Alcohol use disorder (AUD) is a severe medical problem characterized by an inability to control alcohol consumption despite adverse physical, psychological, and social consequences [1-6]. Early alcohol consumption accelerates AUD development, leading to academic decline, injuries, and premature mortality among young adults aged 20‐39 years, accounting for 13.5% of global alcohol-related deaths in this age group [7-15].

AUD remains a significant public health issue, disproportionately affecting ethnic minority young adults compared with the general population [16-18]. In Hong Kong, ethnic minorities constitute about 8.4% of the population [19,20]. However, unlike those with the majority ethnicity, that is, Hong Kong and/or Chinese populations, ethnic minorities encounter various challenges, including stigma, discrimination, acculturation stress, low self-esteem, social integration difficulties, and loneliness, contributing to unhealthy drinking behaviors and AUD [7,21-27]. Previous studies showed that over 50% of ethnic minorities have less access to education and employment than the Chinese population, and about 60% of Chinese respondents believed that ethnic minorities should be limited to low-skilled occupations [28]. Our previous empirical evidence further identifies cultural factors, family influences, hedonistic motives, curiosity, low-risk perception, coping motives, social pressures, and subjective cravings as reasons behind AUD development among young adult ethnic minorities [29].

In this regard, existing research has evaluated several psychosocial interventions for AUD [30-36], including motivational interviewing [30,37-39], integrated family and cognitive behavioral therapy [40], combined motivational enhancement therapy and cognitive behavioral therapy [41], common elements treatment approaches [42], guided self-change [35], and home-based ecologically based family therapy [36]. However, our systematic review and meta-analysis revealed limited efficacy of these approaches, highlighting the need for other, more effective treatments [43]. Consistent with this knowledge gap, our qualitative findings demonstrated that ethnic minority adults with AUD frequently engaged in experiential avoidance, using alcohol to cope with acculturation stress, anxiety, depression, and discrimination. This pattern is further corroborated by our cross-sectional study, which identified discrimination experiences and avoidance coping behaviors as significant contributors to AUD. Together, these maladaptive behaviors underscore the challenge in achieving a sustained behavioral change. Consequently, our findings indicate a need for third-wave therapies that promote acceptance of such negative internal experiences [44-47]. Third-wave therapies are a kind of psychosocial intervention focusing on accepting rather than changing negative thoughts and emotions [48].

Acceptance and commitment therapy (ACT) has emerged as a promising intervention for substance use and mental health disorders, enhancing psychological flexibility [45,49]. Grounded in Relational Frame Theory [48,50,51], ACT enhances conscious behavioral adaptation aligned with personal values while accepting negative thoughts and emotions [52]. Unlike traditional therapies, ACT focuses on behavioral change over symptom reduction, making it particularly suitable for avoidance-driven conditions such as AUD [48,51,53]. Literature review suggests that ACT may be a viable treatment option for ethnic minority young adults with AUD [47,54,55].

However, our review identified methodological limitations in existing studies on the effectiveness of ACT on AUD and substantial challenges in engagement and adherence to ACT [56-58]. First, many studies lacked appropriate control groups and failed to directly assess alcohol-related outcomes (eg, drinking days, drinks per drinking day, heavy drinking days, and abstinence duration). Second, for the adherence issues, participants showed low engagement with value-based activities for guiding committed action toward a meaningful life [56,57]. This is potentially attributed to difficulty in identifying enjoyable, rewarding alcohol-free activities, and overemphasis on internal experiences (motions or thoughts) without sufficient attention to overt behaviors. Finally, ACT does not contain any daily activity for practice, which is a critical challenge for ethnic minorities with AUD to identify pleasurable alternatives to replace drinking, leading to relapse despite quitting attempts. Consequently, these limitations warrant integrating novel strategies into ACT to enhance engagement and adherence.

Behavioral activation (BA) may be a possible method to overcome engagement and adherence problems in ACT. BA is a concept from behaviorism [59,60] and is developed from a concept that behavior can directly affect emotions for better or worse [59]. Hence, individuals can put themselves in a situation that is likely to foster positive emotions. For example, a positive event such as listening to uplifting music can lead to happiness and vice versa. BA is a widely recognized and theoretically sound approach that has been applied to treat depression [61,62]. Its application has been extended to substance use disorders (SUDs) via providing rewarding experiences in daily life, different from substance use, thus achieving abstinence and preventing relapse [63-65]. Evidence suggests that BA is a promising intervention for SUD as a stand-alone treatment or being integrated with other therapeutic modalities. The emphasis on rewarding experiences in BA exactly addresses the low engagement and adherence of ACT due to the failure to identify enjoyable, rewarding, and positively reinforcing activities in the intervention process. Because of the distinctive benefits of BA, it appears to be a treatment component complementary to the shortcomings of ACT, thus enhancing the intervention effectiveness if they both are integrated into the treatment protocol. Also, BA is a brief intervention that can be delivered within 4 weeks, which will not significantly prolong the duration of each intervention session, ensuring the participants’ adherence and engagement. In sum, ACT simply encourages an individual to ACT, which includes Accepting the reactions and being present, Choosing a valued direction, and Take actions, while BA motivates an individual to take real actions. Two studies have integrated BA into ACT to manage different conditions in various populations, including an internet-based BA and acceptance-based treatment for depression [66], and an open trial of a new acceptance-based behavioral treatment for major depression and psychotic features [64]. The results were promising, which supported that the integration is not only theoretically sound but also feasible and, importantly, can translate into observable benefits.

Regarding the mode of intervention delivery, our qualitative findings revealed significant treatment barriers among ethnic minority young adults, in which no participant had accessed treatments for AUD due to their engagement in work. Apart from work, their time and financial constraints also hindered them from seeking hospital-based care [26]. Internet-based interventions may be a viable option to address these challenges by enhancing accessibility [67-69], reducing stigma [70], minimizing work disruption [71], and improving cost-effectiveness [68,72,73]. The ACT’s efficacy for depression further supports this modality [72,74]. However, no studies have evaluated internet-based self-help interventions integrating ACT and BA for these populations. Therefore, this study examined the preliminary effectiveness and feasibility of an internet-based self-help–integrated ACT and BA intervention in improving alcohol use abstinence among ethnic minority young adults with AUD.

The integration of ACT and BA is grounded in principles of functional contextualism [75]. ACT derives from Relational Frame Theory [76,77]. ACT enhances psychological flexibility through the 6 ACT core processes (ie, acceptance, cognitive defusion, being present, self as context, values, and committed action). These processes represent 6 facets of 1 diamond, and the diamond itself is psychological flexibility. Psychological flexibility is defined as the capacity to persist or change one’s behavior to be open and accepting of personal thoughts and feelings, to appreciate and adapt to different situations, and respond to situations that facilitate one’s values and goals [51,76]. Each process contributes to psychological flexibility for AUD treatment as follows: (1) Being present: Individuals with AUD often drink alcohol in response to cravings and emotional triggers. By learning the present moment, they can mindfully notice these triggers and create a space for acceptance rather than avoidance. (2) Acceptance: This involves opening and making room for such painful feelings, sensations, urges, and emotions that trigger drinking. Instead of struggling with these experiences, individuals learn to allow them to come and go, just like cars driving past outside our house, reducing the urge to escape through drinking. (3) Cognitive defusion: Individuals with AUD may be hooked on thoughts like “I need a drink to cope.” Defusion helps them step back and watch thoughts for what they are—nothing more or less than words or pictures. (4) Self-as-context: This process assists individuals in recognizing that they are more than their thoughts and feelings and makes individuals aware of whatever they are thinking, feeling, and sensing. (5) Values: AUD leads individuals away from what is important to them. Clarifying values helps identify what matters to them in their lives. (6) Committed action: This is taking effective actions, guided by our values. For people with AUD, it refers to committing to value-congruent activities despite pain and discomfort.

BA is derived from the behavioral theory of substance use [78]. According to this theory, substance use persists partly because of the limited access to competing rewarding activities and a lack of positive reinforcement from behaviors that do not involve substance use. To tackle AUD by BA, individuals have to identify adaptive alternatives to drinking that provide comparable rewards (eg, exercise for stress relief) without adverse consequences. The integration of ACT and BA provides a synergetic efficacy. ACT encourages patients to change their relationship with triggering factors rather than avoiding them, and creates a metacognitive space to tolerate distress, for example, acceptance of cravings [77], while BA complements ACT by increasing engagement in value-congruent activities that compete with drinking [75] and provides a structured method (eg, scheduling exercises and social events) to modify external behavior, that is, alcohol use [79]. For example, a client who values health and connection might replace bar visits with an alcohol-free gathering. This integrated approach targets both internal and external factors that drive AUD [75]. ACT equips individuals with skills to accept negative emotions, whereas BA provides a concrete strategy to disrupt reinforcement cycles and build healthy habits. Together, they provide a sustained psychological flexibility [75] (Figure 1,1).

This study examined the preliminary effectiveness and feasibility of an internet-based self-help program integrating ACT and BA on alcohol abstinence among Hong Kong ethnic minority young adults with AUD. The primary objective of this study was to evaluate the intervention’s feasibility in terms of various feasibility outcomes under CONSORT (Consolidated Standards of Reporting Trials) 2025, including rates of eligibility, consent, randomization, adherence, retention, completion, missing data, and adverse events. The secondary objective was (1) to evaluate the preliminary effectiveness of this internet-based self-help program integrating ACT and BA on several outcomes, including drinking days, drinks per drinking day, heavy drinking days, cumulative abstinence duration (CAD), alcohol abstinence self-efficacy (AASE), psychological flexibility, readiness to change, and everyday discrimination in this population; and (2) to estimate the effect size of the Acceptance and Commitment Therapy with Behavioral Activation (ACT-BA) intervention on these same outcomes.

Due to limited time and the online mode of self-help intervention, patients were not involved in designing and developing the research question, nor in reporting the findings. However, experts participated in developing and reviewing the intervention protocol.

A single-blinded randomized controlled trial (RCT) was conducted among Hong Kong ethnic minority young adults with AUD from September 15, 2024, to May 15, 2025. This study was used according to the protocol. There were no changes to the trial design, participants, analysis, and outcomes after the study commenced.

This study was conducted in Hong Kong, a special Chinese administrative region located east of the Pearl River Delta. In Hong Kong, approximately 7.5 million people live on a land area of 1104 square kilometers [80], of which about 619,552 (8.4%) are ethnic minorities [19].

The data were analyzed using the IBM SPSS (version 23; IBM Corp). Descriptive statistics were used to summarize the social-demographic characteristics of participants and the feasibility measures, including eligibility rate, consent rate, randomization rate, adherence rate, retention rate, completion rate, percentage of missing data, and adverse events. The baseline and posttest outcome variables were checked for normality assumption and the homogeneity of variance using the Shapiro-Wilk and Levene tests, respectively. Mean and SD were used for continuous variables with a normal distribution, while medians and IQRs were used for non–normal distributions. Comparability of participants between groups was examined by independent samples t test or Mann-Whitney U test for continuous variables, and chi-square test or Fisher exact test for categorical variables. Additionally, paired t test (for normally distributed data) and Wilcoxon signed-rank test (non–normally distributed data) were used to examine within-participant mean differences.

Missing data were handled using intention-to-treat (ITT) analysis, considering all participants enrolled and randomized to either 1 of the 2 groups. Little Missing Completely at Random (MCAR) test was used to determine whether the missingness was completely at random or not. Sensitive analyses were also conducted to compare results from per-protocol (PP) (complete-case data) and ITT analyses.

To compare pre-and posttest outcomes, we used the Generalized Estimating Equations (GEE) under the assumption that data missing were completely at random. Additionally, an independent samples t test was performed as a supplementary analysis to evaluate a pre-post change in outcomes. However, to preserve statistical power and reduce bias under the assumption that data were missing at random, GEE with multiple imputations (MIs) was used due to the presence of substantial missing data (>5% missingness in posttest measures). The pooled estimates of MIs were used to compute the analysis. If the data were missing at random, a weighted GEE analysis was used, and the effect of time × group assignment interaction on outcome variables was determined.

Parameter estimations for signs of instability were calculated using the quasi-likelihood estimation method integrated within the GEE procedure of IBM SPSS. Quasi-Information Criterion values were used to select the appropriate correlation structure. Model convergence was examined across all imputations and found that they converged successfully without warnings or errors. Intervention effects were estimated based on the difference in pre-post difference using the estimation-based approach for multiple comparisons, that is, we emphasized 95% CI and effect sizes rather than relying solely on P values to assess the precision and clinical significance of the intervention effects. The effect size was calculated to assess the clinical significance of the outcomes for between-group comparisons of change scores and was interpreted using Cohen d, as follows: small=0.2, moderate=0.5, and large=0.8. Finally, the findings were reported following the CONSORT 2025 guideline [94].

Regarding the qualitative data, a thematic method of analysis was used. First, all audio recordings were transcribed into English verbatim. A descriptive phenomenological design was chosen to explore the experiences of participants regarding the intervention. Two researchers (GMB and KYH) carefully read transcripts and listened to audio multiple times to understand the participants’ experiences. They independently highlighted the key quotes and labeled them with codes. Any disagreement between coders was resolved through discussion until a consensus was reached. Senior researchers (KYH and YWM) were consulted if a consensus was not reached. Similar codes were grouped into subthemes and eventually themes, and all 13 participants reviewed the findings to ensure that the findings matched with their experiences. Finally, the results were reported following the COREQ (Consolidated Criteria for Reporting Qualitative Research) [95].

Ethical approval was obtained from the Human Subjects Ethics Subcommittee of the Hong Kong Polytechnic University (reference number HSEARS20240712003). The protocol of this study was submitted on August 28, 2024, and was subsequently published on January 15, 2025, on ClinicalTrials.gov with National Clinical Trial number NCT06779006. This record is available online at ClinicalTrials.gov. Informed written consent was obtained from each participant before the intervention. To ensure confidentiality and privacy, all participants’ data were kept confidential, anonymous, and stored on a secure password-protected server; no personally identifiable information was used in the study. Each participant has been compensated for costs to their time and internet usage; that is, each of them received a 300 Hong Kong dollars (approximately US $38.46) gift card with an exchange rate of 1 US $=7.8 Hong Kong dollars. Finally, there were no supplementary materials or images containing identifiable information for participants in the manuscript. They were informed of their right to decline or withdraw from the study at any time.

The study participants’ sociodemographic characteristics are presented in Table 1. Participants had a mean age of 32.43 (SD 2.57) years, with 80% (32/40) female and 87% (35/40) as Filipino; 85% (34/40) were Catholic, 52.5% (21/40) were single, and 85% (34/40) were employed. Of those being employed, 87.5% (35/40) were domestic helpers. Approximately 60% (24/40) attained primary or secondary education, and 32.5% (13/40) cohabited with partners. The median duration of residence in Hong Kong was 5 years, with an IQR of 2‐6. Regarding the AUD severity, 47.5% (19/40) had moderate AUD, 32.5% (13/40) had severe AUD, and 20% (8/40) had mild AUD. Baseline comparisons between groups revealed no significant differences in demographics and the AUD severity.

Independent samples t test was used to compare the baseline outcomes between the 2 groups (Shapiro-Wilk test, P>.05; Levene test, P>.05). There was no significant difference between the intervention and control groups regarding the baseline outcome measures (Table 2).

Figure 2 presents the CONSORT 2025 flow diagram. A total of 148 ethnic minority young adults who drank alcohol were approached from the community. After screening, 38.5% (57/148) met the eligibility criteria. Of them, 85.96% (49/57) consented to join the study. Among these, 81.6% (40/49) were randomly assigned either to the ACT-BA (n=20) or TAU (n=20) group. The retention rate was 82.5% (33/40), with 85% (17/20) in the ACT-BA group and 80% (16/20) in the TAU group. The attrition rate was 15% (3/20) for ACT-BA and 20% (4/20) for TAU, but there was no significant difference in attrition between groups (P=.47). Of those who remained in the study, 30 completed the postintervention test, providing a completion rate of 75% (30/40), with 80% (16/20) in the ACT-BA group and 70% (14/20) in the TAU group (Table 3). Participants in the ACT-BA group completed an average of 4.9 out of 6 sessions (adherence rate: 81.7%). No adverse events were reported at postintervention. The trial was not ended or stopped early. The median session duration was 30 (IQR 35) minutes. The median interval between sessions was 7 (IQR 2) days.

This study had a total of 5.6% missing values. Little’s MCAR test indicated that the pattern of missing data followed the MCAR test (χ^2₁₀=0.000; P≥.99). However, due to substantial missingness (>5% missingness in posttest measures), MIs were used to preserve the statistical power and reduce bias, assuming that data were missing at random. Using SPSS, 10 imputed datasets were generated to calculate the pooled estimates. The results from the ITT analysis were compared with the PP analysis.

Participants showed high satisfaction with the ACT-BA intervention for its guidance to manage drinking, living, thoughts, and emotions. For instance, in the semistructured interview, a participant said, “I find it very helpful for everyday living,” while others expressed its benefits as “I am happy as it helps people overcome drinking,” “a guide on how to live,” and a tool to “learn many things about drinking.” Another participant stated that the mindfulness exercise “go to nature, observe, and relax” reduced her stress. Participants also reported self-efficacy in engaging with the program due to its clear instructions, simple intervention content, perceived health benefits, and easy website navigation. For instance, a participant said, “The website is easy without assistance, I just followed the directions, 1 do not need assistance because I do understand the context. So, I did it on my own,” and “I’m confident in doing it. I just do it by myself.”

Most participants (n=11) expressed that the content was simple, with clear instructions. For instance, a participant described the content as “it’s well explained, easy to understand” and instructions as “it is easy because I tried to follow the instructions.” However, 2 of them noted challenges with unfamiliar terminology, particularly “need to search what words mean,” and task difficulty as “some parts are easy, some are hard.” All participants preferred the online delivery mode due to its flexibility and accessibility, and one said, “online is better because you can do it anywhere, anytime.” Despite the majority of participants enjoying the interventions, a few faced challenges related to the content, duration, and access to the intervention. Regarding duration, while most found the duration suitable, 2 participants suggested shortening it and said, “a little bit longer, but the interface is beautiful. Every content is important.” Additionally, one participant expressed difficulty in understanding some of the materials. In the semistructured interview, a participant said, “It’s quite difficult for me to understand some of the content.” Furthermore, 2 participants reported difficulties in accessing the intervention due to poor internet connection and stated that “sometimes there is an error when submitting, and another time there are errors in connection and login.”

As presented in Multimedia Appendix 1, the participants identified 5 themes regarding the benefits of the intervention. Participants reported increased awareness of their health and its consequences. For example, a participant said, “I got sick because of drinking alcohol… we are not getting younger.” Another participant also noted, “I read all the consequences...it helps me realize now what will happen if I keep on drinking...how would it affect me if I didn't control it.” By realizing the consequences, participants applied the BA strategies such as replacing drinking with alcohol-free activities. As reported by one participant, “I concentrate on other activities and don't drink as much as I can. I can do another thing instead of drinking because I know that drinking is not giving me anything more.” Some participants also avoided triggers by “limiting time in the triggering environment.” Besides, some participants learned to accept urges and negative emotions. One participant said, “I go to nature and just think and observe, and the switch can ease my mind, I can relax and reduce some of the stress.” Some also learned cognitive defusion and understood that “drinking too much can bring themselves out of their mind,” and such thinking helped them further detach from their urges. Overall, participants noted that this integrated intervention was a very good program and could be applied to their and others’ lives in the future. For instance, one of the participants said, “Oh, I think my feedback is that your program is very good. It can help other people...Honestly, I was a drinker before. Now, I realize that our body is not healthy because of alcohol. Even though I drink once a week, but this is not good. I suffered for almost 2 years, I got sick because of drinking alcohol.” Despite its usefulness, some participants reported difficulties in managing the withdrawal symptoms, for example, agitation and increased heart and respiratory rates, although they were motivated to quit alcohol use after the intervention.

This study was conducted among Hong Kong ethnic minority young adults with AUD, a population that often encounters barriers to health care access and has low treatment engagement and intervention adherence due to low-risk perception or poor health literacy [29,96]. To address this problem, we designed and evaluated a novel internet-based self-help program integrating ACT and BA among Hong Kong ethnic minority young adults with AUD.

This study provides preliminary evidence that our proposed intervention statistically significantly improved psychological flexibility, with a moderate effect size among ethnic minority young adults with AUD. This finding is consistent with previous studies on ACT for AUD [88,97]. However, our effect size on psychological flexibility is larger than that in previous studies, ranging from small to moderate [98]. The larger effect size might be due to the integration of ACT and BA [64,66]. Particularly, ACT emphasizes enhancing psychological flexibility in which participants learned different skills to allow difficult thoughts, feelings, and sensations, thus reducing experiential avoidance, which might trigger drinking as a coping behavior [50]. BA, as an additional component, further boosted the participants’ psychological flexibility by guiding them to identify a list of alternative value-congruent activities to replace drinking [99].

Our previous cross-sectional study and qualitative interviews identified that experiences of discrimination were a key factor leading to the development and maintenance of AUD among ethnic minority young adults [29]. Our quantitative findings showed a moderate effect on discrimination, revealing its potential efficacy in reducing the mean score of discrimination experiences, although a statistically significant mean difference was not reached between groups at postintervention. The insignificant difference between groups could be that our intervention did not aim to reduce the discrimination encountered by the participants [100]. Instead, the core focus of the intervention was to enhance the participants’ psychological flexibility to cope with the discrimination [50]. Another possible explanation for the insignificant difference is that the sample size of this study was small, resulting in a type 2 error in the analysis [101].

The ACT-BA intervention successfully enhanced participants’ AASE in the intervention group compared with the TAU group, with a large effect size at postintervention. AASE is defined as personal confidence in one’s own ability to abstain from alcohol [102]. Bandura [103] (1977) pointed out that self-efficacy is an important construct in determining whether a person will make attempts and take actions for behavioral changes. Although our intervention did not contain any strategy that could directly improve self-efficacy in alcohol abstinence among our participants, it can be indirectly and gradually built up in our intervention through participants’ successful experience in alcohol reduction when they tried to cut down on drinking [104]. Based on this preliminary finding, we expect that the intervention effect on AASE would be stronger at a later time because of more successful experience in alcohol reduction [105]. This postulation shall be further confirmed by a full-scale RCT with longer follow-up.

In this study, the ACT-BA was demonstrated to have statistically significant effects on some drinking outcomes. Particularly, compared with the TAU group, participants in the intervention group reported an average reduction of 4.12 drinking days per month and 1.15 drinks per drinking day at postintervention, with −0.57 and −1.89 effect sizes, respectively. These reductions are perhaps a reflection of the synergetic efficacy of the integrated intervention on the participants’ psychological flexibility, which emphasized noticing negative and unfavorable feelings without judgment, while adapting alternative behaviors to replace drinking in view of pursuing life goals [106]. Notwithstanding a statistically significant mean difference on the number of drinking days and number of drinks per drinking day, we did not find any statistically significant difference on the number of heavy drinking days and cumulative alcohol abstinence, but still the intervention group showed improvement on those outcomes. The insignificant mean difference for these 2 outcomes was likely because complete abstinence from alcohol could trigger withdrawal symptoms, which were not heavily emphasized in our intervention, and these withdrawal symptoms were hard to overcome by only improving their psychological flexibility and self-efficacy in alcohol abstinence [107,108]. This explanation is also supported by our qualitative findings that some participants in the experimental group reported difficulties in managing withdrawal symptoms, for example, agitation and increased heart and respiratory rates, although they were motivated to quit alcohol after the intervention. In response to this issue, this intervention can be strengthened, on one hand, by increasing the content about the management of withdrawal symptoms and, on the other hand, by referring the participants who may require medications for withdrawal symptoms to health care professionals alongside the intervention [109]. Another possible explanation for the insignificant findings was that reduction in heavy drinking and alcohol abstinence are more distal outcomes, which could only be captured and detected by a longer follow-up, which is 6 months, a widely accepted international standard for evaluation of behavioral changes [110]. Thus, a follow-up duration of 6 months or more is recommended for future full-scale RCTs.

Although no significant difference was observed between groups regarding readiness to change, mainly due to the limited power of the study, the intervention group showed improvement in such outcome at postintervention in ITT analysis (B=2.71). Again, this was likely because our proposed intervention focused on enhancing psychological flexibility and value-driven action skills rather than the shift in stages of readiness to change [77]. Additionally, the nonsignificant group-by-time interaction could be partly attributed to a ceiling effect, as most participants reported a high score of readiness to change at baseline, resulting from the convenience sampling technique [111].

The study revealed that the ACT-BA program was feasible for ethnic minority young adults with AUD, evidenced by a high intervention completion of 80% and a low attrition rate of 15%. This completion rate is higher than those reported in existing studies on various interventions, including ACT for AUD [57,98]. The variation could be attributable to 2 innovations of this study. First, the incorporation of BA strategies into ACT interventions may enhance engagement through value-aligned activity substitution. Second, when compared with the traditional face-to-face method to deliver ACT, our intervention was delivered using an online platform with no fixed schedule, allowing the participants to access the intervention at their convenience throughout the week, minimizing work disruptions [112]. While there was variation in time between sessions (2 median intervals of 7 days; IQR days), ethnic minority young adults confirmed that a 6-week duration was appropriate, consistent with existing literature [56].

Our postintervention semistructured interview revealed an overall high satisfaction among the participants in the experimental group, with all endorsing the intervention as effective in guiding them to reduce alcohol consumption, improve emotional regulation, and enhance life management. The participants particularly highlighted that the intervention has clear instructions, simplified content, and easy website navigation, enhancing their interest in engaging with the intervention. In their interview, most participants described the intervention sessions as “simple, well explained, and easy to understand.” All participants preferred the online delivery format due to its flexibility and accessibility [112]. Importantly, the participants reported various degrees in behavior change in their alcohol use and expressed intent to practice and sustain the learned intervention skills in the long term. These positive responses from the participants provide compelling evidence to support the feasibility of the ACT-BA, apart from the feasibility indicators.

This RCT has several limitations. First, the small sample size may reduce the statistical power, potentially affecting the true effect size of the interventions. Second, the self-help intervention limited direct measurement of the participants’ engagement with exercises, metaphors, and practices of each module. Nevertheless, engagement was assessed indirectly through online tracking of the module completion time and daily exercise logs. Third, the preliminary effectiveness was assessed immediately postintervention. Although psychosocial interventions usually bring behavioral changes over time following the intervention, the absence of follow-up assessment may limit the evaluation of the long-term effect of the ACT-BA. Fourth, this study relied solely on participants’ self-reported data rather than objective assessments, which might introduce bias and potentially lead to an over- or underestimation of the effect size of the outcome variables. Fifth, although the ACT-BA showed significant effects across multiple outcomes, how the effects varied across different ethnic minority groups was not examined. Sixth, although almost all ethnic minority participants were able to speak English, excluding individuals who do not speak English may have introduced selection bias, potentially limiting the generalizability of findings to all ethnic minority groups.

This study addresses the literature gaps by providing evidence on the preliminary effectiveness and feasibility of the ACT-BA on different drinking-related outcomes among ethnic minorities with AUD. Therefore, a key implication for future study was to conduct a fully powered RCT to further examine the effectiveness of such intervention, with longer follow-up to capture the long-term effect on different drinking outcomes and incorporating objective measures to assess the drinking outcomes to minimize the social desirability bias. Given that the intervention can be self-administered in an online platform with only minimal involvement from interventionalists in a flexible schedule for working groups and at a low cost, it is expected that this intervention, if proven to be effective in a large-scale similar RCT, can be further delivered to a large group of ethnic minority young adults in the community.

This study presents an innovative integrated ACT and BA intervention, which has been delivered through an internet-based self-help format for ethnic minority young adults with AUD. Unlike the existing evidence that mainly focuses on a single intervention, either ACT or BA, this study integrated those 2 interventions to enhance improvement in abstinence. The findings contribute to the field by providing preliminary evidence for both the feasibility and the potential efficacy of this novel, self-help, flexible, and low-intensity integrated intervention. Feasibility was supported by the high completion rate and low attrition rate, while preliminary efficacy was reflected in reductions in drinks per drinking day and drinking days, alongside improvements in AASE and psychological flexibility among ethnic minority young adults with AUD. The main implication of this study is that this intervention would serve as a key treatment option for ethnic minorities with AUD who encounter challenges in face-to-face therapy. Given these promising findings, the effectiveness of the ACT-BA should be further confirmed in a fully powered RCT before its implementation for this population group in the community.