Alexithymia in patients with psoriasis

Alexithymia in patients with psoriasis

Journal of Psychosomatic Research 58 (2005) 89 – 96 Alexithymia in patients with psoriasis Clinical correlates and psychometric properties of the Tor...

148KB Sizes 3 Downloads 155 Views

Journal of Psychosomatic Research 58 (2005) 89 – 96

Alexithymia in patients with psoriasis Clinical correlates and psychometric properties of the Toronto Alexithymia Scale-20 Helen L. Richardsa,b,*, Do´nal G. Fortunea, Christopher E.M. Griffithsb, Chris J. Maina a Department of Behavioural Medicine, Hope Hospital, Salford Royal Hospitals NHS Trust, Manchester, UK Dermatology Centre, University of Manchester School of Medicine, Hope Hospital, Salford Royal Hospitals NHS Trust, Manchester, UK

b

Received 29 October 2003; accepted 16 March 2004

Abstract Objective: The objective of this study was to examine the psychometric properties and clinical correlates of the 20-item Toronto Alexithymia Scale (TAS-20) in patients with psoriasis. Methods: Three hundred patients with psoriasis completed the TAS-20 and had the severity of their psoriasis clinically assessed. Test–retest reliability was assessed over 10 weeks in a subset of 71 patients receiving standard medical treatment for their psoriasis. Examination of the factor structure employed both exploratory (EFA) and confirmatory (CFA) factor analysis techniques. Results: The overall prevalence of alexithymia in this population of psoriasis patients, as indicated by the TAS-20, was 33%. The TAS-20 total score was not related to clinical severity, age, age at

onset or duration of psoriasis. The 10-week test–retest reliability on a sample of 71 psoriasis patients, pre- and posttreatment with photochemotherapy, was found to be acceptable (r = .69; P b.001). EFA produced no stable solutions. The three-factor structure of the TAS-20 using CFA was also not replicable in this sample. Conclusions: The findings of this study support the contention that alexithymia is not significantly influenced by either disease severity or chronicity in patients with psoriasis. It is recommended that when employing the TAS-20 in patients with psoriasis, caution is exercised in the interpretation of the scale scores and that further psychometric evaluation of the scale is undertaken as appropriate. D 2005 Elsevier Inc. All rights reserved.

Keywords: Alexithymia; CFA; EFA; Psoriasis

Introduction One principle of psychosomatic medicine is that deficits in the experience and expression of emotions can have a harmful effect on health. In 1973, the label alexithymia was given by Sifneos [1] to assign a group of cognitive and affective characteristics typical of many patients with psychosomatic disorders. The definitions of alexithymia have been refined, and it is viewed currently as a deficit in the self-regulation of affect [2]. It is suggested that individuals with alexithymia have an impaired ability to

* Corresponding author. Department of Behavioural Medicine, University of Manchester, Clinical Sciences Building, Hope Hospital, Salford M6 8HD, UK. Tel.: +44 0161 206 5588; fax: +44 0161 206 5589. E-mail address: [email protected] (H.L. Richards). 0022-3999/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jpsychores.2004.03.009

build mental representations of emotions, that ordinary somatosensory information is magnified and physical symptoms of emotional arousal are misinterpreted as indicators of physical illness [2]. Subsequently, the concept of alexithymia has taken on a role as a key player within psychosomatic medicine. High prevalence rates of alexithymia have been identified in patients with a variety of health problems. Alexithymia has been associated with poorer metabolic control in patients with diabetes [3], reductions in immunocompetence in women with cervical neoplasia [4] and blood pressure elevations in patients with hypertension [5]. However, other studies have reported no such associations between alexithymia and conditions such as ischaemic heart disease [6], insulin-dependent diabetes mellitus [7] and breast cancer patients [8]. Thus, findings tend to be equivocal.

90

H.L. Richards et al. / Journal of Psychosomatic Research 58 (2005) 89 – 96

Psoriasis is a condition that has often been construed as being the result of personality factors and is associated with an array of physiological [9], behavioural [10], cognitive– affective [11] and social [12] difficulties. As such, psoriasis would appear to be a prime candidate for the investigation of alexithymia. Psoriasis is a chronic, inflammatory, currently incurable dermatological condition. It affects 2–3% of the population, has an equal distribution between the sexes, and its onset may occur at any age. Studies have indicated that genetic factors have an important role in the onset of psoriasis [13]. Psychological distress and stressful life events have been shown to be important in the onset, exacerbation and maintenance of the condition [14–16]. Ever-increasing evidence links chronic low-grade stress to maintenance of the condition [17]. Increasingly psoriasis is recognised to be associated with wide-ranging psychological distress, including anxiety [18], depression [19], disability [20] and pathological worrying [21]. For many years, psoriasis was considered to be a psychosomatic condition, predisposed by certain personality types. However, research to date has not confirmed higher prevalence of alexithymia for patients with psoriasis. Fava et al. [22] compared the alexithymic scores of 20 patients with psoriasis, chronic urticaria and cutaneous infections but did not find significantly higher alexithymic scores in psoriasis patients compared with the two other conditions. Rubino et al. [23], using the Toronto Alexithymia Scale (TAS), found that while patients with psoriasis did score higher on this measure relative to controls, the difference did not reach statistical significance. One recent study [24] found a trend for psoriasis patients with recent exacerbations to have higher scores on the 20-item TAS (TAS-20) than those with other skin conditions. Using the total scale score of the TAS, Allegranti et al. [25] found a significant difference between the patients with psoriasis and the controls. Patients with psoriasis were also more likely to score above the cutoff for borderline alexithymia than the controls. Other studies [26] have reported evidence that alexithymia may also be closely tied up with the construct of anxiety sensitivity, and, as such, it has been suggested that it may be important in accounting for poor psychological adjustment in patients with psoriasis [27]. Unfortunately, previous research investigating alexithymia in patients with psoriasis has been rendered unsatisfactory by the use of inadequate assessment instruments and different methods of investigating the construct. The TAS20 [28,29], which was constructed with regard to specific theoretical concerns, has been considered the best and most frequently used method in the measurement of alexithymia [30]. The TAS-20 is said to have a three-factor structure based on theoretical constructs comprising of (i) difficulty identifying feelings (DIF); (ii) difficulty describing feelings (DDF); and (iii) externally orientated thinking (EOT). A number of studies have replicated this factor structure across both clinical and nonclinical populations [31,32]. However, some studies have failed to confirm the factor structure of

the TAS-20 and have suggested exercising caution in its use [33,34]. It has been suggested that some of these confounding results have been due to variations in the factor analytic techniques employed, in that some studies have used exploratory models whereas others have preferred confirmatory models as a means of testing specific hypotheses [31]. A recent study [35] employed the TAS-20 in a community population in Canada and reported results of the confirmatory factor analysis (CFA), which supported the original three-factor structure. They propose that the exploratory factor analysis (EFA) approach is better suited to studies where little is known about the concept being developed, and as such, within the field of alexithymia, CFA is more appropriate. However, there still remains some criticism of the three-factor fixed model identified in CFA studies [33,34]. Haviland and Reise [34], using CFA, reported that none of the various factor solutions produced a simple structure in either a medical student or psychoactive-substance-dependent inpatient sample, although in this study, a number of the goodness-of-fit indices were at or just short of the criterion level. Recently, Mqller et al. [36] tested a range of factor structures of the TAS-20 in different populations of subjects and concluded that the factor structure of the instrument may vary across clinical and normal adult samples, going some way towards explaining the divergence in results across studies. Given the presence of alexithymia as an essential element in the genesis and maintenance of various psychosomatic pathologies, it was considered important to examine the prevalence of alexithymia in patients with psoriasis and investigate the degree to which it is associated with clinical aspects of the condition. For both clinical and research purposes, it was also considered pertinent to examine the psychometric properties of the TAS-20.

Method Participants Three hundred consecutive patients with chronic plaque psoriasis attending a hospital-based psoriasis specialty clinic had their psoriasis assessed by a dermatologist and completed the TAS (TAS-20). Participants were excluded if they had any other general medical condition, with the exception of psoriatic arthritis. Test–retest reliability over 10 weeks was assessed in a subsample of these patients returning to the clinic for standard pharmacological treatment of their condition. The study protocol was approved by Salford and Trafford Local Research Ethics Committee. Clinical assessment Disease severity was assessed by the Psoriasis Area and Severity Index (PASI) [37]. The PASI incorporates the clinical extent of psoriasis (surface area of skin

H.L. Richards et al. / Journal of Psychosomatic Research 58 (2005) 89 – 96

affected) and clinical severity of its manifestations (erythema, desquamation and induration) using a formula that yields a value of between 0 (no psoriasis) and 72 (very severe psoriasis). Psychological assessment The TAS-20 is a 20-item self-report measure developed in an attempt to improve the previous 26-item version of the scale [28,29]. It requires respondents to indicate to what extent they agree with each item. Scores range from 1, indicating that the respondent strongly disagrees with the item, to 5, indicating strong agreement with the item. Items 4, 5, 10, 18 and 19 in the TAS-20 are reverse scored. Statistical analysis Sample characteristics were examined using descriptive statistics. With the exception of disease severity, as measured by PASI, the characteristics of the data permitted the use of parametric tests. PASI scores were transformed using log transformation (base 10). The construct validity of the questionnaire was firstly examined using a series of principal component EFA. Kaiser–Meyer–Olkins measure of sampling adequacy (MSA) and Bartlett’s test of sphericity were calculated to verify the appropriateness of using factor models. Factors were extracted using the eigenvalue-one procedure and Cattell’s scree test [38]. Varimax rotation was employed to rotate the factors to a simple structure. Re-factor analysis of two random split halves of the sample was also undertaken to ensure a more rigorous means of deciding on the number of factors to retain and to examine the stability of the factor solution. Items were assigned to a factor if the square of the loading for a factor was N50% that of its loading on any other factor [39]. A ratio of 10 participants to each item was deemed the least amount to permit a random split, with a minimum of five cases to one variable ratio. Initially, CFA was undertaken on all 300 data sets using the three-factor criteria (DIF, DDF, and EOT) suggested by Bagby et al. [28]. The Analysis of Moment Structures (AMOS) version 4 for SPSS package [40] was employed. A one-factor model was also investigated by allocating all items of the TAS-20 to one variable. In light of the results from previous studies of the TAS-20 factor structure [34], a two-factor model was constructed by assigning the items DIF and DDF to one variable and EOT items to another. Oblique models were tested to correspond with original factor associations. Goodness of fit was evaluated with the following statistics (acceptable criteria level): root-meansquare error of approximation (RMSEA) and its 90% confidence interval (b0.08), standardized root-mean-square residual (SRMR; b0.05); comparative fit index (CFI; N0.90) and Tucker Lewis Index (TLI; N0.90). The criterion values used were in line with those proposed by previous studies as acceptable [35,41]. Comparative fit using chi square (non-

91

significant) was also employed, despite recent criticisms [42] regarding its use in small sample sizes, to allow for comparisons with previous investigations. In comparing one- and two-factor models with the hypothesized threefactor, model chi-square difference was calculated between the models. Internal reliability and consistency of the scale was computed using Cronbach’s alpha (a) coefficient. Differences between groups were assessed by t test and analysis of variance, as appropriate. Pearson’s correlation coefficient was used to process correlations. The test–retest reliability of the TAS-20 was assessed by calculating Pearson’s Correlation Coefficient at two time points over 10 weeks in a sample of 71 patients receiving a standardized treatment protocol (Psoralen and UltraViolet A Irradiation treatment) at the beginning and end of this treatment. As PUVA treatment demands attendance at the centre for treatment [43], PUVA was chosen to control for the potential for interaction between alexithymia and adherence with treatment, given that up to 40% of patients with psoriasis do not comply with treatment [10]. To investigate the potential movements from alexithymic to nonalexithymic categories, the McNemar test was employed. To assess the prevalence of alexithymia, the TAS-20 scores were categorised according to suggested cutoffs: total score z61, indicating alexithymia; 52 to 60, intermediate; and V51, indicating no alexithymia [2]. For the purposes of statistical comparisons, two extreme groups were created, above and below the cutoff of 61 [2].

Results Study I Description of study population Of the 300 patients participating, 46% were female and 54% male. They were aged between 18 and 80 years (mean 43.4, S.D. 12.1). The mean severity of their psoriasis, as assessed by the PASI, was 8.7 (S.D. 6.21) and ranged between 0 and 51.8. The age at disease onset ranged between 2 and 72 years, with a mean of 23.5 (S.D. 14.3). The mean duration of psoriasis was 19.9 years (S.D. 11.8) and ranged between 6 months and 60 years. Twelve percent of the patients had psoriasis for less than 5 years, 11% for 6–10 years, 33% for 10–20 years and 44% for over 20 years. Characteristics associated with alexithymia The mean score on the TAS-20 was 55.62 (S.D. = 11.06). Scores ranged from 25 to 80. The overall presence of alexithymia, according to the TAS-20 cut-off ( z 61), within this population was 35% (n = 104). Males and females showed no significant differences in their total scores on the TAS-20 (t = 0.78, P = ns). Pearson’s correlations did not identify any relationship between TAS20 total scores and age (r = .05, P = ns), age at onset of

92

H.L. Richards et al. / Journal of Psychosomatic Research 58 (2005) 89 – 96

psoriasis (r = .07, P = ns), duration of psoriasis (r = .01, P = ns) and severity (total PASI; r = .04, P = ns). Structure of the measure The internal reliability of the 20-item scale using Cronbach’s alpha (a) was .82 and the mean inter-item correlation coefficient, .18. Coefficient alphas for the three subscales identified by Bagby et al. [28] were .72 (DDF), .84 (DIF) and .44 (EOT). Mean inter-item correlations were .34 (DDF), .44 (DIF) and .09 (EOT). Moderate–strong correlations were observed between the TAS-20 total score and each of its subscales: DIF, .82; DDF, .85; and EOT, .59. The inter-item correlations for the full scale ranged from .003 to .66. Correlations between the individual items and the total scale score ranged from .02 (Item 10) to .65 (Item 2). Two items correlated less than .1 with the total TAS-20. These were Items 5 (bI prefer to analyse problems rather than just describe themQ) and 10 (bBeing in touch with emotions is essentialQ). These two items were contained within the EOT subscale and were negatively keyed; that is, agreement with the items indicated nonalexithymic responses. Previous studies [44] have suggested that negatively keyed items may have an important impact on the homogeneity of the TAS-20 in that they do not differentiate between alexithymics and nonalexithymics due to suggested inflexibility of response style on behalf of the alexithymic group. Due to possible response style inflexibility, the item–total correlations for the other negatively keyed items were examined more closely. Items 18 and 4 had acceptable item–total correlations of .19 and .32 respectively, whereas Item 19 showed an interitem correlation of .11. The negatively keyed item–total correlations were low in comparison to the positively keyed responses, which ranged from .29 to .65, with the exception of Item 16 (.1). We further explored the response patterns of the alexithymic (TAS-20 z61) and nonalexithymic groups for the positively and negatively keyed items by examining the responses of the two groups to the TAS-20 items. The percentage of patients in each group (alexithymic/nonalexithymic) agreeing or agreeing strongly with each of the items was examined. A vast majority of the alexithymic patients agreed with the positively keyed items. For example, 76% of alexithymic patients agreed with Item 6, bWhen I am upset, I don’t know if I am sad, frightened or angryQ, whereas 26% of the nonalexithymics disagreed. In contrast, for the negatively keyed items, the difference in responses was far less marked. For example, with Item 5, bI prefer to analyse problems rather than just describe themQ, 52% and 61% of the alexithymic and nonalexithymic group, respectively, agreed with this item. Similarly, Item 10, bBeing in touch with emotions is essentialQ was agreed with by 68% of the nonalexithymic and 64% of the alexithymic group. This is in line with previous findings [44] and suggests a less adaptable response pattern of the alexithymic group to the positive and negative-keyed items contained within the TAS-20.

Exploratory factor analysis (EFA) Given the controversy in methodologies previously employed, we first undertook an EFA. To determine the factor structure of the TAS-20, principal components factor analysis using eigenvalue-one procedure with varimax rotation was employed to rotate the factors to a simple structure. All original responses to the 20 items from the 300 questionnaires were utilized. The data met Kaiser–Meyer– Olkins criteria for sampling adequacy (MSA=0.86) and Bartlett’s test of sphericity ( Pb.001). EFA extracted four factors accounting for 51% of the variance (Table 1). Factor 1 had an eigenvalue of 5.22 and accounted for 26.1% of the variance, Factor 2 had an eigenvalue of 1.94 (9.7% of the variance), Factor 3 had an eigenvalue of 1.73 (8.7% of the variance) and Factor 4 had an eigenvalue of 1.21 (6.1% of the variance). Item 17 (bIt is difficult for me to reveal my innermost feelings even to close friendsQ) was split across factors. The splitting of this item across factors is commensurate with results reported in a myocardial infarction population [44]. Item factor loadings are illustrated in Table 1. The first component consisted of 10 items; all 7 items originally reported by Bagby et al. [28] from the DIF subscale of the TAS-20. In addition, Items 2, 11 and 12 originally in the DDF subscale were also contained within this factor. Factor 2 contained Items 10, 18 and 19 originally within EOT, along with Item 4 from DDF, all of which are negatively keyed items. The remaining items from the original EOT scale were contained within Factors 3 (Items 15, 16, 20) and 4 (Items 8 and 5). Cronbach’s alphas for each of the four extracted components were .83, .52, .47 and .21, respectively. With the exception of the first component, these were considered low values in terms of internal consistency. Kaiser’s criterion (the eigenvalue-one rule) often overestimates the number of components because of sampling effects [45,46]. To compensate for the possibility of sampling error and because of the relatively low amounts of variance accounted for by Factors 2, 3 and 4 and of the splitting of Item 17 across factors, a scree plot was examined, and both two- and threefactor solutions were investigated for the measure. Examination of both the two- and three-factor solutions identified a number of items that were split between factors. Items that were split between factors were examined and added to factors on the basis of the size of the loadings across the split factors [39]. Where the items could not be assigned to a factor on the basis of size, individually and in combination, the split items were added to each of the individual scales to examine whether they contributed any additional reliability to the scale (Cronbach’s alpha). In the three-factor solution, Factor 1 had an eigenvalue of 5.5 and accounted for 27.6% of the variance. Factor 2 had an eigenvalue of 1.94 and accounted for 9.7% of the variance and Factor 3 had an eigenvalue of 1.5 and accounted for 7% of the variance. Alphas for these factors were .97, .52 and .46 respectively. Factor 1 was relatively in keeping with the DIF subscale identified by Bagby et al.

H.L. Richards et al. / Journal of Psychosomatic Research 58 (2005) 89 – 96

93

Table 1 Factor loadings component matrix Factors Subscale DIF DIF DDF DIF DIF DIF DIF DDF DDF DIF EOT EOT EOT DDF EOT EOT EOT EOT EOT DDF

1 1. I am often confused about what emotion I am feeling 6. When I am upset, I do not know if I am sad, frightened or angry 2. It is difficult for me to find the right words for my feelings 13. I do not know what is going on inside me 7. I am often puzzled by sensations in my body 14. I often do not know why I am angry 3. I have physical sensations that even doctors do not understand 11. I find it hard to describe how I feel about people 12. People tell me to describe my feelings more 9. I have feelings that I cannot quite identify 19. I find examination of my feelings useful in solving personal problems. 10. Being in touch with emotions is essential 18. I can feel close to someone even in moments of silence 4. I am able to describe my feelings easily 16. I prefer to watch light entertainment shows rather than psychological dramas 15. I prefer talking to people about their daily activities rather than their feelings 20. Looking for hidden meanings in movies or plays distracts from their enjoyment 8. I prefer to just let things happen rather than to understand why they turned out that way 5. I prefer to analyse problems rather than just describe them 17. It is difficult for me to reveal my innermost feelings even to close friends

2

3 .06 .02 .14 .04 .18 .10 .09 .09 .08 .07 .69 .68 .58 .49 .09 .20 .22 .03 .33 .36

.78 .75 .74 .73 .72 .70 .69 .68 .47 .44 .08 .16 .14 .39 .05 .19 .28 .26 .09 .50

4 .08 .08 .18 .23 .07 .17 .14 .08 .28 .27 .12 .01 .09 .08 .72 .62 .52 .14 .08 .45

.05 .06 .05 .02 .20 .06 .31 .06 .004 .33 .25 .04 .12 .26 .11 .02 .10 .64 .58 .18

Sub-scales DIF: difficulty in identifying feelings; DDF: difficulty in describing feelings; EOT: externally oriented thinking. Factor loadings above .40 are shown in bold type.

[28] but also contained items from the EOT and DDT scales. Factors 2 and 3 contained a majority of the items identified in the subscale originally labeled bemotionally oriented thinkingQ. Furthermore, Factor 2 contained all the reversescored items. Items labeled as DDF were contained in Factors 1 and 2 in the current analysis. The stability of this three-factor solution was checked by the split-half method (Table 2). An examination of the rotated factor loadings suggested that the solution was unstable. A number of the items (4, 8 and 17) differed in the factors they loaded on in each separate analysis and a number of the items were consistently split between factors. Given these inconsistencies, the fact that the alphas for Factors 2 and 3 were not sufficiently high (b.52), whereas the total scale alpha was (.82), and the strength of the first factor in comparison to the others, it was considered that the three-factor solution was not the most appropriate use of this scale in this sample of patients with psoriasis. In examining the two-factor solution, Factor 1 had an eigenvalue of 5.53 and accounted for 27.6% of the variance, and Factor 2 an eigenvalue of 1.94, accounting for 9.71% of the variance. Again, a number of items were split between the two factors. Having assigned items to factors on the basis outlined previously, the stability of this solution was explored via a random split. This indicated that a number of items inconsistently loaded on each of the factors (4, 15, 16, 17 and 20). Given these findings, the two-factor solution was considered nonviable. Confirmatory factor analysis (CFA) CFA was then performed, building on issues highlighted in previous analyses. The difference between the predicted

three-factor and observed solution was significant (v 2 = 471.4, df = 167, P b.001), indicating that this solution was not a good comparative fit with the three-factor model originally suggested [28,29]. In line with previous studies, one- and two-factor solutions (Factor 1 DIF and DDF; Factor 2 EOT) were also investigated. The difference in the goodness-of-fit chi squares indicated that of the three models, the three-factor solution offered a better fit than did the one-factor model (v 2 difference = 39.71, df = 3, P b.001) and the two-factor solution (v 2 difference = 35.27, df = 2, P b.001). Table 3 illustrates the goodness-of-fit indices for all three-factor solutions. The only criteria index approaching acceptance was the RMSEA. The only significant parameter estimates was between DIF and DDF (0.70, P b.01; DDF and EOT, 0.10, P = .07; DIF and EOT, 0.01, P = .89). Study II: Test–retest reliability Study population A subset of 71 participants was used in the investigation of test–retest reliability. Thirty-eight percent of this sample was female, with a mean age of 42.7 (S.D.

Table 2 Random split of TAS-20 three-factor solution Random split 1

Random split 2

Factor

1

2

3

1

2

3

Eigenvalue Percent of variance

5.76 28.78

1.87 9.36

1.57 7.83

5.69 28.47

2.17 10.86

1.50 7.56

94

H.L. Richards et al. / Journal of Psychosomatic Research 58 (2005) 89 – 96

Table 3 Goodness-of-fit indices for the confirmatory factor analysis of the TAS-20 Psoriasis sample (n=300) Index

Three-factor

Two-factor

One-factor

GFI AGFI CFI RMSR RMSEA

0.86 0.82 0.81 0.14 0.08

0.85 0.81 0.79 0.14 0.08

0.84 0.80 0.79 0.10 0.08

CFI = comparative fit index; RSMR = root-mean-square residual; RMSEA= root-mean-square error of approximation; GFI = goodness-of-fit index; AGFI = adjusted goodness-of-fit index.

13.4) years. The mean duration of psoriasis was 19.6 (S.D. 10.6) years, and at time one (T1), they had a mean PASI score of 12.5 (6.1). There were no significant differences between this group and the overall population in terms of gender (v 2 = 2.57, P = .11), age, duration of psoriasis and age of onset (ts b 0.59, Ps N .56). Patients in the test–retest population did have significantly more severe psoriasis (t = 5.52, P b.001), reflective of their PUVA treatment status. Participants completed the TAS20 10 weeks following first completion (Time 2). At time two (T2), 61 of these patients had achieved clearance of their psoriasis (less than 1% total body surface area affected by psoriasis). Reliability of measure The TAS-20 was found to be stable over this period, with a test–retest correlation of r = .69 ( P b.001) between administration times 1 and 2. The mean TAS-20 score at T1 was 53.8 (10.4) and at T2, 52.2 (10.6). There was not a statistically significant reduction in mean TAS-20 scores over time (t = 1.66, P = .10). The TAS-20 demonstrated strong internal consistency, with an alpha of .71 at Time 1 and .86 at Time 2 (n = 71). Test–retest values for individual items ranged from .04 (Item 9) to .62 (Item 1). Twenty-four percent of this group were alexithymic at Time 1 and 18% at Time 2, which was statistically insignificant ( P = .42). Additionally, this was not related to clearance of psoriasis (m2 = 0.89, P = .35). Factors associated with alexithymia Given the decrease in alexithymic scores at Time 2, TAS20 residual change scores were calculated for the patients, and these were investigated in association with demographic and clinical variables measured in this study. Only an inverse relationship with age ( P b.01) was identified.

Discussion The examination of alexithymia has been increasingly employed within investigations of patients with psoriasis. The findings of the current study have provided important information in relation to the psychometric properties of

the TAS-20 and its relationship with clinical correlates of psoriasis. Thirty-three percent of the current sample of psoriasis patients scored above the cutoff for alexithymia. This is comparable with that reported in inflammatory bowel disease (35%) [47] and myocardial infarction (30%) [44] and higher than figures reported in normal controls (13% [30], 16% [5]). There were no gender differences identified in the prevalence of alexithymia. A slightly higher percentage of men scored above the cutoff for alexithymia, but this did not reach statistical significance. This is at odds with previous research [30], which has suggested that alexithymia is a personality trait perhaps more typical of men that women. No relationship was identified between alexithymia and the chronicity or severity of psoriasis, which has confirmed findings from other medical conditions [47,48]. This finding lends itself to a number of considerations and suggests that in this group of psoriasis patients, alexithymia is not a reaction to chronic illness influenced by duration or severity. This is in keeping with previous studies that suggest alexithymia to be a fairly stable trait [29]. However, others [49] have suggested that acute or chronic secondary alexithymia may be reactions to stressful situations, such as a disease episode. The low numbers of patients presenting at first onset of psoriasis in the current study do not permit comment on the concept of acute secondary alexithymia as a response to an acute disease episode, and this warrants further investigation. The preliminary results from the prospective treatment study allow us to comment on the issue of whether alexithymic scores reflect a personality trait or a coping response to a chronic disease state. Our patient group on whom the test–retest reliability was undertaken had just completed a period of PUVA treatment in which the majority of patients cleared their psoriasis. In terms of test–retest reliability in this subsample of psoriasis patients, high stability was demonstrated over 10 weeks (r = .69, P b.001), and decrements in the assessment of alexithymia in this subgroup were not found. This suggests that alexithymia in patients with psoriasis is perhaps not a coping strategy in response to an active illness episode and rather a more stable characteristic. It is of note however, that we have recently observed [50] that although factors specific to psoriasis (disability and stress) significantly improve following disease clearance, by comparison, there are no significant improvements in anxiety, depression or worrying. We believe that the inflexibility that is associated with psychological distress does not permit such objective changes to a patient’s life to become immediately apparent. Thus as Kojima et al. [44] note from their observations of myocardial infarction patients, it may be that a period of time is required to allow the individual to adjust to the changes imposed by the improvement of their condition. Clearly, further longitudinal studies are required to substantiate this hypothesis.

H.L. Richards et al. / Journal of Psychosomatic Research 58 (2005) 89 – 96

Good internal consistency and homogeneity were confirmed for the full-scale TAS-20 in our sample of 300 psoriasis patients. However, this was not the case when the three subscales were examined separately. While the DIF and DDF scales were considered acceptable, the EOT subscale had a very low mean inter-item correlation coefficient. This is in keeping with results from a number of other studies suggesting that the EOT dimension is unreliable [32,36,44]. The results of the CFA and EFA factor analysis investigation did not support findings from previous research [35]. The EFA resulted in no clean, stable structures for this sample of patients, and from the CFA, the hypothesized three-factor model provided a poor fit to the data, although was superior to both the one- and two-factor solutions, with the RMSEA approaching acceptance. Thus, despite the findings from the recent large community study supporting the factorial validity of the TAS-20 [35], the results of this study suggest that caution should be taken when employing the TAS-20 in psoriasis. Indeed, it is recommended, in line with other studies [33,34] of the replicability of the TAS-20 factor structure, that further investigations are undertaken of its psychometric properties when used in new clinical populations. Commensurate with previous studies, we found that there were fewer differences between alexithymics and nonalexithymics in relation to the negatively keyed items [44]. Although the use of positively and negatively worded items is often encouraged in the design of psychometric scales, it is increasingly recognised that the inclusion of such items may contribute to significant difficulties [51] and it is recommended that when negatively worded items are used in conjunction with positively worded items, their proportions should be equal to enable a distinction between method effects and underlying constructs. Indeed, this is an interesting finding and suggests the need to consider whether such response patterns are an artifact of response style associated with the wording of the items or represent a domain in its own right, with substantive implications for the measurement of alexithymia more generally. It is suggested that further research is required to ascertain whether the five reverse-worded items make a meaningful contribution to the psychometric measurement properties of the TAS-20. In summary, this study has illustrated levels of alexithymia in patients with psoriasis comparable with that seen in other chronic medical conditions. This was unrelated to the clinical severity of their condition, and results of the test– retest investigation are suggestive that alexithymia is a trait rather than a state construct in this group. Psychometric evaluation of the TAS-20 suggests careful use of the instrument for research purposes in patients with psoriasis. Clearly, the evaluation of the relationship of alexithymia to clinical and psychological variables in patients with psoriasis warrants more a detailed examination, and further research is required to ascertain whether alexithymia is an antecedent or a consequence of the development of psoriasis.

95

Acknowledgments This study was supported by the NHS Executive Research and Development Programme for Physical and Complex Disabilities (Grant No: PCD2/A1/284).

References [1] Sifneos PE. The prevalence of balexithymicQ characteristics in psychosomatic patients. Psychother Psychosom 1973;22:255 – 62. [2] Taylor GJ, Bagby RM, Parker JDA. Disorders of affect regulation. Cambridge7 Cambridge Univ. Press, 1997. [3] Abramson L, McClelland DC, Brown D, Kelner S. Alexithymic characteristics and metabolic control in diabetic and healthy adults. J Nerv Ment Dis 1991;179:490 – 4. [4] Todarello O, Casamassima A, Marinaccio M, La Pesa MW, Caradonna L, Valentino L, Marinaccio L. Alexithymia, immunity and cervical intraepithelial neoplasia: a pilot study. Psychother Psychosom 1994; 61:119 – 204. [5] Todarello O, Taylor GJ, Parker JDA, Fanelli M. Alexithymia in essential hypertensive and psychiatric outpatients: a comparative study. J Psychosom Res 1995;39:987 – 94. [6] Kauhanen J, Kaplan GA, Cohen RD, Salonen R, Salonen JT. Alexithymia may influence the diagnosis of coronary heart disease. Psychosom Med 1994;56:237 – 44. [7] Friedman S, Vila G, Even C, Timsit J, Boitard C, Dardennes R, Guelfi JD, Mouren-Simconi MC. Alexithymia in insulin dependent diabetes mellitus is related to depression and not to somatic variables or compliance. J Psychosom Res 2003;55:283 – 7. [8] Anagnostopoulos F, Vaslamatizis G, Markidid M, Katsouyanni K, Vassilaros S, Stefanis C. An investigation of hostile and alexitymic characteristics in breast cancer patients. Psychother Psychosom 1993;59:179 – 89. [9] Kirby B, Richards HL, Woo P, Hindle E, Main CJ, Griffiths CEM. Physical and psychological measures are necessary to assess overall psoriasis severity. J Am Acad Dermatol 2001;45:72 – 6. [10] Richards HL, Fortune DG, O’Sullivan TM, Main CJ, Griffiths CEM. Psoriasis patients’ compliance with medication. J Am Acad Dermatol 1999;41:581 – 3. [11] Fortune DG, Richards HL, Main CJ, Griffiths CEM. Targeting cognitive–behavioural therapy to patients’ implicit models of psoriasis: results from a patient preference controlled trial. Br J Clin Psychol 2004;43:65 – 82. [12] Fortune DG, Main CJ, O’Sullivan TM, Griffiths CEM. Quality of life in patients with psoriasis: the contribution of clinical variables and psoriasis specific stress. Br J Dermatol 1997;137:755 – 60. [13] Duffy DL, Spelman LS, Martin NG. Psoriasis in Australian twins. J Am Acad Dermatol 1993;29:428 – 34. [14] Ginsburg IM. Psychological and psychophysiological aspects of psoriasis. Dermatol Clin 1995;13:793 – 804. [15] Gaston L, Lassonde M, Bernier-Buzzanga J, Hodgins S, Crombez J-C. Psoriasis and stress: a prospective study. J Am Acad Dermatol 1989;17:82 – 6. [16] Harvima RJ, Viinamaki H, Harvima IT, Naukkarinen A, Savolainen L, Aalton IT, Horsmanheimo M. Association of psychic stress with clinical severity and symptoms of psoriatic patients. Acta Derm Venereol (Stockh) 1996;76:467 – 71. [17] Fortune DG, Richards HL, Kirby B, Mcelhone K, Markham T, Rogers S, Main CJ, Griffiths CEM. Psychological distress impairs clearance of psoriasis in patients treated with photochemotherapy. Arch Dermatol 2003;139:752 – 6. [18] Richards HL, Fortune DG, Main CJ, Griffiths CEM. The contribution of perceptions of stigmatisation to disability in patients with psoriasis. J Psychosom Res 2001;50:10 – 5.

96

H.L. Richards et al. / Journal of Psychosomatic Research 58 (2005) 89 – 96

[19] Gupta MA, Gupta AK. Depression and suicidal ideation in dermatology patients with acne, alopecia areata, atopic dermatitis and psoriasis. Br J Dermatol 1998;139:846 – 50. [20] Rapp SR, Feldman SR, Exum ML, Fleisher AB, Reboussin DM. Psoriasis causes as much disability as other major medical diseases. J Am Acad Dermatol 1999;41:401 – 7. [21] Fortune DG, Richards HL, Main CJ, Griffiths CEM. Pathological worrying, illness perceptions and disease severity in patients with psoriasis. Br J Health Psychol 2000;5:71 – 82. [22] Fava GA, Perini GI, Santonastaso P, Fornasa CV. Life events and psychological distress in dermatologic disorders: psoriasis, chronic urticaria and fungal infections. Br J Med Psychol 1980;53:277 – 82. [23] Rubino IA, Sonnino A, Stefanato C, Pezzarossa B, Ciani N. Separation-individuation, aggression and alexithymia in psoriasis. Acta Derm Venerol 1989;146:87 – 90. [24] Picardi A, Pasquinnni P, Cattaruzza MS, Gaetano P, Balika G, Melchi C, Tiagi A, Camaioni D, Abeni D, Biondi M. Only limited support for a role of psychosomatic factors in psoriasis Results from a case control study. J Psychosom Res 2003;55:189 – 96. [25] Allegranti I, Gon T, Magaton-Rizzi G, Aguglia E. Prevelance of alexithymic characteristics in psoriatic patients. Acta Derm Venereol Suppl 1994;186:146 – 7. [26] Zeitlin SB, McNally RJ. Alexithymia and anxiety sensitivity in panic disorder and obsessive compulsive disorder. Am J Psychiatry 1993;150:658 – 60. [27] Fortune DG, Richards HL, Main CJ, Griffiths CEM. Psychological stress, distress and disability in patients with psoriasis: consensus and variation in the contribution of illness perceptions, coping and alexithymia. Br J Clin Psychol 2002;41:157 – 74. [28] Bagby RM, Parker JDA, Taylor GJ. The twenty item Toronto Alexithymia Scale: I Item selection and cross validation of the factor structure. J Psychosom Res 1994;38:23 – 32. [29] Bagby RM, Taylor GJ, Parker JDA. The twenty item Toronto Alexithymia Scale: II Convergent, discriminant and concurrent validity. J Psychosom Res 1994;38:33 – 40. [30] Salminen JK, Saarijarvi S, Aarela E, Toikka T, Kauhanen J. Prevalence of alexithymia and its association with sociodemographic variables in the general population of Finland. J Psychosom Res 1999;46:75 – 82. [31] Loas G, Corocs M, Stephan P, Pellet J, Bizouard P, Venisse JL, PerezDiaz F, Guelfi JD, Flament M, Jeammet P, the Re´seau INSERM no. 494013. Factorial structure of the 20-item Toronto Alexithymia Scale confirmatory factorial analyses in nonclinical and clinical samples. J Psychosom Res 2001;50:255 – 61. [32] Bressi C, Taylor G, Parker J, Bressi S, Brambilla V, Aguglia E, Allegranti I, Bongiorno A, Giberti F, Bucca M, Todarello O, Callegari S, Vender S, Gala C, Invernizzi G. Cross validation of the factor structure of the 20-item Toronto Alexithymia Scale: an Italian multicenter study. J Psychosom Res 1996;41:551 – 9. [33] Zech E, Luminet O, Rime B, Wagner H. Alexithymia and its measurement: confirmatory factor analysis of the 20 item Toronto

[34] [35]

[36]

[37] [38] [39] [40] [41]

[42]

[43]

[44]

[45] [46] [47]

[48]

[49] [50]

[51]

Alexithymia Scale and the Bermond–Vorst Alexithymia Questionnaire. Eur J Pers 1999;13:511 – 32. Haviland MG, Reise SP. Structure of the Twenty-item Toronto Alexithymia Scale. J Pers Assess 1996;61:116 – 25. Parker JDA, Taylor GJ, Bagby RM. The 20-item Toronto Alexithymia Scale: III Reliability and factorial validity in a community population. J Psychosom Res 2003;55:269 – 75. Mqller J, Bqhner M, Ellgring H. Is there a reliable factorial structure in the 20 item Toronto Alexithymia Scale? A comparison of factor models in clinical and normal adult samples. J Psychosom Res 2003;55:561 – 8. Fredriksson T, Pettersson U. Severe psoriasis: oral therapy with a new retinoid. Dermatologica 1978;157:238 – 44. Cattell RB. Factor analysis and introduction to essentials: the purpose and underlying models. Biometrics 1965;21:190 – 215. Kim JO, Meuller CW. Factor analysis: statistical methods and practical issues. London7 Sage Publications, 1978. Arbuckle JL, Wothke W. AMOS 40 users guide. Chicago7 SmallWaters, 1999. Brown TA. Confirmatory factor analysis of the Penn State Worry Questionnaire: multiple factors or method effects? Behav Res Ther 2003;41:1411–26. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus alternatives. Struct Equ Modeling 1999;6:1 – 55. Fukunishi I. Psychosomatic aspects of patients on haemodialysis: 3 Clinical usefulness of alexithymia. Psychother Psychosom 1990;54: 214 – 20. Kojima M, Frasure-Smith N, Lespe´rance F. Alexithymia following myocardial infarction Psychometric properties and correlates of the Toronto Alexithymia Scale. J Psychosom Res 2001;51:487 – 95. Cliff N. The Eigenvalues-greater-than-one rule and the reliability of components. Psychol Bull 1988;103:276 – 9. Comrey AL. Common methodological problems in factor analytic studies. J Consult Clin Psychol 1978;46:648 – 59. Porcelli P, Leoci C, Guerra V, Taylor GJ, Bagby RM. A longitudinal study of alexithymia and psychological distress in inflammatory bowel disease. J Psychosom Res 1996;41:569 – 73. Porcelli P, Zaka S, Leoci C, Centonze S, Taylor GJ. Alexithymia in inflammatory bowel disease: a case-control study. Psychother Psychosom 1995;64:49 – 53. Freyberger H. Supportive psychotherapeutic techniques in primary and secondary alexithymia. Psychother Psychosom 1977;28:337 – 42. Fortune DG, Richards HL, McElhone K, Kirby B, Main CJ, Griffiths CEM. Successful treatment of psoriasis improves psoriasis-specific but not more general aspects of patients’ well-being? Prospective study. Oral Presentation at the 10th International Congress of the European Society for Dermatology and Psychiatry, Brussels, May 2003. Marsh HW. Positive and negative global self-esteem: a substantively meaningful distinction or artifactors? J Pers Soc Psychol 1996;70: 810 – 9.