Comparative Effects of Acceptance and Commitment Therapy, Cognitive Behavioral Therapy, and Treatment-as-Usual on Depression Among People With Marital Crisis

Study site This study adopts a multicenter design and is carried out under the academic guidance of Universiti Sultan Zainal Abidin (UniSZA) of Terengganu, Malaysia. Ethical approval is first obtained from the UniSZA Human Research Ethics Committee (UHREC), based on the ethical guidelines of UniSZA and the Declaration of Helsinki, before any participant-related activities are initiated. After the UniSZA approval is obtained, ethical clearance is then obtained from the ethics committee of each participating clinical institution in China. The study participants came from two psychiatric secondary and primary referral centers in Guangdong Province, China: Yunfu City Yuncheng District People's Hospital and Zhaoqing City Duanzhou District Shangyitang Comprehensive Clinic. Yunfu City Yuncheng District People's Hospital has about 1,000 registered psychiatric patients receiving treatment; About 800 registered psychiatric patients are receiving treatment at Zhaoqing City Duanzhou District Shangyitang Comprehensive Clinic. These two medical institutions accept referrals from psychiatric patients from Guangdong Province and a few nearby provinces such as Guangxi, Hunan, Hubei, Fujian, Jiangxi, and Hainan。 Sample size The sample size is calculated using G\*Power version 3.1 for a mixed analysis of variance (ANOVA) with three between-subjects groups (ACT, CBT, TAU) and four within-subjects assessment time points (T0, T1, T2, T3), specifying a small-to-medium effect size for the group by time interaction of f = 0.25, a two-tailed alpha of 0.05, a power (1 minus beta) of 0.80, a correlation among repeated measures of 0.50, and a non-sphericity correction of 1. The choice of f = 0.25 is informed by the meta-analytic effect size estimates reported by Cuijpers et al. (2023) for adult CBT for depression and by Zhao et al. (2023) for adult ACT for depression, both of which report medium effect sizes for the between-group differences in depressive symptoms at post-intervention. Under these assumptions, the minimum required sample size is 36 participants per group, totaling 108 participants for the trial. To account for an anticipated attrition rate of approximately 10% across the four assessment time points, the planned recruited sample size is inflated to 40 participants per group, totaling 120 participants for the trial. The expected analyzable sample size at T3 is 108 participants. The assumed attrition rate of approximately 10% is a cumulative estimate across the full follow-up period and is consistent with dropout rates reported in comparable trials of group psychotherapy for depression with follow-up of similar length. Because loss may accrue at each of the four measurement occasions rather than only once, this cumulative figure corresponds to an average per-occasion attrition of roughly 3%; the recruitment target of 40 participants per group is therefore retained as a conservative allowance. In addition, the analysis follows the intention-to-treat principle and uses all available observations, which further reduces the impact of intermittent missing data at individual time points. Recruitment of subjects This study employs simple random sampling as the sampling method. All patients who are registered for outpatient psychiatric treatment at the two study sites and who have a documented diagnosis of depressive disorder are listed and assigned a computer-generated random number. The computer then generates a random sequence for the selection of potential study participants from this list. The selected individuals are subsequently screened by a trained research assistant during a face-to-face assessment. During the face-to-face screening, marital crisis status (DAS total score of 97 or below) and all other inclusion and exclusion criteria are assessed. Only those who meet all inclusion criteria and do not meet any exclusion criteria proceed to the informed consent process and enrollment. The recruitment research assistant is independent of the intervention delivery, the outcome assessment, and the data analysis, and is unaware of the trial hypotheses. The recruitment research assistant holds a bachelor's degree in psychology or a related discipline and receives a two-week structured training prior to the commencement of recruitment, covering the eligibility criteria, the procedures of the trial protocol, the principles of informed consent, the safeguarding of participant confidentiality, and the management of distress that may arise during the screening process. Individuals who are randomly selected from the patient registry are contacted by the trained recruitment research assistant by telephone solely to schedule an appointment for a face-to-face screening interview at the study site within two weeks; no eligibility screening is conducted over the telephone. All eligibility screening is conducted in person during the face-to-face assessment. The face-to-face screening is conducted by the trained recruitment research assistant, not by the research team members. During the face-to-face screening, the diagnostic confirmation of major depressive disorder is established using the Mini International Neuropsychiatric Interview (MINI) administered by a board-certified psychiatrist who is independent of the trial. The BDI-II is administered to confirm a baseline score of 20 or above. The DAS is administered to confirm a baseline total score of 97 or below. After screening, individuals who fulfill all inclusion criteria and do not meet any exclusion criteria are explained in detail about the study by the recruitment research assistant. The recruitment research assistant then offers the individual the opportunity to participate and sign the written informed consent form. Participants are given sufficient time to ask questions and consult family members before providing written informed consent and being enrolled in the trial. Inclusion criteria 1. Meets DSM-5-TR diagnostic criteria for major depressive disorder, confirmed by the Mini International Neuropsychiatric Interview (MINI) structured clinical interview (American Psychiatric Association, 2022). 2. Beck Depression Inventory-II (BDI-II) total score of 20 or above, indicating moderate to severe depressive symptoms (Beck et al., 1996). 3. Dyadic Adjustment Scale (DAS) total score of 97 or below, operationally defining marital crisis (Spanier, 1976). 4. Age between 20 and 60 years. 5. Currently taking a single antidepressant medication for a minimum of two weeks. Rather than being required to remain on an unchanged dose, participants agree to have their antidepressant regimen managed by their treating psychiatrist according to clinical need for the duration of the study; any change in medication or dose is documented and included as a time-varying covariate in the analysis. Exclusion criteria 1. Current diagnosis of symptomatic bipolar mood disorder or psychotic disorder which would interfere with their ability to answer questionnaires and participate in the study. 2. History of substance abuse or acute alcohol use disorder. 3. Prior receipt of any structured psychotherapy before study enrollment. 4. Severe physical illness or cognitive impairment that would prevent meaningful participation in group therapy sessions or completion of assessment instruments. 5. Current exposure to active intimate partner violence or domestic abuse. This is screened at the face-to-face assessment using a brief standardized instrument, and it is applied to safeguard participant safety during the disclosure of marital distress in the group setting. Individuals who screen positive are not enrolled and are referred to appropriate local domestic-violence support services. Randomization Eligible participants who have provided written informed consent are randomized to the three intervention arms (ACT, CBT, TAU) in a 1:1:1 ratio using stratified permuted block randomization. The randomization is performed by another trained research assistant who is not part of the research team and is unaware of the trial hypotheses and objectives. Stratification is conducted on three prognostic variables, namely study site Yunfu City Yuncheng District People's Hospital and Zhaoqing City Duanzhou District Shangyitang Comprehensive Clinic, gender (male, female), and baseline depression severity (moderate, defined as a baseline BDI-II score of 20 to 28, and severe, defined as a baseline BDI-II score of 29 to 63), generating eight strata in total (2 sites x 2 genders x 2 severity levels). The block size is set at 6 within each stratum, with the order of intervention assignments within each block randomly determined, to maintain approximate balance among the three arms while preserving the unpredictability of the next allocation. The randomization sequence is computer-generated using the blockrand function in the R statistical environment by an independent statistician who is not otherwise involved in any aspect of the trial. Allocation concealment is achieved using sequentially numbered, sealed, opaque envelopes prepared by the independent statistician (Schulz et al., 2010). Each envelope contains a card indicating the intervention allocation for the corresponding participant identification number within the stratum. The envelopes are stored in a locked cabinet at the study site, accessible only to the randomization research assistant. Upon completion of the baseline assessment and provision of written informed consent, the randomization research assistant opens the next envelope in the sequence within the corresponding stratum and communicates the allocation to the participant and to the assigned therapist. The envelopes are not opened in advance, and the randomization research assistant does not have access to future allocations in the sequence. Stratified permuted block randomization is expected to reduce selection bias and to enhance statistical power by balancing the distribution of gender and baseline depression severity across the three arms. Assessment time points Outcome variables are assessed across four different points in time, including: T0 (baseline, prior to the start of the intervention), T1 (immediately after completion of the intervention at 8 weeks), T2 (12 weeks after completion of the intervention), and T3 (24 weeks after completion of the intervention). The schedule of tests to be performed at each time point is given in Table 1. Information on assessor blinding, test administration conditions, and data processing for each time point can be found in the Data collection subsection below. Intervention Group 1: ACT combined with a single antidepressant. Participants randomized to the ACT arm receive an eight-week manualized group ACT program, with weekly 90-minute sessions delivered in closed groups of approximately 8 to 10 participants per group. The program is conducted in adherence to the ACT protocol developed by Hayes et al. (2012), targeting the six core processes of psychological flexibility, namely acceptance, cognitive defusion, present-moment awareness, self-as-context, values clarification, and committed action. Each session follows a standard structure of a brief mindfulness opening exercise (10 minutes), review of the previous week's home practice (15 minutes), introduction of the session content through didactic teaching and metaphors (20 minutes), experiential and dyadic exercises in pairs or in the whole group (35 minutes), and a closing summary with the assignment of home practice for the coming week (10 minutes). To consider the program successfully completed, each participant is expected to attend at least 6 of the 8 sessions, with a maximum allowable late arrival of 10 minutes per session. Group 2: CBT combined with a single antidepressant. Participants randomized to the CBT arm receive an eight-week manualized group CBT program, with weekly 90-minute sessions delivered in closed groups of approximately 8 to 10 participants per group, conducted in adherence to established cognitive behavioural therapy for depression adapted for delivery in a group format (Bieling et al., 2022). The CBT manual used in this trial is based on the validated cognitive therapy protocol for depression established by Beck, Rush, Shaw, and Emery (1979), comprising cognitive restructuring and behavioural activation, which remains the standard manualized protocol for cognitive therapy of depression and has been extensively validated in randomized controlled trials. The manual is translated and culturally adapted for the present population through a standardized procedure of forward and backward translation and expert review, and the fidelity of its delivery is verified using the Cognitive Therapy Scale-Revised (CTS-R) as described in the Treatment fidelity subsection. Group-based delivery of cognitive behavioural therapy for depression has itself demonstrated efficacy relative to control conditions across 33 randomized controlled trials in a recent systematic review and meta-analysis (Wong et al., 2024). Each session follows a standard structure of agenda setting and mood check (10 minutes), review of the previous week's home practice and feedback (15 minutes), introduction of the session content through didactic teaching and worked examples (20 minutes), in-session practice and group discussion (35 minutes), and a closing summary with the assignment of home practice for the coming week (10 minutes). To consider the program successfully completed, each participant is expected to attend at least 6 of the 8 sessions, with a maximum allowable late arrival of 10 minutes per session. Group 3: Treatment-as-usual (Active Control). Participants randomized to the TAU group continue the antidepressant regimen prescribed by their treating psychiatrist prior to study enrollment, maintained at the same dose throughout the study period. These participants attend routine psychiatric follow-up visits at the study site approximately once every four weeks, with each visit lasting approximately 15 to 20 minutes. The visits are conducted by board-certified psychiatrists who are not involved in the delivery of the ACT or CBT interventions and are focused on monitoring depressive symptoms, evaluating medication adherence and side effects, and providing supportive but non-directive communication. No structured psychotherapy, no manualized treatment components, and no group sessions are delivered to the TAU group. To ensure participant retention and address ethical equity, TAU participants are offered the opportunity to receive the ACT or CBT intervention free of charge after the completion of the T3 follow-up assessment, if they wish to do so. Therapists All intervention sessions are delivered by licensed clinical psychologists who hold a doctoral or master's degree in clinical psychology and possess a minimum of two years of clinical practice in delivering manualized psychotherapy. Four therapists are recruited for the trial, with two therapists assigned exclusively to the ACT arm and two therapists assigned exclusively to the CBT arm to prevent therapist-level cross-over of treatment content between the two active conditions. The TAU psychiatric follow-up visits are conducted by board-certified psychiatrists who are not involved in the ACT or the CBT intervention delivery. Prior to participant recruitment, the ACT therapists complete a 40-hour standardized workshop on ACT for depression, conducted by a senior clinical psychologist with at least five years of experience in delivering and supervising ACT, covering the six core processes of psychological flexibility, experiential exercises, and case formulation. The CBT therapists complete a 40-hour standardized workshop on CBT for depression, conducted by a senior clinical psychologist with at least five years of experience in delivering and supervising CBT, covering cognitive restructuring, behavioral activation, and case conceptualization. Following the workshop, each therapist conducts three supervised pilot cases over a three-month period before being permitted to deliver the intervention to study participants. During the trial, all therapists attend twice-weekly group supervision meetings led by the senior clinical psychologist of their assigned intervention arm to discuss case progress, address procedural questions, and manage clinical concerns that arise within the groups. Treatment fidelity All ACT and CBT intervention sessions are audiovisually recorded with written informed consent from the participants. To evaluate treatment fidelity, 20% of the recorded sessions are randomly selected for independent rating, with the random selection stratified across the eight sessions of the intervention (covering session 1 (early phase), session 3 (mid-early phase), session 5 (mid-late phase), and session 7 (late phase)) so that each phase of the intervention is represented in the fidelity evaluation. Two independent raters are recruited for the fidelity assessment, both of whom hold a doctoral degree in clinical psychology, have at least five years of experience in delivering and supervising the corresponding intervention modality, and are not involved in any other aspect of the trial. The fidelity of the ACT sessions is rated using the Acceptance and Commitment Therapy Fidelity Measure (ACT-FM), a structured observational rating tool that evaluates therapist adherence to the six core processes of psychological flexibility and the absence of inconsistent therapeutic content (O'Neill et al., 2019). The fidelity of the CBT sessions is rated using the Cognitive Therapy Scale-Revised (CTS-R), a structured observational rating tool that evaluates therapist competence in delivering the core components of CBT for depression, including agenda setting, eliciting feedback, conceptual integration, application of cognitive restructuring and behavioral activation, and homework planning (Blackburn et al., 2001). For both measures, each rater independently rates 50% of the randomly selected recordings, with a subset of 30% of the recordings rated by both raters to allow estimation of inter-rater reliability using the intraclass correlation coefficient (ICC). An ICC of 0.70 or above is considered indicative of acceptable inter-rater reliability and adequate treatment fidelity. Any procedural concerns identified during the fidelity ratings are discussed in the twice-weekly group supervision meetings to ensure consistent delivery of the intervention protocols throughout the trial. Withdrawal criteria A participant is withdrawn from the trial, and the data collected up to the point of withdrawal are excluded from the analysis dataset if it cannot be reasonably assumed that the participant's remaining data will provide valid information about the intervention effects, when one of the following criteria is met. (1) The participant reports clinically significant adverse effects that may or may not be related to the study intervention and that, in the judgment of the trial physician, render continued participation unsafe. (2) The participant reports active suicidal ideation, suicidal plan, intent, behavior, or deliberate self-harm, in which case the participant is immediately referred to crisis intervention services and withdrawn from the trial. (3) The participant initiates additional structured psychotherapy outside the trial during the intervention or follow-up period that, in the judgment of the principal investigator, materially alters the intervention exposure. (4) The participant requests withdrawal from the trial for any reason. (5) The participant is lost to follow-up despite at least three documented attempts to re-establish contact across multiple modalities of communication. Statistical analysis All analyses are conducted using appropriate statistical software, with statistical significance set at a two-tailed alpha of 0.05. The primary analysis follows the intention-to-treat principle and includes all randomized participants. The primary hypotheses are tested using a mixed-model analysis of variance with a between-subjects factor of group (ACT, CBT, TAU), a within-subjects factor of time (T0, T1, T2, T3), and their group × time interaction, consistent with the analytic model on which the sample size calculation is based. Because the ACT and CBT interventions are delivered in therapy groups of approximately eight to ten participants, the outcomes of participants within the same group may not be fully independent and may exhibit a degree of within-group correlation. To address this, the intraclass correlation coefficient (ICC) attributable to therapy group is estimated and reported for the primary outcome. Because each active arm comprises only a small number of therapy groups, the stable estimation of a group-level random effect is limited. Clustering is therefore not modelled as a random effect in the primary analysis. Instead, its influence is evaluated in a sensitivity analysis that refits the model with therapy group as an additional level, using a linear mixed-effects model with a random intercept for therapy group or an equivalent cluster-robust standard-error adjustment. If the estimated ICC is low and the sensitivity analysis yields conclusions consistent with the primary analysis, the primary results are interpreted as robust to within-group clustering; if the ICC is substantial, inferences are based on the clustering-adjusted sensitivity analysis. A per-protocol analysis restricted to participants who complete at least six of the eight sessions is conducted as a supportive analysis, and pairwise between-group and within-group contrasts are examined at each time point with correction for multiple comparisons. Psychometric properties of the measurement instruments All primary and secondary outcomes are measured with instruments that have established reliability and validity and for which validated Chinese-language versions are available, namely the Beck Depression Inventory-II (BDI-II), the Generalized Anxiety Disorder 7-item scale (GAD-7), the Dyadic Adjustment Scale (DAS), the Personal Assessment of Intimacy in Relationships (PAIR), the World Health Organization Quality of Life-BREF (WHOQOL-BREF), and the Acceptance and Action Questionnaire-II (AAQ-II). Prior studies have reported acceptable-to-good internal consistency for these Chinese-language versions. To confirm the psychometric adequacy of each instrument in the present sample, the internal consistency (Cronbach's alpha) of every scale is computed at baseline and reported, with a value of 0.70 or above considered acceptable. Where applicable, test-retest reliability and construct validity from prior validation studies are also cited in the final report.