Rumination following bereavement: an overview

Abstract

This contribution provides an overview of rumination (i.e., thinking repetitively and recurrently about negative events and/or negative emotions) in adjustment to bereavement. First, we summarise a growing literature on rumination and mental health outcomes of bereavement. Next, we compare two main theories explaining the maladaptive effects of rumination after loss, which hold conflicting implications for clinical practice. The Response Styles Theory (RST) states that rumination is a maladaptive confrontation strategy that perpetuates distress by increasing negative cognitions, impairing problem solving and instrumental behaviour and reducing social support. Conversely, the Rumination as Avoidance Hypothesis (RAH) holds that rumination may serve to avoid painful aspects of the loss, thereby hampering adjustment to bereavement. Crucially, while RST predicts that distraction reduces rumination, RAH predicts that loss exposure is more effective. We review evidence for RST and RAH and their clinical implications and conclude with a brief exploration of ways to reconcile these theories.

Keywords:

• Rumination
• avoidance
• complicated grief
• prolonged grief disorder
• cognitive behavioural therapy
• counselling

Editorial comment: Colin Murray Parkes

Readers will be familiar with the lasting part played by patterns of attachment to parents in early childhood on attachments to others later in life. A distinguishing feature of separation distress and grief is the repeated pining and search for some way to recover the lost person. If, as attachment theory postulates, the search is part of a pattern of attachment behaviour that develops early in life, we can expect that children whose parents responded consistently to the child’s search will soon have learned when their crying and searching is likely to be rewarded, whereas children whose parents are inconsistent, sometimes rewarding and other times ignoring them, will learn to persist for much longer periods of time. Psychologists term this ‘intermittent reinforcement’ and it is my belief that it gives rise to the kind of persistent searching that Eisma and Stroebe call ‘rumination’ after bereavement.

In this paper, invited by the Editorial Board of Bereavement Care, the authors describe and test two theories and two methods of treatment for persistent rumination. Of the two, the Response Style Theory comes closest to the model described above.

Precisely a century ago, Sigmund Freud (1917/1957) observed that bereaved individuals often experience recurring thoughts about the deceased and the changed world in which they now live. He believed such thoughts to be an integral part of ‘grief work’, an adaptive process whereby people gradually come to terms with bereavement through repeated confrontation with the loss. Despite early clinical interest in the role of recurrent thought in adjustment to bereavement, it did not become the focus of systematic empirical attention until Nolen-Hoeksema’s work on rumination and depression in the early nineties. Nowadays, high levels of recurrent thought may be considered a hindrance in the process of coming to terms with the death of a loved one (i.e., chronic rumination may imply a person ‘gets stuck’ in one’s grief). Not surprisingly, then, it is regarded as a phenomenon that merits fine-grained investigation by researchers, as part of their endeavor to inform practitioners how best to support those bereaved who experience severe distress after loss.

In the present paper, our aim is therefore to provide an overview of rumination in adjustment to bereavement, summarising research that has been conducted over the span of three decades. We review definitions of ruminative coping, consider its negative and positive consequences, and evaluate the evidence for two main theories used to understand the effects of rumination after bereavement. Interestingly, as we shall see, these theories hold opposing ideas about how ruminative thought leads to poorer adaptation to bereavement. We conclude by presenting the potential implications of this body of research for professionals and volunteers working with bereaved persons.

Rumination

Broadly defined, rumination is the process of thinking repetitively and/or recurrently about the causes and consequences of negative events and/or negative emotions (Michael, Halligan, Clark & Ehlers, 2007; for a review: Nolen-Hoeksema, 2001). There is a remarkably diverse literature on rumination, and ideas on what rumination is have changed over the past decades¹, which makes it necessary to set a clearly-defined goal for the present review. Specifically, we aim to provide an overview of current understanding of the most studied conceptualisations and theories of ruminative coping following bereavement.

Depressive rumination

As mentioned, rumination research in bereaved samples was first done by Nolen-Hoeksema and colleagues who aimed to understand how people cope with depressed mood. Therefore, the first type of rumination that was investigated systematically in bereaved individuals was depressive rumination, which involves repetitively and passively focusing on one’s depressive symptoms and on the possible causes and consequences of these symptoms (e.g., Eisma et al, 2015a, 2014a; Torges, Stewart & Nolen-Hoeksema, 2008; Nolen-Hoeksema, 2001; Nolen-Hoeksema, McBride & Larson, 1997; Nolen-Hoeksema Parker & Larson, 1994). The most-used questionnaire to assess the extent to which people engage in depressive rumination is the Ruminative Response Scale of the Response Style Questionnaire (RRS-RSQ: Nolen-Hoeksema & Morrow, 1991). The RRS assesses the types of thoughts people typically experience when they are sad, blue or depressed. Typical depressive ruminative thoughts are: ‘Why do I have problems other people don’t have’? and ‘I think about my feelings of fatigue and achiness’.

Notably, the RRS has been criticised for content overlap with depressive symptoms. Therefore, based on the original RRS, two 5-item subscales were developed. First, the brooding subscale assesses (presumed) maladaptive comparisons of one’s current situation with some unachieved standard (e.g., ‘Why do I always react this way’?). Second, the reflection subscale, assesses (presumed) adaptive purposeful turning inward to engage in cognitive problem solving to alleviate one’s depressive symptoms (e.g., ‘I analyse recent events to understand why I feel so depressed’) (Treynor, Gonzales & Nolen-Hoeksema, 2003). The brooding and reflection subscales of the RRS have been shown to be reliable and valid ways of assessing subtypes of depressive rumination (Schoofs, Hermans & Raes, 2010; Treynor et al, 2003).

Engaging in depressive rumination has a negative impact on how people adjust to their loss. For example, ruminators generally experience higher levels of depression, anxiety, posttraumatic stress, and complicated grief symptoms, and more general distress (Eisma et al, 2015a, 2014a, 2012; Morina, 2011; Ito et al, 2003; Nolen-Hoeksema et al, 1997, 1994). There is also some evidence that depressive rumination actually leads to poorer adaptation to bereavement over time. For example, in one study Nolen-Hoeksema and colleagues (1994) showed that ruminating more after losing a family member predicted higher levels of depression 5 months later, even after controlling for depressive symptoms and social support at baseline. There is less evidence for the maladaptive effects of brooding and the adaptive effects of reflection after loss, with no recent studies showing a longitudinal effect of brooding on mental health (e.g., Boelen, Reijntjes & Smid, 2016; Eisma et al, 2015a, 2012), and only one study suggesting that reflection may be beneficial after bereavement (Eisma et al, 2015a). This investigation showed that engaging in more reflection on one’s feelings led to lower symptom levels of complicated grief and depression over the course of a year, even after controlling for baseline symptoms.

Grief rumination

More recently, researchers have become interested in grief rumination, repetitive and recurrent thinking about causes and consequences of the loss and loss-related emotions (e.g., Eisma et al, 2015a, 2014b, 2013, 2012; van der Houwen, Stroebe, Schut, Stroebe & van den Bout, 2010; Boelen & van den Hout, 2008; Boelen, van den Bout & van den Hout, 2006). As argued by Eisma and colleagues (2014b), grief rumination could be more relevant than depressive rumination after bereavement for two main reasons. First, rumination is commonly focused on discrepancies between the world as it is, and the world as one would like it to be (e.g., Martin & Tesser, 1996). After bereavement, the most important discrepancy is not that one feels depressed but wants to feel happy, but the fact that the one’s loved one is dead and not alive. Therefore, a person will be more likely to ruminate about causes and consequences of the loss-event instead of the causes and consequences of depressive symptoms. Second, if a bereaved individual does think repetitively about one’s negative emotions, there are many different emotional experiences after loss that one may focus on rather than depressive feelings, including yearning, guilt, anger, loneliness or anxiety (e.g., O’Connor & Sussman, 2014; Li, Stroebe, Chan & Chow, 2013; Maciejewski, Zhang, Block & Prigerson, 2007; Nolen-Hoeksema, 2001).

What then, specifically characterises grief rumination? Typically, grief rumination may consist of so-called ‘counterfactual thinking’ about events leading up to the death, that is, imagining alternative past realities in which the person would not have died (e.g., ‘Could I have done something different so that the loss could have been prevented’?), ruminative thinking about the unfairness of the loss (e.g., ‘I wondered why it happened to me and not someone else’), the meaning of the loss (e.g., ‘I analyse what the personal meaning of the loss is for me’), one’s emotional reactions to the loss (e.g., ‘I ask myself whether I react normally to the loss’) and others’ responses to the loss (e.g., ‘I think about how I would like other people to respond to my loss’). Our research group has recently developed a Dutch and English version of a scale to assess grief rumination, the Utrecht Grief Rumination Scale (UGRS: Eisma et al, 2014a). Multiple studies have provided strong support for the reliability and validity of the UGRS (Eisma et al, 2015a, 2014a, 2012).

Similar to depressive rumination, more grief rumination is concurrently and longitudinally associated with more symptoms of depression, anxiety, posttraumatic stress and complicated grief (Eisma et al, 2015a, 2014a, 2013, 2012; van der Houwen et al, 2010; Boelen & van den Hout, 2008; Boelen et al, 2006). Moreover, grief rumination has repeatedly been shown to be a better longitudinal predictor of mental health problems after bereavement than depressive rumination (Eisma et al, 2015a, 2014a, 2013, 2012). However, this does not imply that all forms of grief rumination are maladaptive. Specifically, in a longitudinal investigation of grief rumination subtypes it was demonstrated that rumination about understanding one’s emotional reactions longitudinally predicted the experience of less severe complicated grief symptoms (Eisma et al, 2015a).

Theories of rumination in adjustment to bereavement

We have so far demonstrated that rumination can have a negative impact on recovery after bereavement (although there are some preliminary clues that some types of ruminative thought may have positive consequences). So, the next step is to understand how this thought process works. This is important to establish, as it may guide the development and improvement of interventions for people with high levels of rumination and loss-related distress. As mentioned, there are two different – one might say competing – theories that have been systematically investigated in the context of bereavement. The first theory is known as the Response Styles Theory (RST: Nolen-Hoeksema, 2001) and the second as the Rumination as Avoidance Hypothesis (RAH)) (Stroebe et al, 2007). We will now consider these two theories, their evidence base, and their clinical implications.

Response Styles Theory (RST)

A key assumption of the RST is that rumination leads to poorer adaptation to bereavement because it serves as a confrontation strategy, whereby people repeatedly think about their loss-related emotions. In the recent past, rumination has even been considered ‘the opposite to denial and suppression’ (Nolen-Hoeksema, 2001; see also: Michael & Snyder, 2005; Tait & Silver, 1989; but see Nolen-Hoeksema, Wisco & Lyubomirsky, 2008, for a more nuanced viewpoint). Such recurrent focus on negative emotions is assumed to prolong distress in four ways. First, rumination may worsen one’s mood, by increasing accessibility of negative cognitions about oneself, the world and the future. Second, rumination may make problem solving less effective, in part because people are thinking more negatively about themselves and their lives. Third, rumination may impair engaging in activities that could improve one’s situation. For example, ruminative coping could take up time and increase feelings of hopelessness, making people less inclined to engage in activities that might lift their mood and increase their sense of control. Fourth, repeatedly going over ruminative thoughts with friends and family members may make people less inclined to provide social support, thereby increasing ruminators’ depressed mood.

Many of the basic assumptions of RST have been supported in research studies in bereaved and non-bereaved samples (for a review: Nolen-Hoeksema et al, 2008). For example, in a recent investigation, it was found that rumination indeed reduced involvement in social, occupational and recreational activities, leading to more depressive symptoms over the course of a year (Eisma et al, 2013). In a similar vein, bereaved persons who ruminate more were shown to reach out for social support more but at the same time be less satisfied with the support that they received (Nolen-Hoeksema & Davis, 1999). This supports a link between rumination and inadequate support by friends and family members.

Rumination as Avoidance Hypothesis (RAH)

A fundamentally different theory was put forward by Stroebe and colleagues (2007) who proposed their Rumination as Avoidance Hypothesis (RAH). Instead of assuming ruminative coping to be a maladaptive confrontation mechanism, RAH holds that (chronic) rumination could serve as cognitive avoidance and may thereby lead to poorer emotional adaptation to bereavement. For instance, someone might repeatedly engage in counterfactual thinking (i.e., ‘if-then statements’ such as ‘If only he had taken his medicine, he would still be alive’) to avoid thinking about the permanence of the separation with the deceased. This may have negative consequences because such avoidance could interfere with acceptance of the loss and/or could impede integration of one’s personal memories about the loss with existing memories, thereby fueling grief complications (Stroebe et al, 2007; cf. Boelen et al, 2006).

Around the time that this theory was first published, there was very little research on this topic. In the years to follow, survey studies in non-bereaved samples clearly supported a link between rumination on the one hand and cognitive avoidance (e.g., thought suppression) and emotional avoidance on the other hand (e.g., Dickson, Ciesla & Reily, 2012; Liverant, Kamholz, Sloan & Brown, 2011; Giorgio et al, 2010; Moulds, Kandris, Starr & Wong, 2007; Cribb, Moulds & Carter, 2006; Wenzlaff & Luxton, 2003). There was also one study that demonstrated that widows who ruminate more also show a stronger tendency to engage in experiential avoidance (i.e., avoidance of internal experiences such as emotions, bodily sensations, thoughts, and memories) (Morina, 2011). Nevertheless, there were no longitudinal studies in bereaved individuals that examined if avoidance processes could explain the relation between rumination and mental health problems. Moreover, there was a clear need for investigations examining an association between rumination and behavioural (as opposed to self-reported) measures of (loss) avoidance.

To fill this gap in knowledge, we conducted three studies. In the first study, we set out to test whether avoidance processes could explain why ruminative coping leads to mental health problems. We conducted a three-wave longitudinal survey in over 300 recently bereaved people (Eisma et al, 2013). A main finding was that experiential avoidance at 6 months fully explained the relationship between grief rumination and complicated grief symptoms at 12 months, even after controlling for baseline symptoms. Additionally (after removing experiential avoidance from the model) we found that thought suppression could similarly explain the relationship between rumination and complicated grief symptoms. This study thereby provided the first evidence that cognitive avoidance can explain the relationship between rumination and mental health problems following loss. It also suggested that ruminative thoughts might serve to suppress more threatening loss-related thoughts.

Next, we wanted to investigate the link between grief rumination and behavioural measures of avoidance (Eisma et al, 2015b, 2014b). In psychology, the study of attention is an important and often used method to assess approach and avoidance tendencies for specific cues. In this specific study, we wanted to assess high and low ruminators’ attention for cues that represent the separation with the deceased (‘loss-reality cues’) and stimuli that do not signal this separation. Using eye-tracking, an advanced research method, we could assess participants’ attention patterns for different paired combinations of pictures (deceased or stranger) and words (loss-related, negative, neutral) on multiple 10-second trials. A main finding was that people who ruminate more looked less at pictures of the deceased, only when this picture was combined with a loss-related word (e.g., death, loss). This effect was maintained even after controlling for concurrent levels of depression, post-traumatic stress and complicated grief, supporting a unique association between rumination and attentional loss avoidance (Eisma et al, 2014b).

In our third study, we conducted an Approach Avoidance Task (AAT) to assess automatic avoidance tendencies for similar cues as used in the eye-tracking investigation (Eisma et al, 2015b). In our AAT, people pushed picture-word cues on a computer screen away from themselves or pulled these towards themselves as fast as possible based on an irrelevant characteristic of the picture (i.e., the colour of the word), using a joystick (shrinking or enlarging stimuli on screen, respectively) (Rinck & Becker, 2007). If participants are faster at pushing a cue away and slower at pulling it towards themselves, this is assumed to indicate automatic avoidance of a specific cue. In our investigation among 72 bereaved individuals, we found that participants who engaged in more grief rumination showed stronger automatic avoidance of pictures of the deceased combined with a loss word, but not for other types of cues. This effect was maintained even after controlling for symptoms of depression and post-traumatic stress, but not when controlling for complicated grief symptoms.

In summary, these findings appear to imply that ruminative thinking after bereavement is linked with cognitive and emotional avoidance and specifically with avoidance of painful aspects of the loss, such as the separation with the deceased. These results thereby suggest that rumination may be targeted effectively with exposure techniques. In the final part of our project we therefore aimed to test the clinical implications of RST versus RAH.

Clinical interventions to target rumination after loss: RST versus RAH

As mentioned, RST and RAH hold contrasting ideas about how ruminative coping influences adaptation to bereavement. While RST assumes rumination to be a maladaptive confrontation process, RAH conceptualises it as an avoidance process. These different notions result in diverging recommendations for the treatment of bereaved people experiencing high levels of rumination and loss-related distress. According to RST, the ruminative cycle can best be broken by letting people distract themselves from their recurring thoughts. One way of doing so is by encouraging someone to increase the number of rewarding activities they undertake. This will not only reduce the time one has available to ruminate, but could also increase positive mood and disconfirm negative cognitions, thereby providing an ‘antidote’ to the effects of rumination (e.g., Papa, Rummel, Garrison-Diehn & Sewell, 2013; Nolen-Hoeksema et al, 2008; Martell, Addis, & Jacobson, 2001). Conversely, according to RAH, the best treatment strategy for rumination and loss-related distress would be to systematically confront someone with avoided aspects of the loss (Stroebe et al, 2007).

To test these competing ideas, we recently conducted a treatment effectiveness study among 47 bereaved individuals with elevated levels of grief rumination and complicated grief (Eisma et al, 2015c). Participants were randomised to 6 weeks of e-mail delivered ‘behavioural activation’, 6 weeks of e-mail delivered ‘exposure therapy’ or a waitlist control group.

Behavioural activation was based on a protocol for depressed patients (Lejuez, Hopko, Acierno, Daughters and Potago, 2011) that was shortened and adapted for administration via e-mail among bereaved persons. In the behavioural activation treatment, participants kept a 7-day activity diary, in which they indicated how pleasurable and important they found the activities they undertook. Next, participants were encouraged to continue keeping a diary, identify life-values they find important (e.g., spending quality time with family), and develop new meaningful and pleasurable activities based on these values (e.g., visit one’s mother every week). Ultimately, the intervention aim was to gradually engage in more value-based activities, while reducing the number of activities that were experienced as unimportant and unfulfilling.

Exposure therapy was based on a treatment protocol for face-to-face exposure therapy for complicated grief, which was adapted for online administration (Boelen et al, 2007). A first assignment for all participants was to list situations, objects, or memories related to the loss that they tended to avoid in daily life. Next, participants were encouraged to gradually expose themselves to those aspects of the loss that they tended to avoid most. This was achieved by engaging in a combination of writing assignments and imaginal or real-life exposure exercises. For example, people who avoided a memory about a stressful event leading up to the death of their loved one were asked to describe the event and the emotions evoked by this event in a realistic, detailed manner. Next, they were asked to engage in a guided imagination exercise in which they repeatedly imagined the event and experienced the distress evoked by remembering the event, until distress dissipated.

Intriguingly, this study provided some evidence for the effectiveness of both behavioral activation and exposure therapy. A main finding was that both treatments strongly reduced grief rumination and complicated grief symptom levels at post-test and 3 month follow-up. However, when comparing these results with results obtained using a different analysis technique, this showed that the findings were more robust for exposure therapy than for behavioural activation. Additionally, exposure was judged to be more acceptable and feasible by participants and appeared to lead to fewer people dropping out from the treatment. In support of RAH, online exposure, but not online behavioural activation, turned out to be the most viable treatment strategy for rumination and complicated grief.

Interpreting these findings within a broader treatment literature, evidence is accumulating that both behavioural activation and exposure treatments may be effective in ameliorating repetitive thought and related mental health problems (e.g., Chen, Liu, Rapee & Pillay, 2013; Papa et al, 2013; Wisco, Sloan & Marx, 2013; Watkins et al, 2011). For example, rumination-focused treatment for depression, an evidence-based treatment for rumination and depression, includes behavioural activation techniques as part of an integrated treatment model (Watkins et al, 2011). In line with our findings on exposure therapy, Wisco and colleagues (2013) demonstrated that written exposure treatment for post-traumatic stress disorder also ameliorates ruminative coping.

Conclusion – implications for counsellors

As Freud noted a century ago, bereaved individuals often experience recurring thoughts about the deceased and the changed world in which they now live. Our review has highlighted the importance of such recurrent thought by providing a comprehensive review of rumination after bereavement. Much of the research we have discussed was conducted in controlled study settings and the mentioned treatment effectiveness research evaluated protocolled interventions delivered by mental health professionals. How these findings generalise to other settings is difficult to assess. Applying this knowledge in a counselling context – for assessment and for treatment – should be therefore be done with caution.

Regarding the assessment of rumination: identification of problematic forms of ruminative coping is difficult. While persistently high levels of rumination have been shown to lead to poorer adaptation to bereavement, some rumination after loss is a natural response to a major negative life-event (Stroebe et al, 2007; Nolen-Hoeksema, 2001). We therefore emphasise: preventing any rumination is not being advocated here. Moreover, although counsellors could potentially use the newly developed UGRS to assess grief rumination in clients (Eisma et al, 2014a), it needs to be taken into account that the UGRS was developed for research in groups; any individual’s score on this scale should only be interpreted as a rough indication of the extent to which a person engages in grief rumination. Regarding intervention, in brief: counsellors who are interested in applying behavioural activation or exposure techniques should ideally do so using protocolled treatment manuals, under supervision from certified health professionals with a background in cognitive-behavioural therapy. In order to facilitate this process, we are now working on translating and publishing our exposure treatment manual.

Putting findings in a broader perspective: the research described above illustrates that the two main theories on rumination in adjustment to bereavement, RST and RAH, are both – to some extent – valid. One might say that rumination is per definition a confrontational process, as ruminators repeatedly focus on their thoughts about the causes and consequences of (feelings about) the loss. However, by ruminating, bereaved individuals also appear to distract themselves from the most painful aspects of the loss, such as the permanence of the separation with the deceased. Our review leads to the optimistic conclusion that there may therefore be more than one way to target ruminative coping following bereavement. On the one hand, rumination is a cognitive process that may be disrupted by engaging in value-based, rewarding activities; on the other hand, rumination is an avoidance strategy that can be ameliorated by systematic exposure to avoided aspects of the loss. Both these processes can be taken into account in efforts to alleviate the suffering of those who relentlessly ruminate about the death of a loved person. 

Notes

¹ Notably, some early theorists characterised rumination after negative life-events as the experience of intrusive memories (e.g., Michael & Snyder, 2005; Bodnar & Kiecolt-Glazer, 1994; Tait & Silver, 1989). However, interviews with people who have experienced a major negative life-event have shown that intrusive memories primarily consist of brief sensory experiences, whereas rumination is primarily a verbal thought process of longer duration (Speckens, Ehlers, Hackmann, Ruths & Clark, 2007). Because of these phenomenological differences between both constructs, we believe that distinguishing rumination from intrusive memories is theoretically important. Therefore, this review only considers research that exclusively defines repetitive thought as a (predominantly) verbal thought process.