A Preliminary Test of the Efficacy of Brief Self-Administered Behavioral Interventions for Rumination

Abstract

Rumination is theorized to be a cognitive avoidance process that is implicated in several manifestations of psychopathology. Few interventions directly target rumination as a core process maintaining emotional disorder symptoms. This pilot study compared the feasibility and preliminary efficacy of self-directed behavioral approaches for reducing rumination. Participants (N = 60) with elevations in rumination were randomized to 3 weeks of self-directed interventions: (a) scheduled rumination time; (b) a strategy combining mindfulness, shaping, and disengagement strategies; or (c) self-monitoring control. Both active treatment conditions outperformed self-monitoring control on post-treatment depression scores. Scheduled rumination time significantly outperformed the other two conditions on measures of rumination and worry. No between-group differences emerged on the secondary outcome (i.e., anxiety symptoms). Brief, self-directed, behavioral interventions targeting rumination are feasible and demonstrate preliminary efficacy. Scheduled rumination time shows moderate to large effects. The use of a small, non-treatment seeking sample was the primary limitation.

Keywords

rumination self-administered interventions behavior therapy

Rumination is a maladaptive emotion regulation strategy characterized by engagement in passive and repetitive thinking about emotional symptoms and their consequences (Aldao et al., 2010). Individuals who react to negative affect with ruminative thinking may do so to understand the causes and consequences of their own negative affect in an effort to reduce it. However, rumination typically does not have the desired effect of providing insight or solving a problem that will lead to symptom reduction. Instead, rumination generates and exacerbates negative affect (see Nolen-Hoeksema et al., 2008, for a thorough review).

Some cognitive and behavioral interventions for depression address rumination, such as behavioral activation (BA) treatment for depression (Martell et al., 2010). Conceptualizing rumination as avoidance of engaging in positively reinforcing activities, in BA, patients are instructed to direct attention to the scheduled activity during times of rumination, rather than to engage with it. Although effects of BA on depressive symptoms moderate to large (Cuijpers et al., 2007; Dimidjian et al., 2006), rumination has only been measured and observed to decrease occasionally as an outcome in clinical trials investigating the efficacy of BA (e.g., Moshier & Otto, 2017). Therefore, little is known about the success of this strategy specifically on rumination, as neither the ability to disengage from rumination and shift to a pleasurable activity nor reductions in rumination itself over time are explicitly or systematically assessed as outcomes in clinical trials of BA. Rumination-focused cognitive behavioral therapy (RFCBT) is another treatment for depression (Watkins et al., 2007) that focuses on using functional analysis to help patients identify unhelpful rumination, and incorporates imagery to recall mental states when a more helpful thinking style was used. Effect sizes for RFCBT are moderate for depressive symptoms, and modest to moderate for the presumed mechanism of change (i.e., rumination; Jacobs et al., 2016; Watkins et al., 2011). Finally, meta-cognitive therapy addresses beliefs about rumination in depressed patients and has received some preliminary empirical support for reducing rumination in a small, multiple-baseline study (Wells et al., 2009) and in a small pilot sample (Wells et al., 2012). Taken together, more work is needed to develop and evaluate strategies for targeting rumination, particularly in a transdiagnostic sample.

Despite the prevalence of rumination in depression and other manifestations of psychopathology, little work has been done to directly address it, and there is room for improvement among the few evidence-based treatments for depression that do address rumination. Moreover, no interventions for rumination have been evaluated outside of the context of treating major depressive disorder, leaving a significant proportion of patients not receiving these treatments who may benefit. Given that rumination is a process associated with a variety of high risk symptoms and behaviors such as substance use (Caselli et al., 2010; Nolen-Hoeksema & Harrell, 2002), suicidal ideation and self-injurious behavior (Gardner et al., 2014; Smith et al., 2006), and binge eating (Gordon et al., 2012), developing and evaluating transdiagnostic interventions that target the process of rumination can have a significant public health impact. Finally, interventions that are straightforward and lend themselves to self-administration have greater potential to increase access to treatment and improve the reach of evidence-based interventions to a larger number of individuals who may benefit (Ahl et al., 2013; Cuijpers et al., 2010).

Several distinct strategies that are drawn from the principles of change in behavior therapy have the potential to improve rumination. Although the principles of change that map onto these strategies are often considered mutually exclusive, many of these approaches share overlapping mechanisms (Arch & Craske, 2008) and draw from similar principles (e.g., approach v. avoidance). The elements of these strategies are based on distinct but complementary evidence-based approaches. For example, helpful disengagement strategies that discourage thought suppression (which has paradoxical effects; Wegner et al., 1987; Wenzlaff & Luxton, 2003) and mindfulness (see Hofmann et al., 2010, for meta-analytic review) each have potential to reduce ruminative thinking. Developing an approach that combines these strategies by encouraging thoughtful and mindful disengagement from rumination may still remain simple yet more potent. Other strategies based in principles of behavior change may also be promising: BA has a strong evidence-base for reducing depressive symptoms (Cuijpers et al., 2007), and stimulus control and shaping are strategies based on foundational principles of learning that have been included in many behavioral therapies for emotional disorders for decades (Borkovec et al., 1983; Werry & Wollersheim, 1967). However, these strategies have yet to be integrated and translated into a clinically meaningful framework for targeting rumination. Novel yet simple behavioral interventions can involve weaving together existing behavioral approaches that are typically utilized in their own “silos” or interventions (e.g., combining shaping or stimulus control with mindfulness and behavioral activation approaches). Moreover, many of these strategies are relatively straightforward to teach and therefore may be amenable to being delivered with minimal therapist contact.

This pilot study aimed to evaluate the preliminary efficacy of several behavioral strategies that have the potential to reduce rumination. In order to increase reach and accessibility, we focused on strategies that could be taught in one brief session and then self-administered during a brief (3-week) period. We aimed to simply and meaningfully link strategies to create processes by which individuals could respond to rumination. Our preliminary study evaluates the potential utility of specific components (or meaningfully integrated sets of components). We focused on brief, simple, streamlined approaches, which require fewer patient and provider training and resources and are well-suited for self-administration.

Thus, we randomized individuals with elevations in rumination to one of three brief, self-administered interventions: (1) a strategy that incorporates elements of mindfulness, shaping, and the spirit of responding to rumination as seen in BA (Jacobson et al., 2001), called “Accept, Limit, Redirect” (ALR;); (2) a strategy that incorporates mindfulness with stimulus control (akin to scheduled worry time for generalized anxiety disorder; Borkovec et al., 1983; McGowan & Behar, 2013), called “Save it for Rumination Time” (SRT); and (3) a self-monitoring only group (SMC). Participants attended one laboratory visit to determine eligibility and receive a brief intervention instructional session. The intervention was self-administered for 3 weeks.

Although each of these strategies described above appear multi-faceted, they are integrated in a simple, seamless way, creating a single process for an individual to practice in responding to rumination (see Methods). Furthermore, although each of these individual components (e.g., mindfulness, shifting attention to pleasurable activities, stimulus control) has its own evidence-base, as an innovative pilot study, we meaningfully linked some of these strategies together in ways that we have anecdotally seen benefit among our patients who engage in ruminative thinking. Our approach in clinical practice has been in response to patients who report difficulty disengaging from rumination, either in the context of behavioral activation or throughout the day. We sought to empirically evaluate some of the combinations of strategies we have been employing in practice and with which we have seen within-patient benefit. We hypothesized that ALR- and SRT-treated participants would show greater decreases in rumination and depressive symptoms than those assigned to SMC by post-treatment, as well as greater decreases in secondary outcomes of worry and anxiety. We also expected that ALR- and SRT-treated participants would show greater within-group change from pre- to post-treatment than those in SMC.

Secondarily, we assessed the effects of these behavioral interventions on worry. Although worry and rumination are distinct processes, they share overlapping features of repetitive and unhelpful thinking styles (Aldao et al., 2010). Although rumination has been historically linked with depressive disorders and worry with anxiety disorders (e.g., Nolen-Hoeksema et al., 1993, 1994; Robinson & Alloy, 2003), rumination is observed among individuals with anxiety symptoms and disorders (Abbott & Rapee, 2004; Kocovski et al., 2005; Nolen-Hoeksema, 2000; Olatunji et al., 2013) and is thus a transdiagnostic construct relevant across emotional disorders (McLaughlin & Nolen-Hoeksema, 2011; Wolitzky-Taylor et al., in press). Therefore, although the interventions did not explicitly intervene on worry, we anticipated that effects of these self-guided interventions may have generalized from rumination to worry given the high degree of overlap between the two constructs in combination with emerging evidence that both are transdiagnostic processes related to multiple emotional disorders.

Finally, given the self-directed nature and novelty of the interventions, we evaluated the feasibility of individuals successfully self-administering these strategies.

Methods

Participants

Participants were recruited from an undergraduate psychology subject pool at a large university. Students scoring 1 SD > M for a normative sample (i.e., M = 9.40, SD = 2.96) on a shortened, 10-item version of the Rumination Responses Scale (RRS; Treynor et al., 2003) were invited to the laboratory for an in-person assessment to further determine eligibility. During the consent process, they were informed that they met initial eligibility criteria due to their scores on a rumination measure. Because the intervention was self-directed, risk was minimized by excluding participants who endorsed moderate to severe suicidality on the Mini International Neuropsychiatric Interview (MINI) suicidality module (note that although this was an exclusion criterion, no participants were actually excluded based on this criterion). Sample (N = 60) demographics are reported in Table 1. See Figure 1 for flow through the study.

Table 1.

Baseline Demographics.

Variable	% or M (SD)
Gender	69 Female
Age	19.41 (2.55)
Race	47.5 Asian-American
Race	22 White, 15.3 Hispanic/Latino, 5 Black/African-American, 1.7 Native-American, 3.4 multiracial, 5.1 other race/ethnicity
Baseline DASS-D	5.19 (3.90)
Baseline DASS-A	5.17 (4.02)
Baseline RRS	52.69 (12.72)
Baseline PSWQ	59.47 (10.39)

Figure 1.

Participant flow through the study.

Measures

Depression and anxiety symptoms

The Depression Anxiety and Stress Scale (DASS-21) is a reliable and well-validated measure (Brown et al., 1997; Lovibond & Lovibond, 1995). Participants indicate the degree to which they experienced symptoms of depression, anxiety, and stress during the last week, on a 4-point scale (0 = “did not apply to me at all” to 3 = “applied to me very much, or most of the time”). Depression (DASS-D) and anxiety (DASS-A) subscales were used in the current analyses, as our primary interest was in these symptoms (depression primarily, and anxiety secondarily) and not in general stress. Internal consistency was a = 0.80 and a = 0.87 at pre-assessment and a = 0.78 and a = 0.89 at post-assessment for the anxiety and depression subscale, respectively.

Rumination

The Ruminative Response Scale (RRS) measures responses to depressed mood that are focused internally on symptoms and possible causes or consequences of symptoms (Nolen-Hoeksema, 1991). Participants rate how often they respond in accordance with each item when they feel depressed on a scale from 1 (almost never) to 4 (almost always). Internal consistency was a = 0.87 at pre-assessment and a = 0.89 at post-assessment. As described earlier, the brief, 10-item RRS was used for the purposes of initial screening through the undergraduate psychology pool. The full RRS was used for the pre- and post-treatment assessment.

Worry

The Penn State Worry Questionnaire (PSWQ; Meyer et al., 1990) measures the tendency to worry excessively and uncontrollably. Participants rate items on a 5-point scale (1 = “not at all typical of me” to 5 = “very typical of me”). The PSWQ demonstrates high internal consistency and good test-retest reliability (Molina & Borkovec, 1994). Internal consistency was a = 0.89 at both pre- and post-treatment assessment.

Nightly ratings

Participants were asked to record their use and success with their prescribed intervention at the end of each day. These data were conceptualized as indices of feasibility, adherence, and a manipulation check. Participants recorded their adherence, success with the intervention, number of minutes of rumination, and mood (0 = very low mood to 100 = very good mood) each night on a web-based survey. Specifically, all participants were asked the following questions: (1) Adding up all the times I ruminated today, I ruminated for this many minutes: (free number response); and (2) Rate your mood today on a scale from 0 to 100, where 0 is very low mood, 50 is neutral, and 100 is very high mood. Mood ratings were primarily used to determine whether a phone contact was needed (e.g., if mood worsened by 50% or more from baseline on any given rating, which never ended up happening).

Due to the different instructions and foci of each intervention, separate, intervention-specific questions were also asked. In SRT, additional ratings included (1) the number of times the participant noticed rumination, (2) whether the participant completed rumination time that day, and 3) how many times the participant successfully saved rumination during the day for the specified rumination time. In the ALR condition, additional ratings included (1) how many times that day the participant was able to limit rumination to <5 minutes each time it was noticed, (2) the number of times the participant was able to successfully redirect attention to one of the specified pleasurable or mastery activities identified during the in-person visit, and (3) a brief description of the themes of rumination for that day.

Procedures

Design and general overview of procedures

The study was approved by the institution’s Institutional Review Board. The general procedures of the study included: (1) web-based screening; (2) in-person baseline assessment to assess for eligibility and collect pre-treatment data; (3) in-person instructions for completing the assigned intervention (at the same visit as the baseline assessment for those eligible and randomized to condition); (4) 3 weeks of self-administered intervention, along with self-reported nightly ratings of adherence; and (5) a web-based post-treatment assessment. All assessments and instructions were highly standardized and scripted and delivered by a highly trained undergraduate-level research assistant (author HB), who was trained by an advanced doctoral student in clinical psychology (author JR), under the supervision of first author (KWT).

Screening

Students participating in the undergraduate subject pool completed a shortened, 10-item version of the RRS (Treynor et al., 2003). Those who scored at least 1 SD > normative mean on this scale (M = 9.40, SD = 2.96) were invited to the laboratory for further screening. At the laboratory visit, participants underwent the informed consent process and then completed the baseline assessment which consisted of the MINI suicide risk module, a demographic questionnaire, and the web-based questionnaires (i.e., DASS-21, PSWQ, and the full version of the RRS). Those who met eligibility criteria and elected to participate were randomized to one of the intervention conditions using a web-based randomization program, randomizer.org. Of note, no participants were identified as being at high risk for suicide, and thus none were excluded based on this criterion.

One laboratory visit included the above assessments as well as the one-time in-person instructions for completing the self-administered intervention (for those eligible and randomized to an intervention condition; see below).

Interventions

General procedures

After randomization, participants underwent a short training by a trained undergraduate-level research assistant in the laboratory with research staff. This training visit included the rationale for the intervention, an explanation/definition of rumination, and instructions for completing home practice, online nightly ratings, and the post-treatment assessment. The research assistant provided clear instructions for each condition using scripts, which were also provided to the participants via hardcopy before leaving the laboratory. Full scripts for each intervention are available by request. Participants underwent the intervention at home and completed nightly surveys each night for 3 weeks. No direct contact was made with the study team during this period, and no feedback was given by the study team to the participants regarding their responses to the nightly surveys or the completion of their interventions. At the end of the 3 weeks, participants were contacted for an online post-treatment assessment consisting of the same questionnaires given in the initial laboratory visit. Interventions are described below.

Save it for rumination time (SRT)

Participants worked with the experimenter during training to: (a) choose a place and time to ruminate on purpose each day for a 15 to 20 minute period; and (b) create a list of 5 simple, pleasurable, distracting activities. They were instructed to limit rumination to that 15 to 20-minute period only. If they found themselves exceeding the 15 to 20-minute allotted period for rumination or ruminating at other points during, they were instructed to notice and briefly observe their worry, and then divert their attention to an activity (e.g., working on homework, spending time with friends) using the list created in the laboratory. Participants were encouraged to attempt to decrease the length of their rumination time by 1 minute each day, though this was not a requirement. In sum, the general approach was that of stimulus control, drawing from behavioral activation strategies and with a “light touch” of the spirit of mindfulness.

Accept, limit, redirect (ALR)

This strategy was designed to be used to address rumination as it came up throughout the day. During training, participants were instructed to: (a) notice when they began to ruminate without judgement (“Accept”); (b) allow themselves up to 5 minutes to continue to ruminate (“Limit”); and (c) once that 5-minute period was over, shift attention away from ruminating by either returning to what they were doing or engaging in a distracting activity (“Redirect”). As in SRT, participants worked with the experimenter to create a list of simple, pleasurable, distracting activities. Participants were also instructed to attempt to reduce the amount the length of the rumination time by a minute each week, though that was not a requirement. In sum, this intervention drew from mindfulness/acceptance and behavioral activation approaches.

Self-monitoring control (SMC)

Participants were instructed to monitor their rumination and mood throughout the day and record it on the web-based nightly survey.

Statistical Analysis

One-way ANOVAs and χ² tests were used to examine baseline differences among the three groups. Variables in which there were significant differences among groups at baseline were planned to be included as covariates in subsequent analyses, but there were no differences among groups on any key baseline variables. Nightly ratings were used to determine the degree to which participants adhered to their self-directed interventions and were successfully able to implement the strategies. In SRT, the number of times a participant was successfully able to save rumination for rumination time per day was averaged across daily ratings for each participant, and in ALR, the number of times the participant was able to limit rumination to 5 minutes or less and the number of times the participant was able to redirect attention to another activity were averaged across nightly ratings for each participant. These averages were used to compute composite adherence variables for each of these two conditions. The SRT composite index was the average number of times the participant saved rumination for rumination time each day divided by the average number of times the participant noticed rumination each day. ALR composite indices included the average number of times the participant was able to limit rumination divided by the average number of times the participant noticed rumination, as well as the average number of times the participant was able to redirect attention to another activity divided by the number of times the participant noticed rumination. These indices thus provided a percentage of times that the SRT-treated participants successfully saved rumination for rumination time, and percentages in ALR for how often participants were able to limit rumination and redirect attention.

Outcome analyses on the primary measures (DASS-D, DASS-A, RRS, and PSWQ) were conducted with a series of 2 (time) × 3 (condition) repeated measures ANOVAs to examine change over time within each group and time x condition interactions. Additional tests at post-treatment using univariate ANCOVA (covarying for pre-treatment scores) were conducted with LSD post hoc tests to examine differences among the three conditions when the time x condition omnibus test was significant. Because (1) there were so few participants who did not complete a pre- and post-treatment assessment (i.e., 10 out of 60, or 16.66%, completed a pre-treatment but not post-treatment assessment), (2) we collected post-treatment data on as many participants as possible even if they did not complete their intervention (and included all available post-treatment data in analyses, regardless of number of intervention sessions completed), and (3) this study represents a preliminary test of these self-directed interventions targeting rumination, we report on the available data at pre- and post-treatment as our primary analyses. As a secondary set of analyses, we imputed missing values at post-treatment using maximum likelihood estimation and report those outcomes when they differ from the available data sample. We do not refer to this sample as a true “completers” sample because we included all available data regardless of how many nightly ratings were completed. Cohen’s d effect sizes were calculated and reported in Table 1 to provide information on the magnitude of effects.

Results

Baseline Descriptives and Differences among Groups

Baseline descriptives are reported in Table 1. Baseline scores across conditions on the DASS-D and DASS-A subscales were in the mild range for depression and anxiety (Crawford & Henry, 2003). The mean scores on the RRS and PSWQ were both >1 SD above the mean for a normative population (Brown et al., 1992; Nolen-Hoeksema et al., 1999). Thus, despite the use of a non-treatment seeking sample, participants were mildly symptomatic and a rough analogue of a sample that would benefit from, and be appropriate for, self-administered interventions (Van Straten et al., 2010).

Randomization was successful: there were no significant differences among groups on gender, age, medical condition status, medication status, or pre-treatment scores on any of the measures (all ps > 0.24). Table 2 shows the descriptive statistics for each variable by condition, including pre- to post-treatment effect sizes in each condition. Table 3 shows the correlations among variables.

Table 2.

Descriptive Statistics by Condition and Effect Sizes (Cohen’s d) for Pre- to Post-Treatment Effects by Condition.

Variable	Pre-treatment			Post-treatment
	SMC (N = 20)	ALR (N = 21)	SRT (N = 19)	SMC (N = 17)		ALR (N = 17)		SRT (N = 17)
	M (SD)	M (SD)	M (SD)	M (SD)	d	M (SD)	d	M (SD)	d
DASS-D	5.58 (4.62)	5.10 (3.53)	4.89 (3.68)	5.24 (4.70)	.07	4.47 (4.46)	.16	3.35 (2.60)	.50
DASS-A	4.42 (2.69)	5.05 (4.25)	6.05 (4.84)	3.29 (2.47)	.45	3.65 (3.59)	.37	3.12 (3.22)	.73
RRS	53.74 (14.40)	51.14 (10.88)	53.37 (13.33)	51.94 (14.35)	.13	48.88 (10.92)	.27	45.41 (16.46)	.54
PSWQ	61.89 (10.91)	56.62 (10.36)	60.21 (9.65)	56.88 (10.62)	.48	55.29 (12.52)	.12	48.12 (14.61)	.90

Note. d = Cohen’s d from pre- to post-treatment for that particular condition; M = mean; SD = standard deviation; DASS-D = Depression Anxiety Stress Scale-Depression Subscale; DASS-A = Depression Anxiety Stress Scale-Anxiety Subscale; PSWQ = Penn State Worry Questionnaire; RRS = Ruminative Response Scale; SMC = self-monitoring control; SRT = save it for rumination time condition; ALR = Accept, Limit, Redirect condition.

Table 3.

Correlations Among Pre- and Post-Treatment Variables.

Variable	1	2	3	4	5	6	7	8	Mean	SD	Range	α
(1) Pre DASS-Depression	1	—	—	—	—	—	—	—	5.19	3.90	0–16	.87
(2) Pre DASS-Anxiety	.60**	1	—	—	—	—	—	—	5.17	4.02	0–18	.8
(3) Pre PSWQ	.26*	.30*	1	—		—	—	—	59.47	10.39	33–76	.89
(4) Pre RRS	.59**	.47**	.32*	1	—	—	—	—	52.69	12.72	22–78	.91
(5) Post DASS-Depression	.73**	.35*	.04	.45**	1	—	—	—	4.35	4.02	0–18	.89
(6) Post DASS-Anxiety	.62**	.54**	.25	.47**	.70**	1	—	—	3.35	3.07	0–14	.78
(7) Post PSWQ	.23	.15	.66**	.15	.25	.37**	1	—	53.43	13.02	0–76	.89
(8) Post RRS	.53**	.29*	.17	.72**	.61**	.44**	.39**	1	48.75	14.07	0–79	.92

p < .05, **p < .01

Adherence to Intervention/Manipulation Check

Participants across all three conditions completed an average of 15.08 (SD = 6.55) of their nightly ratings (out of 21). There were no significant differences among groups on number of nightly session ratings completed [F (2,60) = 0.94, p = .397, partial η² = .03]. In SRT, 67% of participants reported that they were able to save rumination for scheduled rumination time across the study period. To explore whether this skill may have improved over time, the same index was calculated with just the final week of the self-directed intervention nightly ratings. During the last week, participants were able to save rumination for rumination time 66% of the time that they noticed themselves ruminating.

In ALR, on average across the intervention, participants were able to limit their rumination to <5 minutes per occasion 64% of the time and were able to redirect their attention to another distracting activity 72% of the time. When examining the last week’s ratings only (in order to assess whether this skill improved over time), ALR-treated participants were, on average, able to limit their rumination to <5 minutes per occasion 72% of the time, and were able to redirect their attention to another activity 79% of the time.

Symptom Outcomes

Primary: Depression

Using repeated measures ANOVA, the effect of time (across all three conditions together) was not statistically significant [F (1, 47) = 2.74, p = .104, partial η² = .06]. When examining each condition separately, there was a significant effect of time on DASS-D scores (reduction) from pre- to post-treatment for SRT, F (1, 16) = 6.44, p = .022, partial η² = .29, but the effect of time from pre- to post-treatment on DASS-D scores only approached significance in SMC [F (1, 15) = 3.46, p = .083, partial η² = .19] and ALR [F (1, 16) = 3.96, p = .064, partial η² = .20], respectively. The time x condition effect was statistically significant, F (2, 47) = 5.80, p = .006, partial η² = .20. Univariate ANCOVAs covarying for pre-treatment scores were conducted with post-treatment DASS-D as the dependent variable in order to examine inter-group differences using LSD post hoc tests at post-treatment. Not surprisingly, given the repeated measures ANOVA findings, there was significant effect of condition on post-treatment DASS-D scores, F (2, 49) = 5.19, p < .009, partial η² = .18. LSD post hoc tests to examine inter-group differences revealed that ALR and SRT both had significantly lower DASS-D scores at post-treatment than SMC (p = .027 for ALR v SMC and p = .003 for SRT v. SMC) after accounting for pre-treatment scores, with no significant differences in post-treatment DASS-D scores between SRT and ALR (p = .405). With the imputed sample accounting for missing data, the effect of condition effect was no longer statistically significant (p = .568).

As a potentially more meaningful index of the magnitude of within-group pre- to post-treatment changes than inferential statistics in a small sample, Cohen’s d effect sizes were calculated (separately from the above analyses). Effect sizes were negligible for SMC (d = 0.07), small for ALR (d = 0.16), and medium for SRT (d = 0.50).

Secondary: Anxiety

A parallel set of analyses were conducted with DASS-A scores as the dependent variable. Using repeated measures ANOVA, there was a significant effect of time from pre- to post-treatment across conditions, F (1, 47) = 11.47, p = .001, partial η² = .20. Within each group, there was a significant effect of time for SRT, F (1, 15) = 15.75, p = .001, partial η² = .50, indicating significant reduction in DASS-A scores from pre- to post-treatment. In contrast, there was no significant effect of time in either ALR [F (1.16) = 2.06, p = .170, partial η² = .11] or SMC [F(1,15) = 0.28, p = .604, partial η² = .02). The time x condition effect approached but did not attain statistical significance, F (2, 46) = 2.62, p = .074, partial η² = .11. Accordingly, a univariate ANCOVA to examine cross-sectional differences between groups at post-treatment while covarying for pre-treatment scores revealed no significant condition effect, F (2,50) = 1.11, p = .338, partial η² = .05], and LSD post hoc tests conducted within this analysis revealed no statistically significant differences between conditions on post-treatment DASS-A scores when covarying for pre-treatment DASS-A scores (all ps > 0.162). A similar pattern emerged with the imputed sample estimating missing data.

Within-group pre- to post-treatment Cohen’s d effect sizes were also calculated and were moderate for SMC and ALR (d = 0.45 and 0.37, respectively) and large for SRT (d = 0.73).

Treatment Process/Mechanism Measures

Primary: Rumination

Cohen’s d effect sizes from pre- to post-treatment were small for SMC and ALR (d = 0.13 and d = 0.27, respectively), and moderate for SRT (d = 0.54). The effect of time from pre- to post-treatment on the RRS across conditions approached significance in the 2 (time) × 3 (condition) repeated measures ANOVA, F (1, 47) = 3.85, p = .056, partial η² = .08. Pre- to post-treatment change (i.e., reduction in rumination) was significant for SRT, F (1, 16) = 9.17, p = .008, partial η² = .36, but was not significant for ALR or SMC (ps > 0.574). The time x condition effect was statistically significant, F (2, 47) = 4.28, p = .020, partial η² = .15. The condition effect in the univariate ANCOVA was statistically significant, F (2, 50) = 3.91, p = .027, η² = .15. LSD post hoc tests from the univariate ANCOVA with post-treatment RRS as the dependent variable and pre-treatment RRS as the covariate revealed that SRT outperformed SMC (p = .011) and ALR (p = .046) on RRS at post-treatment, with lower RRS scores in SRT than the other two conditions after accounting for pre-treatment scores. No significant differences were observed between SMC and ALR (p = .528). A similar pattern was observed with the imputed sample, but the condition effect in the ANCOVA was no longer statistically significant (p = .517).

Secondary: Worry

Cohen’s d effect sizes from pre- to post-treatment were moderate for SMC (d = 0.48), small for ALR (d = 0.12), and large for SRT (d = 0.90).The effect of time was significant across conditions from pre- to post-treatment on the PSWQ, F (1, 46) = 16.69, p < .001, partial η² = .27. There was significant effect of time within SRT, F (1, 15) = 17.54, p = .001, partial η² = .52, indicating significant decreases in PSWQ scores, but no significant effect of time from pre- to post-treatment on PSWQ within ALR [F (1,16) = 0.92, p = .353, partial η² = .05] or SMC [F (1,15) = 1.75, p = .206, partial η² = .10]. There was also a significant time x condition effect, F (2, 46) = 5.34, p = .003, partial η² = .22. The univariate ANCOVA with post-treatment PSWQ scores as the dependent variable (covarying for pre-treatment PSWQ) revealed a significant condition effect, F (1, 49) = 6.30, p = .004, η² = .22. Post hoc LSD tests conducted within the ANCOVA revealed that at post-treatment (while covarying for pre-treatment scores), SRT outperformed ALR (p = .003) and SMC (p = .005) on PSWQ scores (i.e., lower PSWQ scores at post-treatment), with no differences between ALR and SMC (p = .857). A similar pattern emerged with the imputed sample, with one difference: the effect of time from pre- to post-treatment within SMC was statistically significant (p = .037).

Discussion

This study examined the preliminary efficacy of self-directed behavioral strategies for managing rumination in a non-treatment seeking sample of undergraduate students with elevated scores on a gold standard self-report measure of rumination. Approximately two-thirds of participants who were instructed to save rumination for a prescribed rumination time reported successfully being able to do so. Participants who were instructed to engage in a somewhat more complex strategy involving mindfulness and attentional refocus to a pleasurable or distracting activity were able to develop this skill over the course of the 3 week intervention period, with approximately three-quarters of those assigned to this strategy reporting they were able to limit rumination and redirect their attention to another activity by the latter portion of the intervention period. These preliminary findings demonstrate the feasibility of utilizing these behavioral strategies as self-administered interventions without therapist contact.

With regard to preliminary efficacy of these approaches, our hypotheses were partially supported. The active conditions showed greater reduction in depressive symptoms from pre- to post-treatment compared to the self-monitoring control, but this finding did not generalize to differential reductions in anxiety symptoms. Only those who engaged in scheduled rumination time yielded within-group improvement in anxiety symptoms, but the between-group difference in anxiety reduction was not significant. With regard to the treatment target (i.e., rumination), participants who were instructed to schedule rumination time showed greater reductions on the Ruminative Response Scale than those in the other two conditions, and these effects generalized to future-oriented worries, as measured by the Penn State Worry Questionnaire.

Taken together, these findings suggest that (a) behavioral interventions that combine elements of self-monitoring, mindfulness, basic activity scheduling (in the context of attentional refocusing), and stimulus control can remain simple, straightforward, and feasible for self-administration; (b) these interventions may be promising for the reduction of rumination; and (c) in particular, scheduling rumination time and providing strategies for saving rumination for rumination time may be the most effective of the evaluated strategies, showing the most consistent improvement in symptoms compared to the other conditions. Indeed, moderate to large pre- to post-treatment Cohen’s d effect sizes were observed on all outcome measures in the SRT condition. This finding is consistent with treatments for generalized anxiety disorder that use stimulus control to schedule worry time (e.g., McGowan & Behar, 2013), and adds to the literature in that it is the first study to our knowledge to apply this principle to the scheduling of rumination. This diverges from common conventional clinical wisdom that it is therapeutic to schedule time for future-oriented worries, but may be contraindicated to have individuals ruminate intentionally on a daily basis. Indeed, some instructional sets specifically tell patients not to think about past events or engage in post-event rumination during worry time. This finding suggests that scheduled worry time as seen in treatment for generalized anxiety disorder can be applied to rumination as well.

Presumably, limiting the amount of time an individual ruminates should reduce depressive symptoms by breaking the cycle in which increased rumination leads to increased depressive symptoms, and perhaps greater inactivity, which then leads to more rumination. Confining rumination to a specific time and place may also “free up” other time to engage in more positively reinforcing activities, which should also exert an impact on depressive symptoms. Future research should explore the pathways by which rumination, activation, and depressive symptoms impact one another during the course of scheduled rumination time. However, these findings should be considered preliminary and should be replicated with large clinical samples with higher symptom severity. Also, since scheduled rumination had a significant impact on worry but did not significantly outperform other conditions on anxiety symptom reduction, although trended in this way, future research with larger samples can elucidate whether scheduled rumination time exerts a differential effect on anxiety compared to the other strategies.

Those assigned to an intervention that instructed individuals to notice their rumination, allow it for a brief period, and then redirect to a pleasurable or distracting activity showed pre- to post-treatment reductions in depressive symptoms (albeit with a small effect size), but no significant improvement in pre- to post-treatment rumination, and minimal outperformance of this condition over basic self-monitoring. Therefore, inconsistent with hypothesis, this approach did not appear to exert an effect on rumination. Thus, the mechanism by which this approach exerted its effect on depressive symptoms is unknown, and should be the subject of future research. Perhaps mindfulness-relevant mechanisms, such as increases in experiential willingness, may explain this symptom reduction. Also, the “Accept, Limit, Redirect” condition effects did not extend to secondary outcomes related to anxiety and worry, which suggests that it did not have a broad effect on preservative thinking and corresponding symptoms as seen in the scheduled rumination condition. One plausible explanation is that the 3-week period was not enough time for individuals to practice this skill and see significant improvement. Another possibility was that the nuanced approach of noticing and accepting rumination and then allowing it to go by redirecting attention did not work. Indeed, traditional mindfulness and acceptance approaches would simply encourage individuals to notice and accept rumination without any disengagement strategies. The directive to disengage from rumination by turning attention to one of the pleasurable activities may have unintentionally resulted in thought suppression, which has paradoxical effects (Wegner et al., 1987) and is in fact at odds with typical mindfulness approaches. Taken together, the relatively modest effects of this approach are in need of replication before claims can be made about its efficacy in reducing rumination and associated constructs.

Finally, the moderate effects of the self-monitoring control condition on anxiety and worry are worth noting. These findings suggest that simply monitoring rumination and mood may exert some effect, at least on anxiety. Indeed, self-monitoring is not typically thought of as a stand-alone behavioral treatment, but has a long history of demonstrating behavior change, presumably via increasing awareness (Korotitsch & Nelson-Gray, 1999) and has been shown to reduce depression in the early stages of adolescent depression (Kauer et al., 2012), a potentially similar sample to the one evaluated in this study. Studies with follow-up assessments are needed to ascertain whether these improvements would maintain over time.

The primary limitations of this preliminary investigation include the use of a small sample of undergraduate students, a relatively brief intervention period, and the lack of a follow-up assessment. This study may be underpowered to detect statistically small significant between-group effects among active interventions. Future research with larger, treatment-seeking samples is needed. Indeed, findings from a sample of mildly symptomatic undergraduates may not generalize to more severe, treatment seeking populations, who may require a multi-component treatment package or a larger dose of treatment. Relatedly, given that the sample had mild symptoms at baseline, floor effects and restriction of range are possible problems that may have led to an inability to reveal potentially larger effects. Relatedly, the promising Cohen’s d effect sizes observed represented within-group effect sizes, which are important and appropriate for identifying signals for efficacy in preliminary work, but are not as stringent as an exclusive focus on between-groups effects. Taken together, these limitations call for replication of this study in more severe, clinical samples. As a pilot study, this is a preliminary step. Although the brevity of the intervention is a strength from a cost-effectiveness and resource allocation perspective, larger doses of treatment may yield larger effects, as these are skills that individuals may build and benefit from over time. Importantly, these findings emerged in a small sample that included participants regardless of how many nights they recorded their use of the self-directed intervention, providing a stringent preliminary test of the utility of these self-directed strategies for reducing rumination. Including all participants regardless of adherence is a conservative approach from an efficacy perspective, but is in line with effectiveness research that aims to evaluate how interventions perform in the “real-world.” Self-directed interventions are subject to non-adherence issues, and our findings are promising in spite of potential non-adherence. Relatedly, as with any evaluation of self-directed interventions, it is difficult to ascertain the degree to which participants truly engaged in their interventions with fidelity; that is, our measure of their adherence is self-reported.

Additionally, null findings should be interpreted with caution in a small sample that likely has insufficient power to detect small differences between active treatment conditions. Replication is needed before drawing firm conclusions about the null findings. Notably, our secondary analyses on the imputed sample, in which a few of the statistically significant findings from the primary analyses were no longer statistically significant, provide an unusually highly conservative approach for a preliminary study. Also, we did not include a follow-up assessment, given the preliminary nature of this pilot study. Future studies should examine the longer-term impact of these interventions on rumination and related processes and symptoms. Finally, although the likelihood of assessor bias is extremely low given that all outcome assessments were conducted via online self-report measures, the assessors were not fully blinded to study design. Taken together, these findings are promising but are in need of replication.

This study provides a preliminary test of novel, brief, self-directed interventions for reducing rumination, a transdiagnostic process related to several manifestations of psychopathology. Interestingly, the simplest intervention, scheduled rumination time, was the most effective, and was feasible for self-administration, making it a promising, simple, and potentially cost-effective strategy for reducing rumination. Nonetheless, this study should be considered a preliminary test of an efficacy signal; future research is needed to provide more information about the potential efficacy of these brief, self-directed interventions for rumination in clinical samples, as well as their effect on related constructs such as depression.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Kate Wolitzky-Taylor

Haley Breland

Author Biographies

Kate Wolitzky-Taylor is an associate professor at UCLA, with expertise in the nature and treatment of anxiety and related disorders, with an emphasis on substance use disorder comorbidity and transdiagnostic processes.

Haley Breland is a former UCLA student and current project coordinator at UCLA, working under the direction of Dr. Wolitzky-Taylor at the UCLA Integrated Substance Abuse Programs.

Jaclyn Ross is a doctoral candidate at UCLA, who specializes in emotion dysregulation and behavior therapy.

Amy Sewart is an assistant professor at California State University- Dominguez Hills. She studies the nature and treatment of anxiety disorders.

References

Abbott

M. J.

Rapee

R. M.

(2004). Post-event rumination and negative self-appraisal in social phobia before and after treatment. Journal of Abnormal Psychology, 113(1), 136–144. https://doi.org/10.1037/0021-843X.113.1.136

Ahl

Mikocka-Walus

Gordon

Andrews

J. M.

(2013). Are self-administered or minimal therapist contact psychotherapies an effective treatment for irritable bowel syndrome (IBS): A systematic review. Journal of Psychosomatic Research, 75(2), 113–120.

Aldao

Nolen-Hoeksema

Schweizer

(2010). Emotion-regulation strategies across psychopathology: A meta-analytic review. Clinical Psychology Review, 30(2), 217–237. https://doi.org/10.1016/j.cpr.2009.11.004

Arch

J. J.

Craske

M. G.

(2008). Acceptance and commitment therapy and cognitive behavioral therapy for anxiety disorders: Different treatments, similar mechanisms? Clinical Psychology: Science and Practice, 15(4), 263–279. https://doi.org/10.1111/j.1468-2850.2008.00137

Borkovec

T. D.

Inz

(1990). The nature of worry in generalized anxiety disorder: A predominance of thought activity. Behaviour Research and Therapy, 28(2), 153–158.

Borkovec

T. D.

Wilkinson

Folensbee

Lerman

(1983). Stimulus control applications to the treatment of worry. Behaviour Research and Therapy, 21(3), 247–251. https://doi.org/10.1016/0005-7967(83)90206-1

Brown

T. A.

Chorpita

B. F.

Korotitsch

Barlow

D. H.

(1997). Psychometric properties of the Depression Anxiety Stress Scales (DASS) in clinical samples. Behaviour Research and Therapy, 35(1), 79–89.

Brown

T. A.

Antony

M. M.

Barlow

D. H.

(1992). Psychometric properties of the Penn State Worry Questionnaire in a clinical anxiety disorders sample. Behaviour research and therapy, 30(1), 33–37.

Caselli

Ferretti

Leoni

Rebecchi

Rovetto

Spada

M. M.

(2010). Rumination as a predictor of drinking behaviour in alcohol abusers: A prospective study. Addiction, 105(6), 1041–1048.

10.

Crawford

J. R.

Henry

J. D.

(2003). The Depression Anxiety Stress Scales (DASS): Normative data and latent structure in a large non-clinical sample. British Journal of Clinical Psychology, 42(2), 111–131.

11.

Cuijpers

Donker

van Straten

Andersson

(2010). Is guided self-help as effective as face-to-face psychotherapy for depression and anxiety disorders? A systematic review and meta-analysis of comparative outcome studies. Psychological Medicine, 40(12), 1943–1957.

12.

Cuijpers

Van Straten

Warmerdam

(2007). Behavioral activation treatments of depression: A meta-analysis. Clinical Psychology Review, 27(3), 318–326. https://doi.org/10.1016/j.cpr.2006.11.001

13.

Dimidjian

Hollon

S. D.

Dobson

K. S.

Schmaling

K. B.

Kohlenberg

R. J.

Addis

M. E.

Gallop

McGlinchey

J. B.

Markley

D. K.

Gollan

J. K.

Atkins

D. C.

(2006). Randomized trial of behavioral activation, cognitive therapy, and antidepressant medication in the acute treatment of adults with major depression. Journal of Consulting and Clinical Psychology, 74(4), 658–670. https://doi.org/10.1037/0022-006X.74.4.658

14.

Gardner

K. J.

Dodsworth

Selby

E. A.

(2014). Borderline personality traits, rumination, and self-injurious behavior: An empirical test of the emotional cascades model in adult male offenders. Journal of Forensic Psychology Practice, 14(5), 398–417.

15.

Gordon

K. H.

Holm-Denoma

J. M.

Troop-Gordon

Sand

(2012). Rumination and body dissatisfaction interact to predict concurrent binge eating. Body Image, 9(3), 352–357.

16.

Hofmann

S. G.

Sawyer

A. T.

Witt

A. A.

(2010). The effect of mindfulness-based therapy on anxiety and depression: A meta-analytic review. Journal of Consulting and Clinical Psychology, 78(2), 169–183. https://doi.org/10.1037/a0018555

17.

Jacobs

R. H.

Watkins

E. R.

Peters

A. T.

Feldhaus

C. G.

Barba

Carbray

Langenecker

S. A.

(2016). Targeting ruminative thinking in adolescents at risk for depressive relapse: Rumination-focused cognitive behavior therapy in a pilot randomized controlled trial with resting state fMRI. PLoS One, 11(11), e0163952.

18.

Jacobson

N. S.

Martell

C. R.

Dimidjian

(2001). Behavioral activation treatment for depression: Returning to contextual roots. Clinical Psychology: Science and Practice, 8(3), 255–270. https://doi.org/10.1093/clipsy.8.3.255

19.

Kauer

S. D.

Reid

S. C.

Crooke

A. H. D.

Khor

Hearps

S. J. C.

Jorm

A. F.

Sanci

Patton

(2012). Self-monitoring using mobile phones in the early stages of adolescent depression: Randomized controlled trial. Journal of Medical Internet Research, 14(3), e67.

20.

Kocovski

N. L.

Endler

N. S.

Rector

N. A.

Flett

G. L.

(2005). Ruminative coping and post-event processing in social anxiety. Behaviour Research and Therapy, 43(8), 971–984. https://doi.org/10.1016/j.brat.2004.06.015

21.

Korotitsch

W. J.

Nelson-Gray

R. O.

(1999). An overview of self-monitoring research in assessment and treatment. Psychological Assessment, 11(4), 415.

22.

Lovibond

P. F.

Lovibond

S. H.

(1995). The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behaviour Research and Therapy, 33(3), 335–343.

23.

McGowan

S. K.

Behar

(2013). A preliminary investigation of stimulus control training for worry: Effects on anxiety and insomnia. Behavior Modification, 37(1), 90–112.

24.

McLaughlin

K. A.

Nolen-Hoeksema

(2011). Rumination as a transdiagnostic factor in depression and anxiety. Behaviour Research and Therapy, 49(3), 186–193. https://doi.org/10.1016/j.brat.2010.12.006

25.

Martell

C. R.

Dimidjian

Herman-Dunn

(2010). Behavioral activation for depression. Guilford.

26.

Meyer

T. J.

Miller

M. L.

Metzger

R. L.

Borkovec

T. D.

(1990). Development and validation of the penn state worry questionnaire. Behaviour Research and Therapy, 28(6), 487–495.

27.

Molina

Borkovec

T. D.

(1994). The Penn State Worry Questionnaire: Psychometric properties and associated characteristics. In Davey

G. C. L.

Tallis

(Eds.), Wiley series in clinical psychology. Worrying: Perspectives on theory, assessment and treatment (pp. 265–283). John Wiley & Sons.

28.

Moshier

S. J.

Otto

M. W.

(2017). Behavioral activation treatment for major depression: A randomized trial of the efficacy of augmentation with cognitive control training. Journal of Affective Disorders, 210, 265–268.

29.

Nolen-Hoeksema

(1991). Responses to depression and their effects on the duration of depressive episodes. Journal of Abnormal Psychology, 100(4), 569–582.

30.

Nolen-Hoeksema

(2000). The role of rumination in depressive disorders and mixed anxiety/depressive symptoms. Journal of Abnormal Psychology, 109(3), 504–511. https://doi.org/10.1037/0021-843X.109.3.504

31.

Nolen-Hoeksema

Harrell

Z. A.

(2002). Rumination, depression, and alcohol use: Tests of gender differences. Journal of Cognitive Psychotherapy, 16(4), 391–402.

32.

Nolen-Hoeksema

Larson

Grayson

(1999). Explaining the gender difference in depressive symptoms. Journal of Personality and Social Psychology, 77(5), 1061–1072.

33.

Nolen-Hoeksema

Morrow

Fredrickson

B. L.

(1993). Response styles and the duration of episodes of depressed mood. Journal of Abnormal Psychology, 102(1), 20–28.

34.

Nolen-Hoeksema

Parker

L. E.

Larson

(1994). Ruminative coping with depressed mood following loss. Journal of Personality and Social Psychology, 67(1), 92–104.

35.

Nolen-Hoeksema

Wisco

B. E.

Lyubomirsky

(2008). Rethinking rumination. Perspectives on Psychological Science, 3(5), 400–424. https://doi.org/10.1111/j.1745-6924.2008.00088

36.

Olatunji

B. O.

Naragon-Gainey

Wolitzky-Taylor

K. B.

(2013). Specificity of rumination in anxiety and depression: A multimodal meta-analysis. Clinical Psychology: Science and Practice, 20(3), 225–257. https://doi.org/10.1111/cpsp.12037

37.

Robinson

M. S.

Alloy

L. B.

(2003). Negative cognitive styles and stress-reactive rumination interact to predict depression: A prospective study. Cognitive Therapy and Research, 27(3), 275–291.

38.

Smith

J. M.

Alloy

L. B.

Abramson

L. Y.

(2006). Cognitive vulnerability to depression, rumination, hopelessness, and suicidal ideation: Multiple pathways to self-injurious thinking. Suicide and Life-Threatening Behavior, 36(4), 443–454.

39.

Treynor

Gonzalez

Nolen-Hoeksema

(2003). Rumination reconsidered: A psychometric analysis. Cognitive Therapy and Research, 27(3), 247–259. https://doi.org/10.1023/A:1023910315561

40.

Van Straten

Seekles

Van ‘t Veer-Tazelaar

N. J.

Beekman

A. T.

Cuijpers

. (2010). Stepped care for depression in primary care: What should be offered and how? Medical Journal of Australia, 192, S36–S39.

41.

Watkins

Scott

Wingrove

Rimes

Bathurst

Steiner

Kennell-Webb

Moulds

Malliaris

(2007). Rumination-focused cognitive behaviour therapy for residual depression: A case series. Behaviour Research and Therapy, 45(9), 2144–2154. https://doi.org/10.1016/j.brat.2006.09.018

42.

Watkins

E. R.

Mullan

Wingrove

Rimes

Steiner

Bathurst

Eastman

Scott

(2011). Rumination-focused cognitive–behavioural therapy for residual depression: Phase II randomised controlled trial. The British Journal of Psychiatry, 199(4), 317–322. https://doi.org/10.1192/bjp.bp.110.090282

43.

Wegner

D. M.

Schneider

D. J.

Carter

S. R.

White

T. L.

(1987). Paradoxical effects of thought suppression. Journal of Personality and Social Psychology, 53(1), 5–13.

44.

Wells

Fisher

Myers

Wheatley

Patel

Brewin

C. R.

(2009). Metacognitive therapy in recurrent and persistent depression: A multiple-baseline study of a new treatment. Cognitive Therapy and Research, 33(3), 291–300.

45.

Wells

Fisher

Myers

Wheatley

Patel

Brewin

C. R.

(2012). Metacognitive therapy in treatment-resistant depression: A platform trial. Behaviour Research and Therapy, 50(6), 367–373.

46.

Wenzlaff

R. M.

Luxton

D. D.

(2003). The role of thought suppression in depressive rumination. Cognitive Therapy and Research, 27(3), 293–308.

47.

Werry

J. S.

Wollersheim

J. P.

(1967). Behavior therapy with children: A broad overview. Journal of the American Academy of Child Psychiatry, 6(2), 346–370.

48.

Wolitzky-Taylor

Sewart

Zinbarg

Mineka

Craske

M.G.

(in press). Rumination and worry as putative mediators of the association between emotional disorders and substance use disorders in a longitudinal investigation. Addictive Behaviors.