زمینه عزت نفس در سندرم خستگی مزمن

Objective: It has been suggested that people with chronic fatigue syndrome (CFS) have low self-esteem; however, this is not necessarily apparent when self-esteem is measured overtly. This study is the first to investigate underlying self-esteem using information-processing measures and overtly administered measures of self-esteem with this population. Methods: The study comprised 68 participants (24 CFS, 24 healthy volunteers, and 20 chronic illness volunteers). A Self-Statements Questionnaire (SSQ) and an Emotional Stroop Test (EST) using neutral, positive, and negative trait words were administered. Results: Participants with CFS reported lower self-esteem than the two comparison groups on overt measures. Overt responses, however, did not fully account for the full extent of the interference effect from the negative word Stroop compared to the positive word Stroop. Conclusion: In contrast to previous studies, participants with CFS reported lower levels of self-esteem on overt measures than two comparison groups. It is suggested, however, that the extent to which participants reported low self-esteem did not fully reflect their underlying low self-esteem and that this may result from the use of rigidly held defence mechanisms. Further use of information-processing measures, in contrast to relying only on self-report measures, is advocated for future research.

Chronic fatigue syndrome (CFS) is a disorder characterised by a principal complaint of fatigue accompanied by substantial functional impairment [1]. The cognitive behavioural model of CFS described by Sharpe [2], following Surawy et al. [3], suggests that core beliefs reflecting low self-esteem lead to the development of rigidly held beliefs that predispose a person to developing a chronic illness when confronted by a stressful trigger (including physical stressors). However, rather than suggesting that people with CFS hold core beliefs of inadequacy and low self-esteem, studies have tended to report unimpaired self-esteem in this group. Even when people with postinfectious syndromes and CFS fulfil diagnostic criteria for depression they have tended to report few feelings of guilt and preserved levels of self-esteem [4], [5] and [6]. In their description of common themes that arise when carrying out therapy with people with CFS, Surawy et al. [3] note a relative lack of expressed distress by participants. Similarly, a number of studies report participants' self-descriptions as ‘not the sort of person to become depressed’ [3], [6], [7] and [8]. Moss-Morris and Petrie [9] found reduced levels of self-esteem amongst people with CFS who were also depressed. They concluded, however, that on the whole people with CFS have good self-esteem unless they are depressed. Similarly, Johnson et al. [10] reported a nonsignificant tendency for depressed fibromyalgia participants to have lower basic self-esteem, but enhanced earning self-esteem (a sense of self-worth that is earned by competence). The evidence to date, therefore is inconsistent regarding levels of self-esteem amongst people with CFS who are also depressed. Moss-Morris [11] drew attention to the limitations of research to date, which has restricted itself to the investigation of conscious cognitive processes that are vulnerable to response biases. It might be predicted therefore that the development of rigid defence mechanisms might lead to a tendency for participants with CFS to report intact self-esteem despite having a vulnerable underlying self-concept. This study is the first to describe a comparison between overt and covert measures of self-esteem amongst people with CFS. Information-processing tasks have been used with other groups of participants to investigate discrepancies between self-report and underlying self-esteem (e.g., [13] and [14]). One such method is the Emotional Stroop Test (EST), developed from the Stroop Test [15], which requires participants to name the colour of ink that a word is written in. Attentional bias towards particular words or classes of words can be inferred from the degree of interference with the participant's performance, as measured by the speed of colour naming. Williams et al. [15] review the numerous applications of an emotional analogue of the Stroop test in relation to psychopathology and concluded that it can be regarded as a valid measure of individuals' covert concerns that is not dependent on conscious strategies. Furthermore, the test appears to detect differences in accessing constructs relating to the psychopathology rather than reflecting a consequence of the associated affect [16]. Kinderman [12] studied attention to positive and negative trait words using an EST with participants suffering from persecutory delusions and found that, despite them endorsing more positive than negative adjectives with reference to themselves, they showed a marked degree of interference when colour-naming negative words. This method was applied in the present study in order to investigate the hypothesis that people with CFS have underlying low self-esteem, which is not apparent when using overt measures. Patients with a number of illness features in common with patients with CFS are those who experience chronic pain. The use of modified versions of the Stroop Test have been somewhat inconsistent, however, and appear to suggest that when depression and anxiety are taken in to account, there are no specific effects of chronic pain on attentional bias in relation to affective stimuli [17]. The sensitivity of the methodology used by Pincus et al. [17] may have been reduced by the use of a manual rather than verbal response to the modified Stroop. Furthermore, this study used affective words relating to threat rather than low self-esteem. In order to evaluate whether any effect found is specific to CFS or a reflection of a response to chronic illness, a comparison group of people with a different chronic illness was included. Patients with insulin-treated diabetes mellitus formed the comparison group given that as well as representing a chronic illness group that affects people of all ages, there are a number of similarities between CFS and diabetes. In particular, diabetes is a ‘hidden’ condition that is not readily apparent to other people. Furthermore, diabetes is a demanding condition that requires significant regulation of behaviour and imposes restrictions on lifestyle. Although higher than in the general population, the lifetime prevalence of psychiatric disorders amongst people with diabetes is approximately equivalent to other chronic illness groups [18]. The hypotheses of the current study were as follows:

Table 1 represents the participants' age, gender, ethnicity, occupation, education level reached, and duration of illness. A significant difference was found between the groups with regard gender, reflecting a greater proportion of male participants within the diabetes group. No gender differences were found on the SSQ or EST, so gender was not included in the analyses described below. A significantly higher proportion of participants with CFS described their ethnic background as ‘White British’. There were no significant differences between the proportions of the groups who had participated in further education (beyond GCSE/O level). Although all participants were between the ages of 18 and 30 years, participants in the diabetes group tended to be older than the other groups. The diabetes group had also been ill for significantly longer than the CFS group. As would be expected, significantly smaller proportions of the CFS group were in employment or education than either of the other two groups. On the HADS, significant differences were found between the groups on anxiety and depression, with highest scores in the CFS group, followed by the diabetes group, then the healthy comparison group. Anxiety and depression were not significantly correlated with length of illness. Table 1. Demographic information CFS (n=24) Diabetes (n=20) Healthy (n=24) Statistic Gender, % female (n) 83.3 (20) 50 (10) 83.3 (20) χ2(2)=7.90 * Ethnicity, % White British (n) 95.8 (23) 75 (15) 62.5 (15) χ2(2)=7.90 * Occupation, % in employment/education (n) 62.5 (15) 100 (20) 95.8 (23) χ2(2)=15.51 *** Education, % further education (n) 79.2 (19) 90 (23) 95.8 (18) χ2(2)=3.30 Age, mean years (S.D., range) 24.73 (4.07, 18–30.67) 26.51 (2.89, 19.42–30.58) 23.08 (3.35, 18.75–30.58) F(2,65)=5.21 ** Illness duration, mean months (S.D., range) 43.42 (26.40, 5–108) 131.55 (82.57,22–271) N/A F(1,42)=24.45 *** HADS-D/21, mean (S.D., range) 6.29 (3.37, 1–13) 3.65 (3.18, 0–10) 1.96 (1.92, 0–8) F(2,65)=13.4 *** HADS-A/21, mean (S.D., range) 9.42 (3.94, 2–18) 6.35 (3.18, 0–11) 5.71 (2.49, 1–10) F(2,65)=9.08 *** * P<.05. *** P<.001. ** P<.01. Table options SSQ The distribution of responses to negative words on the SSQ differed significantly from the normal distribution (Kolmogorov–Smirnoff Z=3.07, P<.001). Log and square root transformations were attempted but the data still did not approach a normal distribution. For analyses involving this variable, nonparametric statistics were used. No significant differences were found in the number of positive words that were endorsed by the groups. The CFS participants, however, endorsed the most negative words, followed by the diabetes group (see Table 2). The group effect for the composite self-esteem score approached significance [Kruskal–Wallis χ2(2)=4.96, P=.08]. Tests of simple effects demonstrated that participants with CFS negative descriptor scores (and composite) were significantly different from both the healthy participants (Mann–Whitney U=147, P=.001) and the participants with diabetes (U=141.5, P=.01). The healthy and diabetes participants did not differ from each other (U=220, P=.50). Table 2. Number of positive and negative words endorsed on the SSQ Self-esteem score=positive word endorsed+negative words not endorsed (maximum=24). CFS (n=24) Diabetes (n=20) Healthy (n=24) Kruskal–Wallis χ2(2) Positive words endorsed, mean (S.D., range) 8.08 (2.99, 3–12) 9.5 (2.35, 4–12) 9.25 (2.38, 5–12) 2.92 Negative words not endorsed, mean (S.D., range) 11.0 (0.93, 9–12) 11.55 (1.15, 7–12) 11.79 (0.51, 10–12) 13.44*** Self-esteem score, mean (S.D., range) 19.08 (3.49, 13–24) 20.85 (3.12, 11–24) 21.04 (2.56, 16–24) 4.96 *** P<.001. Table options EST Repeated measures analysis of variance were carried out to compare the time each group took to colour-name in each condition. A significant group effect was found using a Group×Word type (neutral, negative, positive) design (F(2,63)=18.30, P<.001). Post hoc tests controlling for Type 1 error indicated that the CFS participants were significantly slower to colour-name all three word types than the two comparison groups (Tukey HSD, P<.001) (see Table 3). Table 3. Time to colour-name (seconds) on the EST CFS (n=24) Diabetes (n=20) Healthy (n=24) O's time (s), mean (S.D., range) 30.91 (5.12, 22.22–43.47) 25.40 (4.16, 18.94–35.75) 25.35 (4.08. 19.91–34.16) Neutral words time (s), mean (S.D., range) 37.71 (7.89, 24.81–56.21) 29.97 (4.36, 22.88–39.97) 28.39 (5.04, 20.0–39.06) Positive words time (s), mean (S.D., range) 36.29 (6.78, 25.25–55.6) 29.46 (3.81, 23.53–36.44) 28.80 (5.06, 19.56–39.37) Negative words time (s), mean (S.D., range) 38.42 (7.89, 24.81–56.21) 29.97 (4.36, 22.88–39.97) 28.23 (4.19, 20.72–35.75) Table options In order to investigate group differences further, an analysis of covariance was conducted using a Group×Word Type (positive or negative), covarying with the composite self-esteem score, HADS-A, and HADS-D total scores. A significant interaction was found between word type and group [F(2,65)=5.937, P=.004]. There was a greater increase in the time to name negative words compared to positive words for the CFS group and a slight reduction in the time to name negative words compared to positive words for the healthy comparison group. Post hoc tests using Tukey's HSD found a significant difference between the CFS and healthy groups but not between the CFS and diabetes groups or diabetes and healthy groups. The effect of self-esteem, HADS-D, and HADS-A did not have significant independent effects, although the effect of depression approached significance [F(9)=3.051, P=.086]. After covarying for these factors, the Group×Word Type interaction effect was reduced but continued to approach significance [F(2.62)=2.97, P=.059].