2016) and Sierra and Ruiz (2022). It is based on the conceptualization of psychological inflexibility as the core of the ACT pathological model, where experiential avoidance is a typical manifestation leading to thought suppression, worry, and rumination. From a functional analysis perspective, RNT is considered a strategy for experiential avoidance.

14-Day Program: Participants completed 8 modules. On the first day, modules 1 and 2 were unlocked, with a new module unlocked every two days thereafter, until module 8 was unlocked on day 13.

Module Themes:

Introduction
Identifying and understanding RNT
Recognizing value-driven actions and making choices
Further exploring RNT
Developing self-as-context and discrimination of RNT
Life direction
Committed action
Review

Intervention Strategies include: psychoeducation, mindfulness and defusion exercises, metaphors, visualization exercises, and character story prototypes. Metaphors such as "leaves on a stream" and "the bus" are used. Character story prototypes involve an office worker, a single mother, and a student, providing their ideal life goals, current state, negative thinking patterns, and typical behaviors to help participants understand and identify their own RNT patterns.

Wait-list Control Group: During the study, participants in this group did not receive any intervention but underwent the same questionnaire assessments at baseline, mid-point, and post-intervention as the intervention group. After the study concluded, the wait-list control group began the intervention, following the same protocol as the intervention group.

Administration Procedures: Researchers sent access to the WeChat applet to participants in the intervention group, instructing them to complete the introductory and Unit 1 content within two days. Participants could log in to the applet via WeChat. Subsequently, reminders were sent every 48 hours to complete the next unit. On the 7th day, both the intervention and wait-list control groups received the mid-point assessment (T2). The intervention continued with regular reminders. On the 14th day, at the end of the intervention, both groups received the post-intervention assessment (T3). The wait-list control group began the intervention after completing the post-intervention assessments.

Outcome Measures

Primary Outcome: Anxiety

Generalized Anxiety Disorder-7 (GAD-7): Used to assess individual anxiety symptoms. Consists of 7 items (e.g., "Feeling nervous, anxious, or on edge"). Each item is scored on a 4-point scale, with 1 indicating "not at all" and 4 indicating "nearly every day." The total score ranges from 4 to 28, with higher scores indicating more severe anxiety symptoms. In this study, Cronbach's α = 0.891.

Primary Outcome: Depression

Patient Health Questionnaire-9 (PHQ-9): Used to measure individual depressive symptoms. Consists of 9 items (e.g., "Little interest or pleasure in doing things"). Each item is scored on a 4-point scale, with 1 indicating "not at all" and 4 indicating "nearly every day." The total score ranges from 4 to 36, with higher scores indicating more severe depressive symptoms. In this study, Cronbach's α = 0.864.

Secondary Outcome: Repetitive Negative Thinking

Perseverative Thinking Questionnaire (PTQ): Developed by Ehring et al. and revised by Yao et al. for the Chinese population. Consists of 15 items covering three dimensions: core characteristics of repetitive negative thinking (PTQ-C), perceived unproductiveness (PTQ-P), and mental capacity (PTQ-M). Each item is scored on a 5-point scale, with 1 indicating "never" and 5 indicating "almost always." The total score ranges from 15 to 75, with higher scores indicating more severe levels of repetitive negative thinking. In this study, Cronbach's α = 0.928.

Data Analysis

Sample Size Calculation: The sample size was calculated using G*Power 3.1.9.7. Based on an expected between-group effect size of f = 0.25 and a significance level of α = 0.01, 66 participants were needed to achieve a statistical power of (1 - β) = 0.80. Considering an anticipated dropout rate of 20% for psychological interventions, at least 80 participants were required, with 40 in each group.

Effect Analysis: The intent-to-treat (ITT) principle was employed, and all participants were included in the analysis. Missing data due to attrition were addressed using multiple imputation (MI). Normality tests indicated that the dependent variables did not follow a normal distribution (Shapiro-Wilk test, ps < 0.005). Therefore, a generalized linear mixed model (GLMM), which accommodates non-normal distributions of outcome variables, was utilized to evaluate the intervention effects of RNT-ACT. Specifically, participants were set as random effects, while group (intervention, wait-list control), time (T1, T2, T3), and their interaction were fixed effects. Given the significant baseline differences in PTQ and marginal differences in GAD-7 between groups, baseline PTQ and GAD-7 scores were included as covariates. Post-intervention scores of PTQ, GAD-7, and PHQ-9 were examined separately as outcome variables to assess the intervention effect.

Mediation Analysis: The PROCESS 4.2 plugin in SPSS was utilized to conduct the mediation analysis. To adequately demonstrate the role and pathway of a mediator, it is crucial that the mediator precedes changes in the outcome. Therefore, this study considered changes in RNT as the mediator. Specifically, the difference in PTQ scores from baseline (T1) to mid-intervention (T2), denoted as ΔPTQ1 (T2-T1), and from baseline (T1) to post-intervention (T3), denoted as ΔPTQ2 (T3-T1), were used as mediator values. The impacts of changes in PTQ from baseline to midpoint (ΔPTQ1) and from baseline to post-intervention (ΔPTQ2) on the intervention's effect at post-intervention were examined. The group was set as the independent variable, with the intervention group coded as 1 and the wait-list control group coded as 0. Post-intervention scores of GAD-7 and PHQ-9 were used as dependent variables. Baseline PTQ and GAD-7 scores were included as covariates. The significance of the mediation effect was tested using the Bootstrap method. A 95% confidence interval for the indirect effect that did not include zero indicates the presence of a mediation effect.

RESULTS

Sample Characteristics

A total of 94 participants met the criteria and agreed to participate in the study, with 49 assigned to the intervention group and 45 to the wait-list control group. Of these, 89 completed the study, comprising 45 in the intervention group and 44 in the control group, resulting in a dropout rate of 5.32%. In the intervention group, 48 participants (97.96%) completed the first four units, 46 (93.88%) completed the first five units, and 45 (91.84%) completed all intervention modules. No adverse reactions were reported by the participants.

Intervention Effects Analysis

Results showed that when post-intervention PTQ (F(2, 272) = 45.44, p < 0.001), GAD-7 (F(2, 174) = 25.43, p < 0.001), and PHQ-9 (F(2,103) = 12.17, p < 0.001) were the outcome variables, the interaction effect between group and time was significant.

Changes in Scores (Mean ± SD):

Repetitive Negative Thinking (PTQ):

Intervention group: 47.31 ± 11.67 (baseline) → 38.09 ± 9.58 (mid) → 35.13 ± 9.69 (post) → 34.04 ± 11.08 (1-month follow-up)
Control group: 41.87 ± 9.27 (baseline) → 40.84 ± 10.75 (mid) → 37.97 ± 10.88 (post)
Cohen's d (pre vs. post) = 1.136; (post vs. follow-up) = 0.104

Anxiety (GAD-7):

Intervention group: 15.88 ± 5.21 (baseline) → 12.83 ± 3.54 (mid) → 11.82 ± 3.78 (post) → 11.30 ± 3.68 (1-month follow-up)
Control group: 13.73 ± 3.28 (baseline) → 13.29 ± 3.50 (mid) → 13.16 ± 3.52 (post)
Cohen's d (pre vs. post) = 0.892; (post vs. follow-up) = 0.139

Depression (PHQ-9):

Intervention group: 18.37 ± 5.31 (baseline) → 16.24 ± 4.01 (mid) → 13.88 ± 3.19 (post) → 13.28 ± 3.68 (1-month follow-up)
Control group: 16.60 ± 3.58 (baseline) → 15.87 ± 3.95 (mid) → 15.80 ± 3.77 (post)
Cohen's d (pre vs. post) = 1.025; (post vs. follow-up) = 0.174

In the intervention group, the effect sizes for PTQ, GAD-7, and PHQ-9 were relatively small at the one-month follow-up compared to the post-intervention, indicating that the intervention effects remained stable over the follow-up period.

The wait-list control group did not receive any intervention during the study, yet there was still a small effect size observed in the reduction of RNT and improvement in anxiety and depression symptoms. This may be attributed to the hope effect.

Analysis of the Mediating Effect of RNT

The mediation analysis results indicate that the reduction in RNT mediates the effect of group and RNT-ACT on anxiety and depression.

Effects on GAD-7 (Anxiety):

Through ΔPTQ1 (mid-intervention change): Indirect effect β = -1.11, 95% CI [-1.88, -0.46]
Through ΔPTQ2 (post-intervention change): Indirect effect β = -2.08, 95% CI [-3.08, -1.17]

Effects on PHQ-9 (Depression):

Through ΔPTQ1 (mid-intervention change): Indirect effect β = -0.98, 95% CI [-1.78, -0.35]
Through ΔPTQ2 (post-intervention change): Indirect effect β = -2.02, 95% CI [-2.99, -1.17]

Analysis by PTQ Dimensions: Further analysis was conducted using the differences in the three dimensions of the PTQ (PTQ-C core characteristics, PTQ-P perceived unproductiveness, PTQ-M mental capacity) as mediating variables between time points. The results indicated significant indirect effects of the intervention on improving anxiety and depression symptoms through these dimensions:

For ΔPTQ1 as the mediator: Significant indirect effects via PTQ-C (95% CI [-1.45, -0.23]; [-1.34, -0.18]), PTQ-P (95% CI [-1.51, -0.18]; [-1.40, -0.16]), and PTQ-M (95% CI [-1.56, -0.25]; [-1.52, -0.14])
For ΔPTQ2 as the mediator: Significant indirect effects via PTQ-C (95% CI [-2.41, -0.75]; [-2.28, -0.71]), PTQ-P (95% CI [-2.33, -0.66]; [-2.27, -0.66]), and PTQ-M (95% CI [-2.45, -0.84]; [-2.49, -0.94])

DISCUSSION AND CONCLUSIONS

Interpretation of Main Findings

The present study provides compelling evidence for the efficacy of Acceptance and Commitment Therapy (ACT) specifically designed to target Repetitive Negative Thinking (RNT) in adults with the co-occurrence of elevated anxiety and depressive symptoms.

Reductions in RNT, Anxiety, and Depression: The RNT-ACT intervention led to significant reductions in RNT but also resulted in substantial improvements in anxiety and depression symptoms, with large effect sizes observed (Cohen's d > 0.8). These effects were sustained over a month post-intervention, underscoring the lasting impact of the treatment. The effect on RNT (d = 1.136) is consistent with the study by Sierra and Ortiz, which demonstrated that RNT-ACT effectively reduced RNT, anxiety, and depression symptoms among Colombian young women. The effect on anxiety (d = 0.892) aligns with the meta-analysis by Haller et al., which demonstrated the efficacy of ACT in treating anxiety. Notably, the impact on depression (d = 1.025) surpassed the results reported in the meta-analysis by Bai et al. regarding ACT's efficacy for depression.

Role of RNT as a Mediator: The mediation analysis revealed that mid-intervention reductions in RNT significantly mediated the improvement of anxiety and depression symptoms post-intervention, confirming the intervention's effectiveness. This pattern held true throughout the intervention period and aligns with the transdiagnostic process theory, which identifies RNT as a key mediator. Core features of RNT include repetitive, intrusive thoughts that are hard to disengage from. The dimensions of perceived unproductiveness and mental capacity further emphasize this dysfunction. The analysis of RNT's components revealed significant mediating effects across all dimensions in the intervention's effect on anxiety and depression symptoms.

Clinical and Theoretical Significance

This study reinforces the conceptualization of RNT as a transdiagnostic mechanism underlying anxiety and depression. By demonstrating that reductions in RNT precede symptom improvement, the findings provide robust empirical support for the theory that RNT serves as a shared developmental and maintenance pathway for these disorders.

ACT has demonstrated efficacy in treating various psychological disorders, including anxiety and depression. In individuals with comorbid anxiety and depression, RNT functions as a crucial maintaining mechanism. Therefore, an ACT intervention aimed at targeting RNT may enhance and extend its therapeutic effects on these conditions. A possible explanation is that reducing RNT can potentially disrupt the cycle of negative thinking, decrease rumination on negative emotions, and improve emotional regulation. This enables individuals to better manage future stressors and challenges. Such a positive feedback loop may foster more lasting cognitive and behavioral changes, reinforcing the therapy's beneficial effects and prolonging the improvement of anxiety and depression symptoms.

Clinically, this study highlights the potential for interventions targeting RNT modification. The mediation effects of perceived uselessness and mental occupation support Ehring and Behar's theory, emphasizing the difficulty of disengaging from abstract, evaluative thoughts. These patterns consume time and energy, preventing engagement in meaningful activities and creating a vicious cycle. This evidence suggests that future interventions should focus on reducing RNT frequency or transforming it to be more specific and verbal, promoting a more functional thinking style.

Additionally, the use of network-based interventions increases accessibility for individuals in remote areas or those with physical limitations, reducing the cost of seeking help. This highlights the potential for more personalized and effective therapeutic strategies, enhancing the reach and efficacy of psychological interventions.

Limitations

This study represents the first attempt to evaluate the impact of a mobile RNT-ACT applet on reducing RNT in individuals with comorbid anxiety and depression. Nevertheless, several limitations warrant consideration:

Limited Mechanism Exploration: The investigation was limited to RNT as a transdiagnostic process, without exploring more complex models and influencing factors. Future research should aim to gather comprehensive participant data to construct detailed mediation and moderation models.
Limited Process Measurements: Process measurements were conducted only at the intervention midpoint, omitting longitudinal developmental information. Subsequent studies could benefit from more granular assessments following each session and employing longitudinal analytical methods, such as latent growth models.
Generalizability Limitations: The current study's focus on individuals with elevated symptoms of both anxiety and depression limits the generalizability of the findings to those with only one type of symptomatology. Although the transdiagnostic nature of RNT suggests potential efficacy for individuals with either elevated anxiety symptoms or elevated depressive symptoms alone, this hypothesis remains to be empirically tested.
Potential Selection Bias: The potential confounding effects of recruiting participants from both clinical and community settings may impact the ecological validity of the findings, necessitating caution in generalization. Future investigations should differentiate between community and clinical populations to examine distinct and shared maintenance mechanisms.
Research Context: Expanding research within clinical environments, in conjunction with pharmacotherapy, would allow for the assessment of the intervention's efficacy as an adjunctive treatment in clinical populations.

Conclusion

This study found that the RNT-ACT applet had significant immediate and one-month follow-up effects on reducing RNT, anxiety, and depression symptoms in adults with the co-occurrence of elevated anxiety and depressive symptoms. RNT served as a mediator in the intervention's impact on anxiety and depression improvement, indicating that RNT-ACT reduces symptoms by lowering RNT levels in this comorbid population.

Significance and contribution

This study contributes to the treatment of comorbid anxiety and depression by demonstrating the efficacy of a mobile application-based acceptance and commitment therapy intervention. The research provides evidence that a brief 14-day intervention can produce clinically meaningful improvements in symptoms of both anxiety and depression. The longitudinal mediation analysis validates the theoretical mechanisms through which the intervention produces change, demonstrating that reductions in patterns of negative thinking precede improvements in emotional symptoms. The findings expand the literature on digital interventions for comorbid emotional disorders and demonstrate the feasibility of mobile application-based interventions for improving access to evidence-based psychological treatments.