Acceptance and commitment therapy focused on repetitive negative thinking for complicated breakup grief: A randomized multiple-baseline evaluation

Context and Objectives

Romantic relationship breakups are considered significant life events that can provoke important behavioral and emotional consequences. These emotional reactions can take the form of grief, with a high frequency of intrusive thoughts, feelings of loneliness and emptiness, sadness, lack of interest in personal activities, insomnia, compromised immune system, and the "broken heart syndrome." More specifically, when going through a breakup, it is common to experience reactions such as anger, distressing others, and causing worry about them, feeling intense sadness when remembering the former partner, being attracted to places and objects related to the ex-partner, or avoiding situations associated with him/her. Complicated breakup grief occurs when the emotional reaction persists for more than three months and is characterized by a progressive increase in anger and feelings of betrayal and mistrust of others. Research has identified grief rumination (known as repetitive negative thinking or RNT in the breakup context) as a key psychological process involved in its onset and maintenance.

In recent years, a version of Acceptance and Commitment Therapy (ACT) focused on disrupting unconstructive worry and rumination has been developed, called RNT-focused ACT (RNT-focused ACT). This approach aims to foster psychological flexibility to produce changes in mental health, quality of life, and behavioral effectiveness. Psychological flexibility can be defined as the ability to nonjudgmentally contact private experiences and orient behavior toward valued directions.

The objective of this study was to explore the potential efficacy of a brief RNT-focused ACT protocol in complicated breakup grief. Given the relevance of grief rumination in complicated bereavement, it was anticipated that ACT protocols centered on RNT should enhance the efficacy of ACT because the cycle of experiential avoidance would be interrupted from the very beginning. This study aimed to conduct a nonconcurrent, randomized, multiple-baseline design on three females showing complicated breakup grief.

Method

Participants

Participants were recruited through advertisements on social media (i.e., Facebook, institutional webpage, etc.). The inclusion criteria were: (a) being 18 years or older; (b) having experienced complicated grief for at least three months due to the breakup of a romantic relationship that had lasted at least 12 months.

The exclusion criteria were: (a) undergoing current psychological or psychiatric treatment, (b) experiencing severe suicide ideation, (c) having a psychotic disorder, and (d) substance abuse.

A total of 72 individuals showed an interest in the research. However, only 3 participants met the inclusion criteria. Of the 72 individuals excluded: 20 individuals had experienced breakup grief for less than three months, 16 had had relationships that lasted less than 12 months, 13 were receiving psychological or psychiatric interventions, 7 were in a cyclic relationship with their ex-partner, after the interview, 5 said that they did not have time to participate, 4 did not respond to the phone calls and emails, 2 got back together with the ex-partner, 1 showed suicide ideation, and 1 was living in another country.

Participant characteristics:

Participant 1 (P1) was a 23-year-old woman who had experienced the breakup of a 5-year romantic relationship two years before treatment. Participant 2 (P2) was also a 23-year-old woman who had been married for five years. The breakup was due to her ex-husband's infidelity six months before recruitment. Participant 3 (P3) was a 32-year-old woman who had maintained an engagement for one year. She terminated it eight months before recruitment when she realized that her boyfriend was losing interest in her.

Design

A nonconcurrent, randomized, multiple-baseline design across participants was implemented. The independent variable of the study was the staggered introduction of a 3-session, RNT-focused ACT protocol. Participants were randomly assigned to receive the intervention after collecting baseline data between 3 to 5 weeks. The protocol was implemented weekly. Afterward, an 8-week follow-up was conducted.

Dependent variables were divided into primary, secondary, and process outcome measures. The primary outcome measure was breakup distress, which includes both emotional symptoms and behavioral inflexibility in response to private events related to the breakup. Secondary outcome measures were scores on emotional symptoms and life satisfaction. Lastly, process outcomes were measures of RNT, experiential avoidance, cognitive fusion, and valued living. These measures were selected as process outcomes because they are the main putative processes in the ACT model to promote mental health, quality of life, and behavioral effectiveness. Blinding procedures were not implemented because the study only involved one intervention, and the dependent measures were taken through automatic emails on the Internet.

Intervention

The protocol consisted of three face-to-face, individual sessions. The first session lasted approximately 90 minutes, and the second and third sessions lasted about 60 minutes. The protocol was based on the one used in Ruiz et al. (2019, 2020).

Session 1: Presentation of the Intervention

• Work procedure: developing the skill to identify entanglement with our thoughts and learn to focus on what is important in our lives (with special focus on the couple relationship).
• Hierarchical relations between the self and thoughts/emotions (remembering thoughts in different life moments and observing the flow of thoughts).
• Dike metaphor: Thoughts are like leaves on a stream, but we can choose to build a dike to stop the leaves and analyze them to the point of creating whirlpools of thoughts.
• Cards exercise: Write down ongoing negative thoughts, observe them, and choose to build a dike or let them pass.

Session 2: Identification of the main triggers to initiate worry/rumination and other experiential avoidance strategies

• Socratic dialogue: (a) In which direction are you going when you worry and try to avoid controlling your thoughts? (b) Are they helpful in the short term? (c) And in the long term? (d) Are they stronger than before?
• Physical metaphor: "Pushing triggers away." The experimenter writes the participant's triggers on a piece of paper and puts them near the participant's face. When participants begin to push the piece of paper away with their hands, the experimenter resists.
• Questions: (a) How much strength would you have if you pushed a more veil away with their hands for MH years? (b) Can you do something important while pushing? (c) How much stronger would they be if you pushed more veil away with hands for MH's years?
• Exercise of alternating RNT and defusion from thoughts in a conversation with the therapist.
• Pink elephant exercise (based on Hayes et al., 1999): The participant was invited to avoid thinking about a pink elephant.

Session 3: Identification of valued actions

• Exploration of RNT and valued actions during the last week.
• Multiple-example training in defusion techniques and RNT and defusion exercises.
• Defusion exercise: focusing on breathing and the surrounding noises while noticing who chose to do it. Then, practice putting ongoing thoughts on balloons and letting them float. Lastly, practice with two situations in which triggers for RNT were activated.
• Daydreaming and worrying exercise: The participant was invited to daydream for 2 minutes. Each 20 s, the therapist asked the participant to notice what she was thinking and how she was able to choose between following or stopping the process. The same process was repeated with worry.
• Exercise of alternating RNT and defusion from thoughts in a conversation with the therapist.
• Pink elephant exercise (based on Hayes et al., 1999): The participant was invited to avoid thinking about a pink elephant.
• The therapist asked the participant to list some valued actions that she could do instead of being entangled with her thoughts and identify her proud at the end of the day.
• The therapist asked the participant to list some valued actions that she could do instead of being entangled with her thoughts on engagement (with a special focus on the couple relationship).
• Identifying possibilities to behave inflexibly and values-oriented in a typical day.
• Summary of the intervention and commitment to respond to the follow-up assessments.

Measurement Instruments

Primary outcome measure:

Breakup Distress Scale (BDS; Field et al., 2009). The BDS is a 16-item, 4-point Likert-type scale that measures the frequency with which the individual experiences breakup-related distress, both in the form of emotional symptoms and inflexible behavior in response to private events related to the breakup. The BDS showed an alpha of .91 in its original validation. The method described in Muñiz et al. (2013) was used to translate the BDS into Spanish.

Secondary outcome measures:

Depression, Anxiety, and Stress Scales – 21 (DASS-21; S. H. Lovibond & P. F. Lovibond, 1995; Spanish version by Ruiz et al., 2017). The DASS-21 is a 21-item, 4-point Likert-type scale that measures the negative emotional states experienced last week. The DASS-21 showed excellent internal consistency and a hierarchical factor structure that justifies computing an overall indicator of emotional symptoms (i.e., DASS-Total). The DASS-21 showed an alpha of .89.

Satisfaction with Life Scale (SWLS; Diener et al., 1985; Spanish version by Ruiz et al., 2019). The SWLS is a 5-item, 7-point Likert-type scale that measures self-perceived well-being. It showed a one-factor structure and an alpha of .89.

Process outcome measures:

Perseverative Thinking Questionnaire (PTQ; Ehring et al., 2011; Spanish version by Ruiz et al., submitted). The PTQ is a 15-item, 5-point Likert-type self-report instrument. It is a content-independent self-report of RNT in response to negative events. The Spanish version of the PTQ has shown excellent internal consistency and a one-factor structure.

Acceptance and Action Questionnaire – II (AAQ-II; Bond et al., 2011; Spanish version by Ruiz et al., 2016b). The AAQ-II is a 7-item, 7-point Likert-type scale that measures experiential avoidance as averaged across contexts. The Spanish version of the AAQ-II showed excellent internal consistency and a one-factor structure.

Cognitive Fusion Questionnaire (CFQ; Gillanders et al., 2014; Spanish version by Ruiz et al., 2017). The CFQ is a 7-item, 7-point Likert-type scale that measures cognitive fusion as averaged across contexts. The Spanish version of the CFQ has shown excellent internal consistency and a one-factor structure.

Valuing Questionnaire (VQ; Smout, Davies, Burns, & Christie, 2014; Spanish version by Ruiz et al., 2022). The VQ is a 10-item, 7-point Likert-type self-report instrument that assesses valued living averaged across life areas during the past week. It comprises two subscales: Progress and Obstruction. The Spanish version of the VQ has shown good internal consistency and a two-factor structure.

Data Analysis

Raw data of this study can be accessed at https://bit.ly/39uZlDI. Visual analyses were conducted to analyze the results of each participant according to the guidelines provided by Lane and Gast (2014). The data were analyzed at the individual and group levels. The nonparametric Tau-U test (Parker et al., 2011) was computed to perform within-participant analysis using the calculator http://singlecaseresearch.org/calculators/tau-u. This test provides a nonoverlapping effect size between baseline and intervention data and computes a p-value that determines the presence of a statistically significant intervention effect. Tau-U values range from -1 and 1 and indicate the percentage of data that improve through the baseline to intervention phases. For the sake of clarity, we present all effect sizes in favor of the intervention phase as positive, regardless of whether its scores should decrease or increase.

The achievement of clinically significant changes was analyzed through the guidelines presented by Ruiz et al. (2018). The criteria to obtain a clinically significant change required: (a) the Tau-U value to be statistically significant in favor of the intervention phase, and (b) to show a score at the 8-week follow-up closer to the mean of the nonclinical than the clinical population. To test the latter criterion, we consulted descriptive data obtained from scale validation studies (see Ruiz et al., 2018).

Overall estimations of the intervention effect sizes were conducted with the standardized mean difference developed for single-case experimental designs by Hedges et al. (2013), which accounts for the autocorrelation typically seen in these designs. This analysis yields an overall estimation of the effect size that shares the same metric as Cohen's d used in group designs. The analysis was computed using the R package schlm.

Results

Within-Participant Results

Figure 1 shows the scores' evolution of primary and secondary outcomes. Participants showed high scores on breakup distress (i.e., BDS), moderate to severe scores on emotional symptoms (i.e., DASS-Total), and scores slightly below the average in life satisfaction (i.e., SWLS). Visual inspection shows that the RNT-focused ACT protocol was very efficacious in decreasing the BDS scores, with large and immediate effects. During the follow-up, the scores on the BDS remained at the minimum level of the scale. The treatment effect was more gradual in the case of emotional symptoms and life satisfaction, but the effects were also clearly visible. During the follow-up, the DASS-Total scores reached minimum levels, and the SWLS scores indicated high life satisfaction.

Table 2 presents the results on Tau-U and the analysis of the presence of clinically significant changes. All three participants showed Tau-U values of 1.00 on breakup distress (i.e., BDS scores), reaching clinically significant changes. Regarding emotional symptoms, P1 and P2 showed statistically significant Tau-U values and reached clinically significant changes. P3 did not obtain a clinically significant change because the change in emotional symptoms was not immediate after introducing the intervention. However, her emotional symptom level was extremely low during the 8-week follow-up. Lastly, all three participants showed statistically significant Tau-U values on life satisfaction and obtained clinically significant changes in this outcome.

Process Outcome Results

Figure 2 presents the scores' evolution in process outcome measures. Visual analyses reveal that the three participants showed decreases in RNT, experiential avoidance, cognitive fusion, and values obstruction. Similarly, all participants showed increases in progress in valued living. Table 2 also shows that all Tau-U values were statistically significant and that all participants obtained clinically significant changes in all process measures.

Between-Participant Results

Table 3 shows that the standardized mean difference for SCED showed that the treatment effect was statistically significant for all variables. Specifically, the effect size for the primary outcome (i.e., breakup distress) was extremely large (d = 7.11). Likewise, the effect sizes for the secondary outcomes were very large, although more modest (DASS-Total: d = 2.46; SWLS: d = 1.25). Lastly, the effect sizes for process measures were very large in all cases (PTQ: d = 4.06; AAQ-II: d = 4.43; CFQ: d = 5.37; VQ-Progress: d = 1.59; and VQ-Obstruction: d = 2.09).

Anecdotal Comments

Anecdotal comments by the participants revealed that, after the intervention, they engaged in more social activities and had fewer arguments with people close to them. Participants also undertook and maintained valued plans, and they abandoned spying through social media. Lastly, participants experienced a decrease in health issues and improved sleep quality.

Discussion and Conclusions

This study aimed to preliminarily analyze the efficacy of a brief RNT-focused ACT protocol in three participants experiencing complicated breakup grief. This type of protocol was considered promising for intervening in complicated breakup grief, given the essential role played by the persistent rumination in response to intrusive thoughts and memories related to the romantic relationship and the engagement in additional experiential avoidance strategies (Field et al., 2009; Luhmann et al., 2012; Murdock et al., 2014; Sharra & Emery, 2005).

The results showed that the three participants experienced very large improvements in breakup distress, which reached a zero level at the end of the 8-week follow-up. The improvement in emotional symptoms was also very large, but the change was produced more gradually. A similar pattern followed in the case of life satisfaction, in which all participants reached high scores at the end of the follow-up. All effect sizes were very large and statistically significant. The intervention was efficacious in all process outcomes. Specifically, we observed reductions in RNT, experiential avoidance, cognitive fusion, and values obstruction, and increases in values progress.

Some limitations of this study are worth noting. Firstly, nonconcurrent multiple-baseline designs cannot control for history or maturation effects that might coincide with the application of the intervention (Harvey et al. 2004). This design was chosen because of the difficulty of recruiting participants. Secondly, the minimum number of participants recommended for conducting a multiple-baseline design is three. Although the effect of the intervention was evident in this study, these findings should be replicated in other studies with a larger sample. Thirdly, dependent variables were measured exclusively through self-reports. Further studies might include data from observers or self-registers. Lastly, the design of this study did not allow analyzing the potential mediators of the intervention. For this purpose, it would be necessary to conduct a more intensive assessment (e.g., ecological momentary assessment).

In conclusion, this study preliminarily indicates that RNT-focused ACT protocols might be efficacious for treating complicated breakup grief. Subsequent studies should replicate these findings with more participants or extend them by conducting a randomized controlled trial comparing the effect of these protocols with a waitlist control condition or a psychological placebo.

Significanceand contribution

This study provides preliminary evidence that brief Acceptance and Commitment Therapy protocols addressing repetitive negative thinking can be efficacious in treating complicated breakup grief. The demonstration that a brief 3-session protocol produces clinically significant changes in breakup distress with large effect sizes and sustained effects at 8-week follow-up extends the applicability of brief psychological interventions to grief and loss contexts. The results show changes not only in primary distress outcomes but also in transdiagnostic processes including cognitive flexibility and values-directed living.

---