Expressive writing as a therapeutic intervention for people with advanced disease: a systematic review

Background

Expressive writing involves writing about stressful or traumatic experiences. Despite trials in people with advanced disease, no systematic review to date has critiqued the evidence on expressive writing in this population. To synthesise the evidence of the effects of expressive writing on pain, sleep, depression and anxiety in people with advanced disease.

Methods

A systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. CINAHL, CENTRAL, PsycINFO and PubMed were searched from January 1986 to March 2018. Other sources included clinical data registers and conference proceedings. Studies were included if they were randomised controlled trials that assessed the impact of an intervention involving expressive writing for adults with advanced disease and/or studies involving linguistic analysis on the expressive writing output. Methodological quality was assessed using the Cochrane risk of bias tool and the Mixed Methods Appraisal Tool. The Grading of Recommendations Assessment, Development and Evaluation tool was used to assess the level of evidence for the outcomes of interest. The protocol of this systematic review has been registered on PROSPERO (CRD42017058193).

Results

Six eligible studies with a total of 288 participants were identified, including four randomised controlled trials. All of the trials were in cancer and recruited predominantly women. None of the interventions were tailored to the population. Studies had methodological shortcomings and evidence was generally of low quality. Combined analysis of the four trials, involving 214 participants in total, showed no clear difference in the effect of expressive writing on sleep, anxiety or depression compared to an active control. Pain was not evaluated in the trials. In contrast, analysis of the four studies that included linguistic analysis alluded to linguistic mechanisms for potential effects.

Conclusion

Although the trial results suggest there is no benefit in expressive writing for people with advanced disease, the current evidence is limited. There is a need for more rigorous trials. It would be of benefit first to undertake exploratory research in trial design including how best to measure impact and in tailoring of the intervention to address the specific needs of people with advanced disease.

Trial registration

The protocol of this systematic review has been registered on PROSPERO, which can be accessed here (registration number: CRD42017058193).

Electronic supplementary material

The online version of this article (10.1186/s12904-019-0449-y) contains supplementary material, which is available to authorized users.

The impact of an emotionally expressive writing intervention on eating pathology in female students

Introduction: Previous research demonstrating emotional influences on eating and weight suggest that emotionally expressive writing may have a significant impact on reducing risk of eating pathology. This study examined the effects of writing about Intensely Positive Experiences on weight and disordered eating during a naturalistic stressor.

Method: Seventy-one female students completed an expressive or a control writing task before a period of exams. Both groups were compared on BMI (kg/m²) and the Eating Disorder Examination – Questionnaire (EDE-Q) before the writing task and at 8-week follow-up. A number of secondary analyses were also examined (to identify potential mediators) including measures of attachment, social rank, self-criticism and self-reassurance, stress and mood.

Results: There was a significant effect of intervention on changes in the subscales of the EDE-Q (p = .03). Specifically, expressive writers significantly reduced their dietary restraint while those in the control group did not. There was no significant effect of the intervention on changes in BMI or the other subscales of the EDE-Q (Eating, Weight and Shape Concern). There was also no effect of writing on any of the potential mediators in the secondary analyses.

Discussion: Emotionally expressive writing may reduce the risk of dietary restraint in women but these findings should be accepted with caution. It is a simple and light touch intervention that has the potential to be widely applied. However, it remains for future research to replicate these results and to identify the mechanisms of action.

Affect systems, changes in body mass index, disordered eating and stress: an 18-month longitudinal study in women

Background: Evidence suggests that stress plays a role in changes in body weight and disordered eating. The present study examined the effect of mood, affect systems (attachment and social rank) and affect regulatory processes (self-criticism, self-reassurance) on the stress process and how this impacts on changes in weight and disordered eating. Methods: A large sample of women participated in a community-based prospective, longitudinal online study in which measures of body mass index (BMI), disordered eating, perceived stress, attachment, social rank, mood and self-criticism/reassurance were measured at 6-monthly intervals over an 18-month period. Results: Latent Growth Curve Modelling showed that BMI increased over 18 months while stress and disordered eating decreased and that these changes were predicted by high baseline levels of these constructs. Independently of this, however, increases in stress predicted a reduction in BMI which was, itself, predicted by baseline levels of self-hatred and unfavourable social comparison. Conclusions: This study adds support to the evidence that stress is important in weight change. In addition, this is the first study to show in a longitudinal design, that social rank and self-criticism (as opposed to self-reassurance) at times of difficulty predict increases in stress and, thus, suggests a role for these constructs in weight regulation.

Establishing a useful distinction between current and anticipated bodily shame in eating disorders

Previous research has suggested a role for bodily shame in the development of bulimia. The purpose of the present study was to extend this research by exploring a temporal perspective on bodily shame and eating pathology. Specifically, we were interested in whether bodily shame associated with the possibility of future weight gain was important in determining eating disorder symptoms independently of any association with bodily shame associated with current body size. A brief questionnaire designed to measure bodily shame was developed for the purposes of this study and administered to four samples (total n=428) of eating disordered and non-eating disordered women who also completed a number of measures of eating pathology. Factor analysis of the bodily shame scale identified three sub-scales, two measuring feelings of bodily shame (one measuring shame associated with current body size and one measuring shame that is anticipated should the individual gain weight) and a third measuring the perceived unattractiveness of being overweight. Anticipated bodily shame made a significant additional contribution to predicting eating disorder symptoms over and above that made by current bodily shame. A focus on feelings of bodily shame as they are currently experienced may limit the usefulness of this construct in eating disorder research. Since shame can be both punishing and prohibitive, consideration of the anticipation of shame as a consequence of weight gain may be a useful addition to understanding eating disordered behaviours, particularly in relation to symptoms concerning the prevention of weight gain rather than just the achievement of weight loss.

Shame, depressive symptoms and eating, weight and shape concerns in a non-clinical sample

Shame has been shown to be related both to symptoms of depression and eating pathology. However, the independence of this relationship has not yet been established. The purpose of the present study was to determine whether the relationship between shame and eating disorder symptoms was independent of the relationships of these variables with depression. Seventy non-clinical female participants completed measures of eating disorder-related concerns using the Eating Disorder Examination-Questionnaire version (EDE-Q), depressive symptoms using the Beck Depression Inventory (BDI-II) and two measures of shame, the Other As Shamer Scale (QAS) and the Test of Self-Conscious Affect (TOSCA). Despite a strong association between BDI-II and EDE-Q scores and a moderate relationship between the shame measures, the two measures of shame showed some specificity in their relationships with symptom measures. The OAS was independently related to levels of BDI-II scores while the TOSCA was independently related to scores on the EDE-Q. There are a number of differences between the two measures of shame used in this study. The fact that each was differentially related to eating concerns and depressive symptoms may give clues as to which aspects of shame are important in each of the two types of pathology.

Specificity in the relationship between depressive and eating disorder symptoms in remitted and nonremitted women

OBJECTIVE

There is a strong association between eating disorders and depression. However, because both eating disorder symptoms and depression are multifactorial, this study explored the relationship between these two disorders in women with eating disorders and women in remission.

METHOD

Two hundred and eight (mostly female) volunteers with a history of eating disorders participated. They completed a self-report questionnaire of eating disorder symptoms, the Short Evaluation for Eating Disorders (SEED), and a questionnaire measuring depression, the Beck Depression Inventory (BDI).

RESULTS

According to the SEED, 57 volunteers were classified as being in remission and 151 were classified as being ill. Those who were in remission were significantly less depressed overall than those who were still ill with 72% of the former falling in the "not depressed" or "mildly depressed" categories and 73% of the latter falling in the "moderately" or "severely depressed" categories. Factor analyses of the SEED and BDI identified three subscales of eating disorder symptoms (dietary restriction, bulimia, and body mass index [BMI]/menstruation) and two subscales of depression (cognitive and somatic/affective). Dietary restriction and bulimia, but not BMI/menstruation, were uniquely associated with the cognitive symptoms of depression. However, none of the eating disorder symptoms were uniquely associated with the somatic/affective symptoms of depression.

DISCUSSION

Although eating disorders and depression share considerable comorbidity, a specific association is restricted to that between the cognitive and behavioral symptoms of eating disorders and the cognitive symptoms of depression.

Differential environmental factors in anorexia nervosa: a sibling pair study

OBJECTIVES

Previous studies have explored differences in psychosocial and familial factors between women who develop anorexia nervosa and those who do not. However, these studies have generally used between-group comparisons. This study looks at the environmental factors which may be antecedents of anorexia nervosa looking at sister pairs where one had anorexia nervosa and the other did not.

DESIGN

A paired design was used to compare anorexic women with an unaffected sister on a number of background variables, including sibling interaction, parental care, peer group characteristics and other events unique to the individual.

METHODS

The Sibling Inventory of Differential Experience (SIDE) was used to determine non-shared environment. Out of an initial sample of 148 women with past or current anorexia nervosa, 28 were identified who had sisters with no reported history of eating disorders and who also consented to complete the questionnaire.

RESULTS

Anorexic sisters perceived more maternal control and reported more antagonism towards and jealousy of their sisters than did unaffected sisters. In addition, anorexic women reported having had fewer friends and boyfriends than their sisters.

CONCLUSIONS

These results confirm the perceived differences in background environment between women with and women without anorexia nervosa. These issues are discussed in relation to behavioural genetics, family dynamics and psychosexual development.

Disgust sensitivity in eating disorders: a preliminary investigation

BACKGROUND

Disgust is a basic emotion that has been relatively neglected in psychiatry in general and in eating disorders in particular. Nevertheless, there are features of disgust and its more complex derivatives (e.g., shame) which suggest that disgust may have a role to play in eating disorders.

METHOD

Seventy-four patients with a DSM-IV diagnosis of anorexia nervosa, bulimia nervosa, eating disorder not otherwise specified, and obese binge eater were compared with 15 control subjects on their levels of disgust sensitivity.

RESULTS

Overall, eating disorder patients did not appear to be more sensitive to disgust-eliciting stimuli than comparison subjects, although there was a tendency for patients to be more disgusted by body products. However, drive for thinness and bulimia scores were related to higher levels of disgust sensitivity to food, death, and magical contagion. General psychopathology did not appear to be related to levels of disgust sensitivity.

DISCUSSION

Although patients are not more sensitive than controls to the disgust-eliciting stimuli measured, disgust still has a positive relationship to eating disorder symptoms. Future studies will need to examine more precisely what this relationship might be.

Events and the onset of eating disorders: correcting an "age old" myth

BACKGROUND

Previous reports have suggested differences between early and late onset cases both in anorexia nervosa and bulimia nervosa. In anorexia nervosa, women with late onset are thought to be more likely to develop symptoms in response to life events than early onset cases. Women developing anorexia in response to life events are also considered to show different background features. In bulimia nervosa, it is not known whether rates of life events before onset differ in bulimics who develop their disorder at different stages in life.

METHOD

Previously published data were reanalyzed to examine differences between women whose eating disorder did or did not develop in response to stress. Seventy-two women with anorexia nervosa (48 with a severe provoking agent) and 29 women with bulimia nervosa (22 with a severe provoking agent) were administered semistructured interviews to assess life events and difficulties in the year before onset, childhood adversity, and other clinical features.

RESULTS

In both anorexia nervosa and bulimia nervosa, subjects with or without a severe provoking agent did not differ from each other on age or childhood variables. The only significant difference found was that, in women developing anorexia nervosa, those who later developed bulimic symptoms alongside their anorexia reported a lower rate of provoking agents (31%) than those anorexic women who either maintained their restriction (76%) or went on to develop normal weight bulimia nervosa (69%).

CONCLUSION

There are no differences between women whose eating disorder develops in response to stress and those who do not. Women presenting for treatment with anorexia nervosa of the binge-purge subtype are unlikely to have developed their illness in response to a severe provoking agent.

Stress, coping, and crisis support in eating disorders

OBJECTIVE

Recent research has supported the role of stress in the development and maintenance of eating disorders. However, coping and crisis support, important aspects of this stress process, have received little systematic attention. The cognitive-transactional approach to coping emphasizes the importance of the relationship between the individual and the particular problematic situation and yet most studies investigating coping in eating disorders have failed to measure situation-specific coping.

METHOD

The present study used semistructured interviews to measure coping and crisis support in response to severe events and/or marked difficulties in 12 women with anorexia nervosa (AN), 21 women with bulimia nervosa (BN), and 21 women without an eating disorder (non-ED).

RESULTS

Women with eating disorders were more likely to use cognitive avoidance or cognitive rumination and were less likely to downplay their problems. In addition, BN subjects were more likely to blame themselves and were less likely to receive crisis support from a core-tie. Overall, women with eating disorders were less likely to be masterful in response to crises than women without eating disorders.

DISCUSSION

It is concluded that women with eating disorders are less effective in their coping than women without eating disorders.

Psychosocial factors in the onset of eating disorders: responses to life-events and difficulties

A number of studies have proposed a role for stress in the onset of eating disorders. Initially the focus was on the life-events and difficulties themselves. However, the aim of this study was to examine the coping and support elicited in response to the crises that precipitate onset of eating disorders. Thirty-two women who had developed an eating disorder within four years of presentation were administered semi-structured interviews asking about events and difficulties in the year before onset, as well as how they coped and the support they received. Twenty women with no history of eating disorders acted as a comparison group. Results show that the onset of anorexic symptoms is associated with cognitive avoidance in response to a crisis while the onset of bulimic symptoms is associated with cognitive rumination. In addition, women who developed an eating disorder were more likely to be helpless in response to the crisis than women who did not develop an eating disorder. It is concluded that coping interventions may be useful in terms of primary and relapse prevention.

Setting the scene for eating disorders, II. Childhood helplessness and mastery

BACKGROUND

Previous studies have examined childhood factors that appear to increase the risk of developing an eating disorder (e.g. sexual abuse and parental care). Studies have not yet examined whether the way in which women cope with these adversities in childhood influences the risk.

METHOD

Using a semi-structured interview, childhood helplessness and mastery were measured (based on behavioural indices) in women with and without a history of eating disorders.

RESULTS

There was a higher rate of childhood helplessness and a lower rate of childhood mastery in women with eating disorders compared to those without. Furthermore, this difference did not appear to be a result of current psychiatric state.

CONCLUSION

It is concluded that it is not simply the presence of adversity in childhood which is of aetiological importance in the development of eating disorders but the way in which these are negotiated.

Predictors of outcome for two treatments for bulimia nervosa: short and long-term

OBJECTIVE

This study examined pretreatment variables to predict outcome in two treatments for bulimia nervosa.

METHOD

Patients were offered either 16 weeks of cognitive-behavioral therapy (CBT) or a self-treatment manual followed by up to 8 weeks of CBT (sequential group). Using complete data, stepwise regression analyses were performed.

RESULTS

It was found that a longer duration of illness and lower binge frequency predicted a better outcome both at the end of treatment (p < .001) and at 18 months of follow-up (p < .005). In the sequential group, lower pretreatment binge frequency predicted better outcome at the end of treatment (p < .05) and at 18 months of follow-up (p < .05). In the CBT group, longer duration of illness predicted better outcome at the end of treatment (p < .02).

DISCUSSION

It is concluded that (1) those with more frequent binging may require a more intense intervention and (2) those who have been ill longer may be more motivated to respond to treatment.

Feelings and fantasy in eating disorders: a factor analysis of the Toronto Alexithymia Scale

The aim of the study was to examine the factor structure of alexithymia in patients with eating disorders and to compare scores on these factors with a non-eating disordered population. The Toronto Alexithymia Scale (TAS) was given to patients with restricting anorexia nervosa (AN/R, n = 29), bulimia nervosa (BN, n = 83), anorexia with a subtype bulimia (AN/BN, n = 15), and 79 female students. Factors were extracted using a principal-components factor analysis. Four factors were found--inability to Identify Feelings, Paucity of Fantasy, Noncommunication of Feelings, and Concrete Thinking. All three eating disorder groups were less able to identify their feelings than the comparison group and AN/R patients had a more diminished fantasy life than BN patients and students. Groups did not differ significantly on concrete thinking but there was a trend towards significance on noncommunication of feelings, with patient groups expressing their feelings less than comparison subjects. Differences between patient groups on factors of the TAS suggest that scores are not simply a result of psychopathology in general. Approaches which promote the identification and expression of feelings may be particularly useful in the treatment of eating disorders.

[Help with self-care]

Instructions for self-treatment, whether printed, presented via computer or by audiovisual means, are effective in the management of phobias, panic disorder, other anxieties, depression, bulimia nervosa, obesity, alcohol problems, nicotine abuse, myocardial infarction, AIDS, compliance problems and the counseling of patients' relatives. A lasting improvement has been shown for up to 7 years. The mechanisms of effective self-change are discussed.

Ways of coping in women with eating disorders

The aim of this study was to investigate the strategies used for coping with stress in eating disorder patients. Twenty-four anorexia nervosa (AN) patients, 66 bulimia nervosa (BN) patients, and 30 female control subjects completed a revised Ways of Coping Checklist, indicating how they dealt with a self-nominated stressor. The AN and BN patients used proportionately more avoidance than control subjects. The BN patients used proportionately more wishful thinking and sought less social support than control subjects but patients with AN did not differ significantly from either BN or control groups. Patient groups did not differ significantly from control subjects on their use of problem-focused coping or self-blame, although the use of problem-focused coping was significantly lower, and self-blame significantly higher, with psychological problems than with relationship and general problems in all groups. Coping failed to predict severity of eating pathology but, in the patient groups, Beck Depression scores were related positively to avoidant coping (avoidance in BN patients and wishful thinking in AN patients) and inversely to problem-focused coping and seeking social support (although the latter just failed to reach significance in the AN group). It is concluded that a treatment approach that teaches coping strategies, as well as removing the obstacles (cognitive, emotional, or practical) that preclude the use of more effective coping, may be a useful component of treatment.