Efficacy of a Brief ACT Intervention in Reducing Interpersonal Problems and Fostering Interpersonal Flourishing: A Randomized Multiple-Baseline Evaluation

Title

Efficacy of a Brief ACT Intervention in Reducing Interpersonal Problems and Fostering Interpersonal Flourishing: A Randomized Multiple-Baseline Evaluation

Complete Reference

González-Cifuentes, C. E., Ruiz, F. J., Bernal-Manrique, K. N., Hernández-Bernal, X., Otálora, D. M., Peña-Vargas, A., Toquica-Orjuela, D. J., & Pérez-Nieves, K. T. (2025). Efficacy of a brief ACT intervention in reducing interpersonal problems and fostering interpersonal flourishing: A randomized multiple-baseline evaluation. The Psychological Record, 75, 265–290. https://doi.org/10.1007/s40732-025-00634-3

Study Type

Randomized, concurrent, multiple-baseline single-case experimental design (SCED) with 15 participants.

---

CONTEXT AND OBJECTIVES

Interpersonal problems are persistent difficulties, challenges, or conflicts that people experience in their social relationships. Although such problems are frequent in multiple psychological disorders and considered a transdiagnostic process, they have received little attention as an outcome of psychological interventions. Repetitive Negative Thinking (RNT) has been identified as a central mechanism in various psychological disorders (e.g., generalized anxiety disorder, depression, eating disorders, substance use disorders).

Recently, an approach has emerged that integrates Acceptance and Commitment Therapy (ACT) with the concept of Early Maladaptive Schemas (EMS). This RNT-focused ACT model proposes that identifying hierarchical triggers of RNT and teaching participants to respond flexibly to aversive private experiences may be particularly effective in reducing interpersonal problems.

General objective: The study evaluated the efficacy of a brief, four-session RNT-focused ACT protocol in reducing interpersonal problems and fostering interpersonal flourishing using a randomized multiple-baseline design with 15 participants experiencing chronic interpersonal difficulties lasting at least 6 months.

---

METHOD

Participants

Fifteen participants (10 females, 5 males; mean age = 31.93 years, SD = 11.48) were recruited through an open call posted on social media (Facebook, WhatsApp) asking for readers experiencing difficulties interacting with others. Inclusion criteria were: (1) Colombian nationality; (2) scoring above the cutoff score established by Salazar et al. (2010) on the 64-item Inventory of Interpersonal Problems (IIP-64; Alden et al., 1990), which was set at 87 points; (3) having experienced interpersonal problems for at least 6 months; and (4) not concurrently undergoing psychological or pharmacological treatment.

Exclusion criteria were: (1) active psychological or pharmacological psychosis; (2) diagnosed with borderline personality disorder, bipolar affective disorder, antisocial personality disorder, addictions, substance abuse disorder, or presenting with high suicide risk. Of 28 potential participants meeting initial inclusion criteria, 15 completed the study (10 females; mean age = 31.93 years, SD = 11.48).

Most participants (10 of 15) presented with depression; others presented with generalized anxiety disorder (6) and social anxiety disorder (5). Only three participants did not have a formal diagnosis. Most achieved a postgraduate degree (2 participants), with others at the undergraduate level (higher education generally), with the lowest being high school education (4), and they were from a medium-to-low socioeconomic background.

Design

A randomized, concurrent, multiple-baseline design was conducted across 15 participants. Participants were randomly assigned to one of three cohorts:

• Cohort 1: intervention after 3 weeks of baseline (at least 21 days)
• Cohort 2: intervention after 4 weeks of baseline (at least 28 days)
• Cohort 3: intervention after 5 weeks of baseline (at least 35 days)

Intervention

RNT-Focused ACT Protocol (4 sessions of approximately 60 minutes each, except the first session which lasted approximately 90 minutes)

Sessions were administered weekly via videoconference (Google Meet or Microsoft Teams). The protocol was implemented by six therapists with master's degrees in clinical psychology, trained in the RNT-focused ACT model.

Session 1: Establishing context and introducing fundamental concepts

• Presentation of the intervention and its rationale focused on interpersonal problems
• Psychoeducation about early maladaptive schemas (EMS) and psychological flexibility
• Assessment of psychological inflexibility patterns linked to maladaptive schemas
• Assessment of interpersonal values and meaningful experiences
• Introduction of the physical metaphor ("pushing metaphor")
• Forest metaphor: presentation of two forests (one positive related to values learning; one negative related to hierarchical aversive self-content learning)
• Assessment and exploration of hierarchical self-contents based on highest YSQ-S3 scores
• Identification of hierarchical triggers and aversive self-contents
• Evaluation of RNT related to triggers and their negative effects on interpersonal relationships
• Assessment of additional experiential avoidance strategies
• Return to forest metaphor and questions about values
• Identification of interpersonal values
• Identification of daily actions that could advance toward interpersonal values
• Pushing metaphor: invitation to write hierarchical triggers on a piece of paper and other associated thoughts and feelings; introduction of the concept of not pushing the piece of paper and facilitating realization that they can choose whether to push or not when triggers surface

Session 2: Training in identification of RNT triggers and flexibility in response to them

• Review of flexible and inflexible reactions to triggers
• Multiple-exemplar training in identifying RNT triggers and defusion exercises
• Identification of valued interpersonal actions and barriers to engaging in them
• Balloon exercise: focusing on breathing, neutral thoughts, and aversive thoughts related to interpersonal problems that occurred recently
• Free association exercise: reading 12 words separated by 7 seconds; the participant must notice what thoughts emerge and place it on a balloon, realizing the hierarchical relationship between them and the possibility of choosing between inflexible and flexible reactions
• Free association exercises alternating "letting go" with "entanglement": similar to the previous exercise, but the therapist uses words related to the hierarchical trigger and asks the participant to let them pass with neutral response and to entangle in response to others
• Worrying exercise alternating entanglement and distancing and engaging in conversation with the therapist
• Request for the therapist to ask the participant to list valued interpersonal actions they could do instead of being entangled with thoughts
• Identification of barriers to engaging in valued interpersonal actions
• Identification of barriers to practicing the defusion exercises introduced earlier

Session 3: Hierarchical self-development and knowledge about the development of values and hierarchical triggers

• Review of flexible and inflexible reactions to triggers
• Development of hierarchical self-understanding and coherent self-knowledge about values and hierarchical trigger development
• Values clarification and identification of valued actions
• Garden metaphor (modified): review of the participant's multiple hierarchical relationships between their lives and private events; exploration of learning history related to values and hierarchical trigger development
• River metaphor (Wilson & Luciano, 2002): summary of information gathered during previous exercises and planning of agreed-upon actions

Session 4: Multiple-exemplar defusion training, review of intervention content

• Review of flexible and inflexible reactions to triggers
• Multiple-exemplar training in defusion
• Review of intervention content and what the participant learned
• Concert exercise: therapist guides the participant to imagine they are at a concert and asks the participant to observe the flow of thoughts and emotions without interfering with them
• Concert exercise with the hierarchical trigger: repetition of the previous exercise, but therapist guides the participant to now imagine past and future occurrences of the hierarchical trigger
• Cliff and waves metaphor (Wilson & Luciano, 2002): as a summary of previous exercises, a metaphor is introduced in which the participant can choose between observing a storm at the edge of a cliff (fighting and judging the waves symbolizing aversive private events) and observing the storm from the seawall promenade (choosing to let private events be and only observe their flow)

Protocol Integrity

Protocol integrity and therapist competence were analyzed similarly to previous RNT-focused ACT protocols. Treatment integrity was analyzed for 12 randomly selected sessions (i.e., 20% of sessions), with visualization order of three sessions for each Session 1, Session 2, Session 3, and Session 4, respectively. Four independent raters scored each session using checklists of clinical competence with two elements per session. All raters were trained by the first author in ACT and the intervention protocol by the second author, who is one of the developers of this type of intervention.

The mean percentage of protocol adherence was 97.35% of relevant clinical interactions contained in the protocol. The Kappa index was used to obtain inter-rater reliability. The mean inter-rater reliability was 95%, which can be interpreted as "almost perfect" inter-rater reliability. Regarding therapist competence, the mean percentage of clinical interactions performed competently was 95.24%, and the Kappa index was .90, which can also be interpreted as "almost perfect" inter-rater reliability.

Measurement Instruments

Primary Outcomes

Problematic Interpersonal Behavior (PIB)

This instrument consisted of three items collaboratively designed with each participant based on the most problematic items of the IIP-64 for them. Responses were on a 5-point Likert scale (0 = not at all, 4 = extremely) according to the instruction: "How many of the following problems have you experienced during the last 24 hours?" Table 1 presents the items for each participant.

Interpersonal Flourishing (IPF)

This instrument consisted of three items: "I have felt proud or satisfied with the way I have interacted with others," "I have behaved with others according to the ideal person I would like to be," and "I have taken actions to build the kind of interpersonal relationships I desire." Responses were on a 5-point Likert scale (0 = not at all, 4 = strongly agree) according to the instruction: "Indicate your degree of agreement with the following statements related to how you have felt and behaved in your interpersonal relationships during the last 24 hours."

A preliminary analysis of the three IPF items was conducted with 27 undergraduates who responded daily to this instrument for 20–24 days. A multilevel confirmatory factor analysis was conducted with a one-factor model. The model showed non-significant longitudinal factorial invariance (S-Bχ²(12) = 9.93, p = .62) and good comparative fit indices (CFI = 1.00) and root mean square error of approximation (RMSE = 0.00).

Secondary Outcomes

Daily Version of the Personal Health Questionnaire-4 (PHQ-4; Kroenke et al., 2009)

The PHQ-4 is a four-item screening instrument for depression and anxiety. The daily version of the PHQ-4 showed good internal consistency, a hierarchical factor structure with two first-order factors (Depression and Anxiety) and a second-order factor representing general emotional symptoms, longitudinal factorial invariance, and convergent validity. In the daily version, the PHQ-4 was responded to on a 5-point Likert scale (0 = not at all, 4 = every day).

Personal Health Questionnaire-9 (PHQ-9; Kroenke et al., 2001)

The PHQ-9 is a screening and severity assessment instrument for depression in primary care and research clinical settings. The scale consists of nine items responded to on a 4-point Likert scale (0 = not at all, 3 = nearly every day). The PHQ-9 showed good psychometric properties in a validation study in primary care settings, with a Cronbach's alpha of .80 and McDonald's omega of .81.

Generalized Anxiety Disorder-7 (GAD-7; Spitzer et al., 2006)

The GAD-7 is a brief questionnaire intended to serve as a screening instrument for generalized anxiety disorder in primary care and research clinical settings. It consists of seven items responded to on a 4-point Likert scale (0 = not at all, 3 = nearly every day), with total scores ranging from 0 to 21 points. The GAD-7 showed good psychometric properties in a Colombian sample, with a Cronbach's alpha of .92.

Process Outcomes

Repetitive Negative Thinking Questionnaire-3 (RNTQ-3; Ruiz et al., 2023)

The RNTQ-3 consists of three items responded to on a 5-point Likert scale (0 = not at all, 4 = every day) and was designed to measure daily engagement with RNT. It has shown excellent internal consistency, a one-factor structure, longitudinal factorial invariance, and convergent validity.

Psychological Flexibility in Interpersonal Relationships (PFIR)

It consisted of two items responded to on the same 5-point Likert scale: "I have been willing to feel vulnerable when I considered it appropriate to build the kind of interpersonal relationships I desire" and "I have acted with others as I really wanted to, even when I felt fear of failure." Preliminary analysis showed that PFIR exhibited longitudinal factorial invariance (S-Bχ²(3) = 3.27, p = .35, CFI = 1.00, RMSE = 0.01).

Awareness of Opportunities for Closeness (AOC)

It also consisted of two items responded to on the same 5-point Likert scale: "I have been aware of times when I could be loving and supportive of others," and "I have been aware of times when others tried to be loving and supportive of me." Preliminary analysis showed that AOC exhibited longitudinal factorial invariance (S-Bχ²(3) = 2.51, p = .47, CFI = 1.00, RMSE = 0.00).

Self-Awareness in Interpersonal Situations (SIS)

It consisted of two items: "I notice my thoughts and feelings when I am with other people" and "I have noticed my reactions and responses when I interact with other people." Preliminary analysis showed that SIS exhibited longitudinal factorial invariance (S-Bχ²(3) = 3.25, p = .36, CFI = 1.00, RMSE = 0.01).

Warm, Connected, and Compassionate (WCS)

It also consisted of two items responded to similarly: "I have created moments of warmth and connection with people I care about," and "I have been compassionate and empathetic with others, without rushing to judge them." In preliminary analysis, WCS exhibited longitudinal factorial invariance (S-Bχ²(3) = 0.16, p = .98, CFI = 1.00, RMSE = 0.00).

Young Schema Questionnaire – Short Form 3rd Edition (YSQ-S3; Young, 2005)

The YSQ-S3 is a questionnaire consisting of 90 items responded to on a 6-point Likert scale (1 = completely false of me, 6 = describes me perfectly). It consists of 15 subscales, and the scales are grouped into five theoretical domains. González-Cifuentes et al. (2024) found that the YSQ-S3 had good psychometric properties and the expected factor structure in two large Colombian samples.

Data Analysis

The raw data and R code used to compute the data analysis can be downloaded at https://osf.io/cz7yf/. This link also contains figures for each participant on each dependent variable.

Analysis and Imputation of Missing Data

The rate of missing data was low (5.92%), although it varied across participants (from 0% for P4 to 12.39% for P5). A single imputation method using an ARIMA model was adopted that performs similarly to more sophisticated methods in the absence of a high rate of missing data. This imputation was conducted for each participant and variable with the imputeTS R package.

Visual Analysis and Selection of Statistical Analyses for SCED

Statistical tests appropriate to the observed data pattern in primary outcomes were selected. Improving trends in baselines were not expected because having ongoing interpersonal difficulties for at least 6 months was an inclusion criterion. Most participants had stable baselines or small trends that might be due to typical daily fluctuations in the variables evaluated.

The only exception was P4, who showed an abrupt improvement in problematic interpersonal behavior (i.e., PIB scores) that nearly reached the non-clinical cutoff. To preserve the internal consistency of the study, this participant was excluded from the daily data analyses.

Several types of SCED analyses that are relevant to analyze, such as level, trend, overlap, immediacy, and variability were selected. Thus, three statistical analyses that control for higher-order autocorrelation or are robust to the presence of autocorrelation (Barnard-Brak et al., 2021) and that cover the main characteristics of SCED data were selected. For clarity, all analysis results are presented with a positive sign indicating intervention effects.

Within-Case Effect Sizes

Piecewise regression consists of estimating the level and trend of each phase of an SCED through regression analyses. The baseline score level and the level and trend during intervention were estimated. Following Maric et al. (2015), the dataset was coded so that the regression analysis examines the same endpoint of the intervention phase, which corresponds to the last observation of the 1-month follow-up. To compute the piecewise regression analysis, the GLS estimator with an autoregressive factor (AR) was adopted as recommended by Maggin et al. (2011) and Swaminathan et al. (2014). The non-parametric Tau-U statistic is a family of indices that computes the non-overlap between baseline and intervention phases while accounting for the trend in each phase. The R package scan was used, which allows specification of the method to compute the U level and compute different options for this statistic depending on the trend to apply.

Some relevant advantages of Tau-U versus other statistical options are that it exploits the presence of autocorrelation, controls Type I error rate, provides sufficient power to detect an intervention effect, and can support clinical decision-making. The Tau-U version that accounts for the non-overlap between baseline and intervention phases over the time each phase extends and the first calculators did not include all versions of this statistic. The R package scan was used, which allows specification of the method to compute U and compute different options for this statistic depending on the trend to apply. Thus, the Tau-U statistic that accounts for non-overlap between baseline and intervention phases was selected.

Between-Case Effect Sizes

The design-comparable standardized mean difference developed by Pustejovsky et al. (2014) provides an overall estimate of the intervention effect for the entire study. In doing so, it fits a hierarchical linear model that restricts maximum likelihood estimation and accounts for the serial dependence of data with an autoregressive factor (AR). The treatment phase was modeled with linear trends with fixed and random effects for level and slope.

Traditional Psychometric Instruments

Results on IIP-64, YSQ-S3, PHQ-9, and GAD-7 were analyzed using paired t-tests through JASP 0.18.1.0 and providing standardized mean differences for pre/post designs. To analyze reliable change for total IIP-64 and YSQ-S3 scores, the psychometric properties of these instruments in Colombian samples were considered.

Cutoff scores for claiming clinically significant change on the IIP-64 and YSQ-S3 were computed following the same procedure indicated for PIB. For the IIP-64, it was calculated and multiplied by 3 (i.e., PIB scores, which would be 3 times). It was interpreted that participants would be experiencing their most problematic interpersonal behavior between "little" and "moderately" daily. The cutoff was set at 4.18 points. This would mean that participants were experiencing their most problematic interpersonal behavior between "little" and "moderately" daily. The cutoff for claiming remission was set at a score ≤ 4.

For PHQ-9 and GAD-7, reliable change, clinically significant changes, and remissions were computed following McMillan et al. (2010) recommendations. In particular, reliable change was defined as a 5-point reduction, clinically significant change as obtaining reliable change and a score ≤ 9, whereas the cutoff for claiming remission was a score ≤ 4.

---

RESULTS

Within-Participant Results

Problematic Interpersonal Behavior (PIB)

Figure 1 shows the daily evolution of problematic interpersonal behavior (i.e., PIB scores) throughout the study. To facilitate visual inspection, the figure includes level and trend lines constructed with the R package scanlhm to compute the design-comparable standardized mean difference.

It can be observed that the introduction of the intervention was associated with score changes for most participants. In particular, Table 4 presents the statistical analyses conducted on level, trend, and overlap.

The piecewise GLS regression showed that 12 of 14 participants experienced statistically significant reductions in the level of problematic interpersonal behavior scores after introducing the intervention (i.e., the mean level of the intervention phase was lower than baseline). P3 experienced a reduction that was near statistical significance, whereas P14 experienced a statistically significant increase in score level. Eight participants showed a statistically significant improvement trend during the intervention phase (i.e., negative coefficient), and one participant (P14) experienced a deteriorating trend. The Tau-U vs. Baseline B results were statistically significant for all participants, indicating that intervention scores non-overlapped with baseline scores in the expected direction except for P14. The median value was 0.41. Eleven of 14 participants (78.57%) obtained clinically significant reductions in problematic interpersonal behavior.

Interpersonal Flourishing (IPF)

Figure 2 depicts the daily evolution of interpersonal flourishing (i.e., IPF scores). Visual inspection indicates that the intervention was associated with score changes for most participants. Table 5 shows that piecewise GLS regression identified statistically significant increases in level change for 10 of 14 participants. However, two participants (P13 and P14) experienced decreases in score levels. Likewise, nine participants showed statistically significant improvement trends during the intervention phase, but P13 and P14 also showed deteriorating trends. The Tau-U vs. Baseline B statistic showed that 11 of 14 participants experienced statistically significant increases in interpersonal flourishing. Two remained the same, while P13 showed a statistically significant decrease. The median Tau-U value was 2.08. Ten participants showed clinically significant increases in interpersonal flourishing (71.42%).

Between-Participants Results

The design-comparable standardized mean differences can be observed in Table 7. Effect sizes for primary outcomes were large: d = 0.97 for problematic interpersonal behavior and d = 0.94 for interpersonal flourishing. Values for most interpersonal process variables were also in the large effect size range: d = 0.97 for self-awareness in relationships, d = 0.90 for psychological inflexibility in relationships, and d = 0.74 for warmth, connection, and compassion. The remaining effect sizes were more modest: d = 0.55 for both RNT symptoms and emotional symptoms, and d = 0.25 for awareness of opportunities for closeness.

Results on Traditional Psychometric Instruments

Results on interpersonal problems (IIP-64) and early maladaptive schemas (YSQ-S3) can be observed in Table 8. Overall, participants showed statistically significant decreases on both instruments at posttreatment (IIP-64: t(14) = 5.63, p < .001; YSQ-S3: t(14) = 4.69, p < .001) and at 1-month follow-up (IIP-64: t(13) = 6.47, p < .001; YSQ-S3: t(13) = 6.73, p < .001). Cohen's d values were very large in both cases at 1-month follow-up (IIP-64: dfollow-up = 1.73, 95% CI [0.88, 2.56]; YSQ-S3: dfollow-up = 1.80, 95% CI [0.53, 1.87]). At 1-month follow-up, 12 of 14 (85.7%) participants obtained reliable change in IIP-64 scores, and 10 of 14 (71.4%) had clinically significant change. Regarding YSQ-S3, all participants obtained reliable change at follow-up, and 11 of 14 (78.6%) had clinically significant change.

Table 9 presents scores on depression (PHQ-9) and anxiety (GAD-7). Changes were statistically significant at both posttreatment (PHQ-9: t(14) = 2.24, p < .05; GAD-7: t(14) = 3.99, p = .001) and 1-month follow-up (PHQ-9: t(13) = 4.14, p = .008; GAD-7: t(13) = 4.93, p < .001). Cohen's d values were large in both cases at 1-month follow-up (PHQ-9: d = 0.84, 95% CI [0.22, 1.44]; GAD-7: d = 1.73, 95% CI [0.88, 2.56]). At 1-month follow-up, seven of nine (77.8%) participants showed reliable change in depression, six of nine (66.7%) clinically significant change, and four of nine (44.4%) complete remission. Regarding anxiety, three of five (60%) participants obtained reliable change, and complete remission of GAD was obtained.

Qualitative Results

Table 10 summarizes participants' comments in the interview conducted at the 1-month follow-up. Two participants (P5 and P9) did not attend the interview due to contact problems. Three thematic categories explained significant aspects of participants' improvements and experiences. First, most participants (11 of 13) expressed improvements in interpersonal relationships due to focus on social interactions, increased awareness in these situations, boundary setting, more assertive behavior, or closer engagement in social relationships. Second, management of negative thoughts, the decrease in RNT, and the emotional strain associated with it stood out as a noteworthy achievement for most participants (10 of 13). Third, most participants also expressed improvement in decision-making by focusing on values and committed actions, and greater self-acceptance and empowerment (9 of 13).

---

DISCUSSION AND CONCLUSIONS

Interpersonal problems are prevalent across various psychological disorders and have been largely overlooked as a distinct category and as an outcome of psychological interventions. The present study sought to evaluate the efficacy of the brief, four-session RNT-focused ACT protocol designed in the present study to reduce interpersonal problems and foster interpersonal flourishing.

The intervention demonstrated high efficacy on primary outcomes. In terms of daily measures, it successfully reduced problematic interpersonal behaviors in most participants (92.86%), with 78.57% experiencing clinically significant change. Furthermore, most participants showed increases in interpersonal flourishing (71.4%), reaching a considerably larger effect size than the IIP-64 (d = 1.73) at 1-month follow-up. These results are comparable to those from previous group ACT studies employing larger protocols (e.g., Lev, 2011; Quilantan et al., 2018).

The intervention proved highly efficacious in achieving clinically significant change. The present protocol is consistent with previous interventions for these disorders (Cuijpers et al., 2021; Loering et al., 2015). However, the intervention presented in this study is similar to those used in previous studies and includes at least one additional process variable from those interventions for an important measure in this study. One explanation for the difference in the effect on emotional symptoms could be related to: (1) the higher heterogeneity of this sample, (2) the different motivations for participants in this study regarding solving interpersonal problems versus suffering from depression and anxiety; or (3) the delivery of the protocol via teleconference.

The results of this study show considerable variability in intervention steps on daily measures. The variability between participants in effect sizes was probably larger due to varied self-report effects. Among participants in the study who exhibited reductions in problematic interpersonal behavior and increases in interpersonal flourishing, also observed were increases in self-awareness and psychological inflexibility in relationships, while changes in RNT appear more related to reductions in emotional symptoms.

---

SIGNIFICANCE AND CONTRIBUTION

This study makes important contributions to the field of interpersonal psychotherapy by demonstrating that a brief four-session intervention based on principles of acceptance and committed action can effectively reduce interpersonal problems and foster flourishing in social relationships. The integration of early maladaptive schema concepts with behavioral principles provides a comprehensive approach to addressing persistent relational difficulties. Notably, the study demonstrates that this intervention can be delivered effectively via videoconference format and maintained its benefits at one-month follow-up, suggesting practical applicability for diverse populations. The use of single-case experimental design with randomized multiple baselines provides rigorous evidence of efficacy, contributing to the literature on brief, scalable interventions for interpersonal difficulties, a commonly overlooked yet clinically important area of psychological treatment.

---

---