Multimethod alexithymia assessment in adolescents and young adults with a cannabis use disorder

Available online at www.sciencedirect.com

Comprehensive Psychiatry 49 (2008) 585 – 592 www.elsevier.com/locate/comppsych

Multimethod alexithymia assessment in adolescents and young adults with a cannabis use disorder Géraldine Dorarda,⁎, Sylvie Berthozb , Mark G. Havilandc , Olivier Phanb , Maurice Corcosb , Catherine Bungenera a Laboratory of Clinical Psychopathology and Neuropsychology, Paris Descartes University, Paris, France Department of Psychiatry for Adolescents and Young Adults, Institut Mutualiste Montsouris, Paris Descartes University, Paris, France c Department of Psychiatry, Loma Linda University School of Medicine, Loma Linda, CA, USA Received 17 March 2008; revised 22 May 2008; accepted 29 May 2008

b

Abstract The value of alexithymia assessments in medical and psychiatric research is well documented, but such assessments in cannabis abusers are scarce. Moreover, despite repeated calls for multimethod alexithymia evaluations, researchers typically use 1 self-report only: the 20-item Toronto Alexithymia Scale. Herein, we evaluated (1) the psychometric properties of the Observer Alexithymia Scale (OAS), (2) the correspondence between 3 alexithymia measures, (3) OAS raters' affect and its relationship to OAS scores, and (4) cannabis abusers' alexithymic features. Eighty-seven cannabis abusers completed self-reports measuring alexithymia (Toronto Alexithymia Scale, BermondVorst Alexithymia Questionnaire–B), depression (13-item Beck Depression Inventory), and anxiety (State and Trait Anxiety Inventory–Form Y) and asked relatives to rate them using the OAS. The raters also completed the self-report scales. The OAS met acceptable reliability and validity standards, with the exception of relatively low interrater reliability for one of its subscales. Rater affect appeared to influence OAS scores, albeit slightly. Patients' OAS scores were higher than scores reported for people-in-general samples and lower than those for outpatient clinical samples. Alexithymia rates were similar to those previously reported in cannabis abusers. Our results demonstrated the adequacy and appropriateness of the OAS in these (and related) clinical samples, which may encourage multimethod alexithymia assessments in both research and clinical practice. © 2008 Elsevier Inc. All rights reserved.

1. Introduction Alexithymia, which literally translates to “no words for mood” (introduced by Sifneos in 1973 [1]), refers to a specific emotional disturbance characterized by an inability to identify internal affective states and describe them verbally [2]. In clinical settings, alexithymic patients present with an inability to distinguish emotions from physical sensations and pensée opératoire [3], a distinctive cognitive

⁎ Corresponding author. Laboratoire de Psychopathologie et Neuropsychologie Cliniques, Université Paris Descartes-Institut Henri Piéron, 92100 Boulogne Billancourt, France. Tel.: +33 1 55 20 54 03; fax: +33 1 55 20 59 56. E-mail addresses: [email protected] (G. Dorard), [email protected] (S. Berthoz), [email protected] (M.G. Haviland), [email protected] (O. Phan), [email protected] (M. Corcos), [email protected] (C. Bungener). 0010-440X/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.comppsych.2008.05.001

style characterized by a preoccupation with the details of external events as well as a lack of fantasy. Alexithymia rates in the general population have been reported to be 9% to 17% for men and 5% to 10% for women [4], whereas estimates are as high as 70% in some clinical groups (eg, women with anorexia nervosa [5]). Alexithymia is considered a risk factor for physical illness and psychiatric disorders [6,7]. Numerous studies have reported high rates and levels of alexithymia in adult substance abusers, and some speculate that substances are used to compensate for deficits in emotional self-awareness (eg, [8]). Using selfreport questionnaires (the 20- and 26-item Toronto Alexithymia Scales [TASs]), alexithymia rates ranged from 39% to 54.5% in adults diagnosed with substance abuse or dependence [6,7]. In a French study [9], an alexithymia rate of 43.9% was found among 68 substance abusers (substance not specified, age range = 15-24 years). To our knowledge, only Troisi et al [10] measured alexithymia in young

586

G. Dorard et al. / Comprehensive Psychiatry 49 (2008) 585–592

cannabis-abusing/dependent subjects; and they reported the prevalence of alexithymia (using the 20-item TAS [TAS-20]) among their 88 subjects with a cannabis abuse or dependence diagnosis to be 30%. This is about twice the rate of alexithymia reported recently by Säkkinen et al [11], who found a prevalence of 14.6% among boys and 17.3% among girls in a sample of “general adolescents” (882 students, age range = 12-17 years). Alexithymia assessment has become increasingly common in medical and psychiatric research, and numerous psychometrically sound tools have been developed [12-14]. Presently, the TAS-20 and the Bermond-Vorst Alexithymia Questionnaire (BVAQ) are the most popular reliable and valid alexithymia self-report measures [15,16]. Considering the nature of alexithymic deficits, it is desirable to gather information from both individuals themselves (self-report) and observers (observer reports); and thus, some researchers have developed and modified observer alexithymia reports (eg, Beth Israel Hospital Psychosomatic Questionnaire [BIQ] [1], the modified BIQ [17], and the California Q-set Alexithymia Prototype (CAQ-AP) [18]). These are not particularly good screening instruments, however. The BIQs are most appropriately used by professional raters, typically clinicians who know their patients very well. The CAQ-AP, on the other hand, can be used by professional or lay raters; but the procedure requires training and practice. To simplify the alexithymia observer rating process, Haviland et al [19-21] developed the Observer Alexithymia Scale (OAS), which may be completed by clinicians who know their patients well or by subjects' acquaintances and relatives in 15 minutes or less. The OAS consists of 33 items (content taken from the CAQ-AP) and captures 5 alexithymic expressions: lacking skill in interpersonal matters and relationships; poor stress tolerance, insight, and self-understanding; health worries and physical problems; lacking humor; and excessive self-control. Previous studies have demonstrated the good psychometric properties of the OAS in the general population [20-23] as well as in clinical samples [19,24]. Importantly, Mueller et al [25] showed that psychosomatic outpatients rated high (vs low) alexithymia by observers using the OAS (but not by TAS-20 self-ratings) had impaired performances on an emotional Stroop task. This finding underscores the potential importance of observer alexithymia reports. The main purpose of the present study was to further evaluate the psychometric properties of the OAS in a clinical sample in which we expected high rates and levels of alexithymia. We first evaluated reliability and hypothesized that the OAS would have good internal consistency and good interrater reliability. Second, we evaluated the correspondence among 3 alexithymia measures: 1 observer report (the OAS) and 2 self-reports, the TAS-20 and BVAQ-B. We expected positive associations between these 3 alexithymia measures, in other words, that the OAS

would have good convergent and discriminant validity. Finally, we explored whether rater affect bore any relationship to OAS scores and described the alexithymic features of young cannabis abusers. 2. Method 2.1. Participants Eighty-seven adolescents and young adults, with a mean age of 18.4 years (SD = 2.8; range = 14-25; 67 male, 20 female), each diagnosed with a substance use (cannabis) disorder, participated in the study. Regarding their educational level, 16.1% were in middle school (14-15 years old), 35.6% were in high school (15-18 years old), 25.2% were in vocational school (14-18 years old), and 23% were attending a university (18 years old and older). The classification of the French National Institute for Statistics and Economic Studies was used to categorize the economic status of the participants' families. When both parents were working (63.2%), the higher professional status was used. Among the 87 participants, 12.6% had parents working as artisans, storekeepers, tradesmen, or heads of business; 48.3% had parents who were high-level white collar workers or intellectual professionals; 18.3% had parents working in intermediate professions (eg, teachers, intermediate workers in health and social sectors, intermediate administrative and business workers); 13.8% had parents working as public employees, police officers, and military staff members, or in sales; 3.4% had both parents unemployed; 1.1% had parents working as building and industrial workers or drivers; 1.1% had both parents retired; and for 1.1%, the professional status of the parents was not specified. Among our patients, 83 met the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), diagnostic criteria for cannabis dependence, and 4 met the DSM-IV criteria for cannabis abuse only (using the appropriate section of the Mini-International Neuropsychiatric Interview [26]). The patients had been using cannabis regularly (average length of time was 40.3 months, SD = 29.2); and during the previous month, they had smoked on average 4.8 (±3.4) joints per day. Participants were outpatients recruited from the addiction unit of the Institut Mutualiste Montsouris, Paris. Exclusion criteria were mental retardation, organic brain disease, chronic or severe somatic disorder, psychotic disorder, and an inability to read or fill out the questionnaires. Among the 87 patients, we collected OAS data for 66 of them. Sixty-two were evaluated by their mother, 2 by their father, 1 by a brother, and 1 by a partner. Of the 66, 39 were evaluated by 2 raters (36 by the 2 parents, 1 by the mother and a close friend, 1 by the mother and the partner, and 1 by the mother and the sister). From the 2-rater assessments, we formed the more reliable (averaged) composite scores [27], a convention we [22] and others [28] have followed. The

G. Dorard et al. / Comprehensive Psychiatry 49 (2008) 585–592

mean age of the raters (n = 105) was 49.5 years (SD = 8.9, range = 20-64). The protocol was approved by the Pitié-Salpêtrière Hospital Ethical Committee, and all subjects (and at least 1 parent for those younger than 18 years) gave their written informed consent. Patients freely and explicitly agreed to have the observer raters participate in the study. 2.2. Materials The OAS [22] is a 33-item, observer-rated alexithymia measure. Each item is rated on a 4-point scale (from 0 = “never, not at all like the person” to 3 = “all of the time, completely like the person”). The OAS has 5 subscales: distant (O1), uninsightful (O2), somatizing (O3), humorless (O4), and rigid (O5). Reliability and validity data of the OAS French version are acceptable [22]. The TAS-20 [29] captures 3 interrelated and core alexithymic features: difficulty identifying feelings (F1), difficulty describing feelings (F2), and external thinking (F3). We used the alexithymia cutoff scores established for French samples (nonalexithymic, TAS-20 ≤ 44; intermediate, 45 ≤ TAS-20 ≤ 55; and alexithymic, TAS-20 ≥ 56) [30]. The BVAQ evolved from the Amsterdam Alexithymia Scale [31]. The original BVAQ is a 40-item self-report questionnaire, with each item rated on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree). Two short versions of the BVAQ are available: the BVAQ-A (items 1-20) and the BVAQ-B (items 21-40). Several studies have demonstrated that the BVAQ-B has better psychometric properties than the BVAQ-A (eg, in French-speaking samples [31,32]); thus, we used the BVAQ-B in the present study. Factor analyses have consistently supported a 5-factor structure: verbalizing (B1), fantasizing (B2), identifying (B3), emotionalizing (B4), and analyzing (B5). Three BVAQ-B factors are assumed to correspond to the TAS-20 factors: B1 and F2, B3 and F1, and B5 and F3. We used the alexithymia cutoff scores established for French samples (nonalexithymic, BVAQ-B ≤ 43; intermediate, 52 ≤ BVAQB ≤ 44; and alexithymic, BVAQ-B ≥ 53) [33]. For depression, we used the 13-item Beck Depression Inventory (BDI-13) [34]. Individuals are asked to respond to statements on the basis of how they have felt over the past week. Depression severity was determined with the validated French cutoff scores: no depressive symptoms ≤ 3; 4 ≤ low depressive symptoms ≤ 7; 8 ≤ moderate depressive symptoms ≤ 15; and severe depressive symptoms ≥6 [34]. For anxiety, we used the State and Trait Anxiety Inventory (STAI)–Form Y [35]. The subjects are asked to report the extent of their anxiety at present (S-STAI) and the intensity of their anxiety in general (T-STAI). 2.3. Statistical analyses Descriptive statistics for continuous and categorical measures were calculated. To evaluate OAS internal consistency, we used Cronbach α; and to determine OAS

587

and subscale relationships, we calculated Pearson correlation coefficients. To establish interrater reliability, we calculated intraclass correlation coefficients with a 2-way mixed model (consistency definition). We used Pearson correlation coefficients to evaluate the correspondence among the 3 alexithymia measures. To test the discriminant validity of the OAS, we used analysis of variance (ANOVA), with alexithymia as the between-group factor (French TAS-20 and BVAQ-B cutoff scores) and OAS total score as the criterion variable (followed by Bonferroni post hoc tests). We used Pearson correlation coefficients to assess the correspondence between OAS scores and the raters' affect scores. Finally, partial correlations were calculated to control for the potential effect of rater affect (BDI-13, S- and T-STAI, TAS-20, and BVAQ-B scores) on the OAS scores. Data were analyzed using SPSS 11.5 (SPSS, Chicago, IL). 3. Results Participants' self-report (n = 87) and OAS scores (n = 66, for which 39 had a composite OAS score) are presented in Table 1. The OAS total scores ranged from 21 to 73 (mean ± SD = 39.9 ± 10.9). The OAS total and subscale intercorrelations (Table 2) ranged from 0.49 (O3: somatizing) to 0.78 (O2: uninsightful). Subscale intercorrelations ranged from 0.04 (distant [O1]/somatizing [O3]) to 0.65 (distant [O1]/humorless [O4]).

Table 1 Patients’ (n = 87) and raters’ (n = 105) depression, anxiety, and alexithymia scores

BDI-13 total score State STAI total score Trait STAI total score TAS-20 total score Difficulties identifying feelings (F1) Difficulties describing feelings (F2) External thinking (F3) BVAQ-B total score Verbalizing (B1) Fantasizing (B2) Identifying (B3) Emotionalizing (B4) Analyzing (B5) OAS total score (n = 66)a Distant (O1) Uninsightful (O2) Somatizing (O3) Humorless (O4) Rigid (O5)

Patients (mean ± SD/range)

Raters (mean ± SD/range)

9.06 ± 5.5/0-22 43.6 ± 12.1/20-76 50.4 ± 11.8/20-75 52.1 ± 10.2/25-76 18.7 ± 5.6/7-31

5.1 ± 4.1/0-21 40.8 ± 12.5/20-73 41.8 ± 10.1/20-72 42.2 ± 11.3/21-74 14.3 ± 5.2/7-31

14.4 ± 4.5/5-25

12.4 ± 4.9/5-24

19.04 ± 4.3/10-30 51.2 ± 9.3/30-74 11.7 ± 3.5/4-20 8.9 ± 3.04/4-17 11.4 ± 3.1/4-17 10.2 ± 2.7/4-17 8.9 ± 3.4/4-20 39.9 ± 10.9/21-73

15.5 ± 4.2/8-25 45.5 ± 9.7/29-72 10.1 ± 3.6/4-20 10.6 ±3.6/4-19 9.0 ± 3.1/4-17 8.7 ± 2.9/4-17 7.1 ± 2.5/4-16

11.7 ± 4.5/2-23 13.7 ± 3.5/7-24 4.9 ± 2.9/0-12 5.3 ± 2.9/0-12 4.2 ± 2.2/0-12

a Among the 66 patients, a composite OAS score was calculated for the 39 patients evaluated by 2 raters (see “Method” section).

588

OAS O1 O2 O3 O4 O5 TAS-20 F1 F2 F3 BVAQ-B B1 B2 B3 B4 B5

OAS totala

O1

O2

O3

O4

O5

TAS-20 total

F1

F2

F3

BVAQ-B total

B1

B2

B3

B4

.758⁎⁎⁎⁎ .700⁎⁎⁎⁎ .491⁎⁎⁎⁎ .780⁎⁎⁎⁎ .627⁎⁎⁎⁎ .410⁎⁎⁎⁎ .396⁎⁎⁎⁎ .223 .211 .464⁎⁎⁎ .214 .289⁎ .319⁎⁎⁎ .173 .351⁎⁎⁎

.299⁎ .039 .652⁎⁎⁎⁎ .327⁎⁎ .308⁎ .221 .142 .283⁎ .394⁎⁎⁎⁎ .200 .190 .266⁎ .184 .305⁎

.319⁎⁎ .412⁎⁎⁎⁎ .313⁎⁎ .305⁎ .408⁎⁎⁎⁎ .067 .115 .306⁎ .137 .233 .256⁎ .095 .175

.147 .343⁎⁎⁎ .259⁎ .268⁎ .282⁎ -.036 .156 .091 .155 .073 -.033 .148

.374⁎⁎⁎ .248⁎ .229 .159 .117 .396⁎⁎⁎⁎ .165 .290⁎ .333⁎⁎ .233 .160

.253⁎ .212 .132 .177 .282⁎ .099 .090 .096 .067 .435⁎⁎⁎⁎

.787⁎⁎⁎⁎ .771⁎⁎⁎⁎ .539⁎⁎⁎⁎ .536⁎⁎⁎⁎ .628⁎⁎⁎⁎ .155 .477⁎⁎⁎⁎ -.035 .288⁎⁎

.461⁎⁎⁎⁎ .073 .311⁎⁎⁎ .379⁎⁎⁎⁎ .002 .564⁎⁎⁎⁎ -.185 .099

.182 .451⁎⁎⁎⁎ .632⁎⁎⁎⁎ .169 .316⁎⁎⁎ .011 .147

.397⁎⁎⁎⁎ .334⁎⁎⁎ .190 .061 .149 .405⁎⁎⁎⁎

.686⁎⁎⁎⁎ .560⁎⁎⁎⁎ .471⁎⁎⁎⁎ .517⁎⁎⁎⁎ .702⁎⁎⁎⁎

.183 .303⁎⁎⁎ .108 .339⁎⁎⁎⁎

-.005 .260⁎⁎ .255⁎

-.048 .114

.317⁎⁎⁎

Among the 66 patients, a composite OAS score was calculated for the 39 patients evaluated by 2 raters (see “Method” section). ⁎ P ≤ .05. ⁎⁎ P ≤ .01. ⁎⁎⁎ P ≤ .005. ⁎⁎⁎⁎ P ≤ .001. a

G. Dorard et al. / Comprehensive Psychiatry 49 (2008) 585–592

Table 2 Patients’ OAS, TAS-20, and BVAQ-B total and subscale intercorrelations

G. Dorard et al. / Comprehensive Psychiatry 49 (2008) 585–592 Table 3 Cronbach α (n = 66) and interrater intraclass coefficients (n = 39) for the OAS total and subscales scores Cronbach α coefficient Interrater intraclass coefficient

OAS total

O1

O2

O3

O4

O5

.85 .70

.78 .59

.59 .62

.64 .76

.75 .73

.54 .35

Cronbach α coefficients (n = 66) and interrater intraclass coefficients (39 pairs) for the OAS total score and the 5 subscales are presented in Table 3. Patients' OAS, TAS-20, and BVAQ-B total and subscale intercorrelations are presented in Table 2 (we note only the significant correlations at P b .05 in this section). The OAS total score was positively and moderately correlated with TAS-20 (0.41) and BVAQ-B (0.46) total scores. Regarding the OAS and TAS-20 subscales, the OAS total and F1 (difficulty identifying feelings) scores were positively correlated. The O1 score was positively correlated with the TAS-20 total and F3 (external thinking) scores. The O2 score was positively correlated with the TAS-20 total and F1 scores. The O3 score was positively correlated with the TAS-20 total score and with the F1 and F2 (difficulty describing feelings) scores. The O4 and O5 scores were positively correlated with the TAS-20 total score. Regarding the OAS and BVAQ-B subscales, the OAS total score was positively correlated with the B2 (fantasizing), B3 (identifying), and B5 (analyzing) scores. The O1 score was positively correlated with the BVAQ-B total, B3, and B5 scores. Positive correlations were observed between O2 and the BVAQ-B total and B3 scores. The O4 and the BVAQ-B total, B2, and B3 scores were positively correlated. The O5 score was positively correlated with the BVAQ-B total and B5 scores. Correlations between OAS scores (n = 66, for which 39 had a composite OAS score) and patients' BDI-13, S-STAI, and T-STAI scores are presented in Table 4. Only the patients' trait anxiety scores were positively and significantly correlated with the OAS total score. Partial correlations (adjusting for patients' BDI-13 and STAI scores) showed that the OAS total score remained significantly correlated with the TAS-20 and BVAQ-B total scores (respectively, 0.33, P = .013 and 0.37, P = .005). The analyses of the correspondence between the 2 alexithymia self-reports (n = 87, Table 2) showed a significant positive correlation between the TAS-20 and BVAQ-B total scores (0.54), as well as among their corresponding factors scores. All of the TAS-20 subscale scores were significantly and positively correlated with the BVAQ-B total score; however, only the BVAQ-B subscale scores corresponding to those of the TAS-20 were significantly and positively correlated to the TAS-20 total score. Using TAS-20 scores as the between-group factor, the ANOVA revealed a significant main effect of alexithymia category (alexithymic, n = 25; intermediate, n = 30; nonalexithymic, n = 11) on the OAS total score (F2,63 = 6.001, P = .004). Post hoc analyses (with Bonferroni

589

corrections) showed that alexithymic patients had a higher mean OAS total score than did nonalexithymic patients (mean difference ± SD = 12.2 ± 3.7, P = .005) but not a significantly higher OAS total score than those classified intermediate (6.5 ± 2.8, P = .067). Intermediate participants did not have a significantly higher OAS total score than nonalexithymics (5.7 ± 3.6, P = .363). Mean OAS total score (±SD) for the nonalexithymic, intermediate, and alexithymic subjects was 32.7 (7.6), 38.4 (12.05), and 44.9 (8.7), respectively. Using BVAQ-B scores for the between-group factor, the ANOVA revealed a significant main effect of alexithymia category (alexithymic, n = 28; intermediate, n = 26; nonalexithymic, n = 12) on the OAS total score (F2,63 = 8.343, P b .001). Post hoc analyses showed that alexithymic patients had a higher mean OAS total score than did nonalexithymic patients (mean difference ± SD = 13.1 ± 3.4, P b .001), as well as those classified intermediates (7.5 ± 2.7, P = .021). Intermediate participants did not have a significantly higher average OAS total score than the nonalexithymics (5.6 ± 3.4, P = .341). Mean OAS total score (±SD) for the nonalexithymic, intermediate, and alexithymic subjects was 32.2 (9.2), 37.7 (8.8), and 45.2 (11.1), respectively. Descriptive statistics for the raters on the various selfreport questionnaires are presented in Table 1. For the following analyses, we used the scores of each individual OAS rater and not the composite scores (n = 105). Using the BDI-13 cutoff scores, we found that 1.9% (n = 2) of the raters had severe depressive symptoms, 21.9% (n = 23) had moderate depressive symptoms, 36.2% (n = 38) had mild depressive symptoms, and 40% (n = 42) had no depressive symptoms. Using the TAS-20 cutoff scores, 13.3% (n = 14) of the raters were categorized as alexithymic, 22.9% (n = 24) were intermediates, and 63.8% (n = 67) were nonalexithymic. Using the BVAQ-B cutoff scores, 24.8% (n = 26) of the raters were categorized as

Table 4 Patients’ (n = 66) and raters’ (n = 105) depression, anxiety, and alexithymia intercorrelations BDI-13

State STAI

Trait STAI

Patients (n = 66) OASa TAS-20 BVAQ-B

.169 .489⁎⁎⁎⁎ .192⁎

.204 .362⁎⁎⁎⁎ .155⁎

.369⁎⁎⁎ .498⁎⁎⁎⁎ .221⁎⁎⁎

Raters (n = 105) OAS TAS-20 BVAQ-B

.352⁎⁎⁎⁎ .248⁎ .005

.313⁎⁎⁎ .111 .076

.267⁎⁎ .262⁎⁎ .074

a Among the 66 patients, a composite OAS score was calculated for the 39 patients evaluated by 2 raters (see “Method” section). ⁎ P ≤ .05. ⁎⁎ P ≤ .01. ⁎⁎⁎ P ≤ .005. ⁎⁎⁎⁎ P ≤ .001.

590

G. Dorard et al. / Comprehensive Psychiatry 49 (2008) 585–592

alexithymic, 30.5% (n =32) were intermediate, and 44.8% (n =47) were nonalexithymic. The OAS total score was positively and significantly correlated with the rater BDI-13 and State and Trait STAI scores (Table 4) but not with TAS-20 or BVAQ-B scores (respectively, r = 0.05, P N .05 and r = 0.03, P N .05). The partial correlations—adjusting for the rater affect (BDI-13, S-STAI, and T-STAI)—showed that the OAS total score remained significantly correlated to the patients' TAS-20 and BVAQ-B total scores (respectively, r = 0.32, P = .001 and r = .36, P b .001). Descriptive statistics for all cannabis abusers (n = 87) on the self-report questionnaires are presented in Table 1. Using the TAS-20 cutoff scores, 35.6% (n = 31) of the patients were categorized as alexithymic, 43.7% (n = 38) were intermediate, and 20.7% (n = 18) were nonalexithymic. Using the BVAQ-B cutoff scores, 39.1% (n = 34) of the patients were categorized as alexithymic, 40.2% (n = 35) were intermediate, and 20.7% (n = 18) were nonalexithymic.

4. Discussion Our first aim was to further evaluate the OAS's psychometric properties in a clinical sample of adolescents/ young adults with a cannabis abuse/dependence disorder. The OAS showed acceptable internal consistency; however, α coefficients for the uninsightful and rigid subscales were lower than those reported in normative and clinical studies [21-23]. The 5 OAS subscale scores were significantly and positively correlated with the OAS total score. Consistent with previous studies, the somatizing factor had the lowest correlation with the OAS total score [19-24]. Regarding the OAS subscale intercorrelations, the lowest (distant/somatizing) was consistent with all previous reported data [19-24]; and the highest (humorless/distant) was consistent with the Haviland et al data [19-21]. With the exception of the somatizing/distant and the somatizing/ humorless correlations, all of the OAS subscale scores were strongly and positively correlated. Our results suggest that the OAS has acceptable interrater reliability, with the 1 exception of the OAS rigid intraclass coefficient (0.35). In previous studies [23,24], however, the O5 interrater reliability coefficients were acceptable (respectively, 0.64 and 0.74). Interestingly, it is the second clinical study (French data) that reported the weakest interrater reliability coefficient, although it was not for the same OAS subscale. Indeed, whereas Yao et al [23] reported relatively high interrater reliability coefficients for the 5 OAS subscales in the Chinese normative sample, Berthoz et al [24] reported a weaker one for the uninsightful subscale (0.32) in a clinical sample of patients with eating disorder. Yet, as these authors have written [24], “discrepant ratings are neither unexpected nor necessarily undesirable; composite ratings are more reliable [27] and agreement is only a proxy for accuracy [36].” For the 4 other subscales, our

coefficients were roughly comparable with those reported in previous studies [24]. With respect to the correspondence among the 3 alexithymia measures, as expected, OAS/TAS-20 and OAS/BVAQ-B total scores were moderately and positively correlated (respectively, 0.41 and 0.46); and the coefficients were slightly higher than those reported previously. In the OAS French validation study [22] (people in general; 132 women, 27 men; mean age = 24.3 years; range = 18-60), the OAS/TAS-20 total r was 0.31. Furthermore, in a French clinical study [24] of 75 women with eating disorders (mean age = 18.8 years, range = 13-31), the OAS/TAS-20 total r was 0.39. This is the second study assessing the correlation between OAS and BVAQ-B scores. In the first [24], the OAS/ BVAQ-B total r was 0.31. We, however, found positive and significant correlations between the OAS total score and the BVAQ-B subscales fantasizing and analyzing. Moreover, the 5 OAS subscale scores were positively and significantly correlated with the 2 alexithymia self-report total scores, with the exception of the somatizing subscale with the BVAQ-B total score. These data on the correspondence between the OAS and the 2 self-report scores suggest that the OAS has good convergent validity. Regarding the correspondence between the 2 alexithymia self-report measures, we replicated the previously reported positive relationship between the TAS-20 and BVAQ-B total scores and their corresponding subscales scores. This is a point of interest only; the appropriateness/adequacy of the BVAQ-B for research or practice is beyond the scope of our study. Consistent with previous data [22,24], which were obtained in general and clinical samples, alexithymic participants (TAS-20 scores) had significantly higher OAS scores than did nonalexithymic individuals. Interestingly, here, using the recently published BVAQ-B French cutoff scores [33], we showed that the OAS discriminated not only the alexithymic participants from the nonalexithymic ones, but also the alexithymic participants from the intermediates. Therefore, we believe that the OAS has shown acceptable discriminant validity. This is the first study assessing the potential influence of rater affect on OAS scores. Considering the nature of alexithymic deficits, we questioned whether a rater's awareness of his/her own feelings would influence his/her judgments about another person's affective state. Interestingly, the OAS score was correlated with the raters' negative affect but not with their alexithymia scores. This latter result might be explained by the fact that, in the present study, only 13.3% of the raters were categorized as alexithymic (on the basis of their TAS-20 scores). Clearly, however, the potential influence of raters' alexithymia scores on OAS scores is a complex issue; and it should be further investigated in future studies. Regarding the effect of raters' negative affect (depression and anxiety), the correlation between the OAS total score and the patients' alexithymia self-report scores remained significant after adjusting for rater affect. This is a

G. Dorard et al. / Comprehensive Psychiatry 49 (2008) 585–592

novel and potentially useful finding. Given the potential effect of substances on the affective and cognitive state of substance users, the validity of patients' alexithymia selfreports may be questionable. Therefore, observer reports should be promoted in work with substance abusers. The OAS may help describe in richer and more nuanced terms the emotional profile of these young drug users, profiles that may be useful for identifying the emotional risk factors underlying addictive disorders during adolescence and young adulthood and, possibly, preventing progression in treatment. More than 30 years ago, Krystal and Raskin [37] suggested that affect deficits play a role in the development of drug dependence. Since then, although alarming data from several epidemiologic studies repeatedly have underscored that the use of illicit substances among adolescents and young adults continues to be a distressing international public health concern (eg, [38]), few researchers have studied emotion regulation deficits in this population. Dorard et al [39] have recently demonstrated that cannabis dependence in adolescents and young adults is related to severe psychological distress and specific emotional dimensions (eg, trait anxiety, anhedonia, sensation seeking). Notably, our patients' average OAS total score was more than 10 points higher than the scores reported in the Frenchand English-speaking normative samples [19,22], albeit lower than average scores reported for other clinical samples (differences ranged from 2 to 9 points) [20,21,24]. Participants' average TAS-20 and BVAQ-B total scores were higher than scores reported in French-speaking normative samples (for the BVAQ-B, the score differences ranged from 7 to 10 points [33,40]; for the TAS-20, the score differences ranged from 5 to 10 points [33,40,41]), yet they were slightly lower than those observed in patients with eating disorder (for the BVAQ-B, the score differences ranged from 1 to 2 points; for the TAS-20, the score differences ranged from 4 to 5 points [33,40]). These data demonstrate the good concordance between the OAS and the self-reports for estimating the level of alexithymia and further support the OAS validity. The prevalence of alexithymia (TAS-20 and BVAQ-B based: 35.6% and 39.1%, respectively) was higher than that recently reported for normal adolescents. As noted earlier, Säkkinen et al [11] reported TAS-20 rates of 14.6% for male subjects and 17.3% for female subjects in the 12- to 17-year age range. Another recent study using the TAS-20, conducted among 6000 Finnish adolescents 15 to 16 years old [42], found a prevalence of 6.9% for male subjects and 9.5% for female subjects. Both of our alexithymia rates (TAS-20 and BVAQ-B based) are greater than that in the Troisi et al [10] study (which is, to our knowledge, the only other study of alexithymia in young people with cannabis disorders); the rate was 30% (with the TAS-20). On the other hand, our alexithymia rates were lower than that in the Farges et al study [9]: 43.9% alexithymic. Despite these small

591

discrepancies, it seems reasonable to conclude that fully one third of the present subjects may be at risk for the many and varied problems associated with poor emotion regulation [8]. The combination of self and observer alexithymia rating scales may help us better understand the links between emotion regulation deficits and mental/addictive disorders and physical illness. A combination may be particularly important for assessing alexithymia in individuals who are under the effects of substances; the validity of their selfreports are often and appropriately criticized. Given the established reliability and validity of the TAS-20 and accumulating evidence of the OAS's psychometric adequacy, it seems reasonable to recommend that the 2 be used together to tap both core alexithymic deficits (TAS-20) and their expressions (OAS) [43]. Acknowledgment Géraldine Dorard was funded by the French Ministry of Research. A part of this study was presented at the 26th European Conference on Psychosomatic Research, Cavtat/ Dubrovnik, September 2006. References [1] Sifneos PE. The prevalence of “alexithymic” characteristics in psychosomatic patients. Psychother Psychosom 1973;22:255-62. [2] Nemiah JC, Freyberger H, Sifneos PE. Alexithymia. A view of the psychosomatic process. In: Hill O, editor. Modern trends in psychosomatic medicine. New York: Appleton-Century-Crofts; 1976. p. 430-9. [3] Marty P, M'Uzan de M. La pensée opératoire. Revue Française de Psychanalyse 1963;27:345-56. [4] Mattila AK, Ahola K, Honkonen T, Salminen JK, Huhtala H, Joukamaa M. Alexithymia and occupational burnout are strongly associated in working population. J Psychosom Res 2007;62:657-65. [5] Bourke MP, Taylor GJ, Parker JDA, Bagby RM. Alexithymia in women with anorexia nervosa. A preliminary investigation. Br J Psychiatry 1992;161:240-3. [6] Taylor GJ, Bagby RM, Parker JDA. Disorders of affect regulation. Cambridge: Cambridge University Press; 1997. [7] Corcos M, Speranza M. Psychopathologie de l'alexithymie. Paris: Dunod; 2003. [8] Taylor GJ, Bagby RM, Parker JDA. Substance use disorders. In: Taylor GJ, Bagby RM, Parker JDA, editors. Disorders of affect regulation. Cambridge: Cambridge University Press; 1997. p. 166-89. [9] Farges F, Corcos M, Speranza M, Loas G, Perez-Diaz F, Venisse JL, et al. Alexithymia, depression and drug addiction. Encephale 2004; 30:201-11. [10] Troisi A, Pasini A, Saracco M, Spalletta G. Psychiatric symptoms in male cannabis users not using other illicit drugs. Addiction 1998;93: 487-92. [11] Säkkinen P, Kaltiala-Heino R, Ranta K, Haataja R, Joukamaa M. Psychometric properties of the 20-item Toronto Alexithymia Scale and prevalence of alexithymia in Finnish adolescent population. Psychosomatics 2007;48:154-61. [12] Bagby RM, Taylor GJ. Measurement and validation of the alexithymia construct. In: Taylor GJ, Bagby RM, Parker JDA, editors. Disorders of affect regulation: alexithymia in medical and psychiatric illness. Cambridge: Cambridge University Press; 1997. p. 46-66.

592

G. Dorard et al. / Comprehensive Psychiatry 49 (2008) 585–592

[13] Lumley MA, Neeley LC, Burger AJ. The assessment of alexithymia in medical settings: implication for understanding and treating health problems. J Pers Assess 2007;89:230-46. [14] Taylor GJ, Bagby RM, Luminet O. Assessment of alexithymia: selfreport and observer-rated measures. In: Bar-On R, Parker JDA, editors. The handbook of emotional intelligence. San Fransisco: Jossey-Bass; 2000. p. 301-19. [15] Kooiman CG, Spinhoven P, Trijsburg RW. The assessment of alexithymia. A critical review of the literature and a psychometric study of the Toronto Alexithymia Scale–20. J Psychosom Res 2002; 53:1083-90. [16] Morera OF, Culhane SE, Watson PJ, Skewes MC. Assessing the reliability and validity of the Bermond-Vorst Alexithymia Questionnaire among US Anglo and US Hispanic samples. J Psychosom Res 2005;58:289-98. [17] 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. [18] Haviland MG, Reise SP. A California Q-set Alexithymia Prototype and its relationship to ego-control and ego-resiliency. J Psychosom Res 1996;41:597-608. [19] Haviland MG, Warren WL, Riggs ML. An observer scale to measure alexithymia. Psychosomatics 2000;41:385-92. [20] Haviland MG, Warren WL, Riggs ML, Gallacher M. Psychometric properties of the Observer Alexithymia Scale in a clinical sample. J Pers Assess 2001;77:176-86. [21] Haviland MG, Warren WL, Riggs ML, Nitch SR. Concurrent validity of two observer-rated alexithymia measures. Psychosomatics 2002;43: 472-7. [22] Berthoz S, Haviland MG, Riggs ML, Perdereau F, Bungener C. Assessing alexithymia in French-speaking samples: psychometric properties of the Observer Alexithymia Scale–French translation. Eur Psychiatry 2005;20:497-502. [23] Yao S, Yi J, Zhu X, Haviland MG. Reliability and factorial validity of the Observer Alexithymia Scale–Chinese translation. Psychiatry Res 2005;134:93-100. [24] Berthoz S, Perdereau F, Godart N, Corcos M, Haviland MG. Observerand self-rated alexithymia in eating disorder patients: levels and correspondence among three measures. J Psychosom Res 2007;62: 341-7. [25] Mueller J, Alpers GW, Reim N. Dissociation of rated emotional valence and Stroop interference in observer-rated alexithymia. J Psychosom Res 2006;61:261-9. [26] Lecrubier Y, Sheehan DV, Weiller E, Amorin P, Bonora I, Harnett Sheehan KH, et al. The Mini International Neuropsychiatric Interview (MINI) a short diagnostic structured interview: reliability and validity according to the CIDI. Eur Psychiatry 1997;5:224-31. [27] Cheek JM. Aggregation, moderator variables, and the validity of personality test: a peer rating study. J Pers Soc Psychol 1982;43: 1254-69. [28] Lumley MA, Gustavson BJ, Partridge T, Labouvie-Vief G. Assessing alexithymia and related emotional ability constructs using multiple

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36] [37] [38] [39]

[40]

[41]

[42]

[43]

methods: interrelationships among measures. Emotion 2005;5: 329-42. Loas G, Fremaux D, Marchand MP. Factorial structure and internal consistency of the French version of the twenty-item Toronto Alexithymia Scale in a group of 183 healthy probands. Encephale 1995;21:117-22. Loas G, Otmani O, Fremaux D, Lecercle C, Duflot M, Delahousse J. Criterion validity, reliability and cutoff scores of the Toronto Alexithymia Scales (TAS and TAS-20). Encephale 1996;12:35-40. Zech E, Luminet O, Rimé B, Wagner H. Alexithymia and its measurement: confirmatory factor analyses of the 20-item Toronto Alexithymia Scale and the Bermond-Vorst Alexithymia Questionnaire. Eur J Pers 1999;13:511-32. Berthoz S, Ouhayoun B, Perez-Diaz F, Consoli SM, Jouvent R. Comparison of the psychometric properties of two self-report questionnaires measuring alexithymia: confirmatory factor analysis of the 20-item Toronto Alexithymia Scale and the Bermond-Vorst Questionnaire. Eur Rev App Psychol 2000;50:359-68. Deborde AS, Berthoz S, Wallier JM, Fermanian J, Falissard B, Jeammet P, et al. The Bermond-Vorst Alexithymia Questionnaire cutoff scores: a study in eating-disordered and control subjects. Psychopathology 2007;41:43-9. Collet L, Cottraux J. The shortened Beck depression inventory (13 items). Study of the concurrent validity with the Hamilton scale and Widlöcher's retardation scale. Encephale 1986;12:77-9. Bruchon Schweitzer M, Paulhan I. Manuel de l'Inventaire d'Anxiété Etat-Trait, Forme Y (STAI-Y). Paris: Editions du Centre de Psychologie Appliquée; 1990. Funder DC. Personality judgment: a realistic approach to person perception. San Diego: Academic Press; 1999. Krystal H, Raskin H. Drug dependence, aspect of ego function. Detroit: Wayne State University Press; 1970. Rey JM, Sawyer MG, Raphael B, Patton GC, Lynskey M. Mental health of teenagers who use cannabis. Br J Psychiatry 2002;180:216-21. Dorard G, Berthoz S, Phan O, Corcos M, Bungener C. Affect dysregulation in cannabis abusers: a study in adolescents and young adults. Eur Child Adolesc Psychiatry (in press). Deborde AS, Berthoz S, Perdereau F, Godart N, Corcos M, Jeammet P. Validity of the BVAQ: a study in eating disorder patients and controls. Encephale 2004;30:464-73. Guilbaud O, Corcos M, Loas G, Speranza M, Dugré-Le Bigre C, Jeammet P. Le concept d'alexithymie: approches compréhensives et données épidémiologiques. In: Corcos M, Flament M, Jeammet P, editors. Les conduites de dépendance. Paris: Masson; 2002. p. 160-72. Joukamaa M, Taanila A, Miettunen J, Karvonen JT, Koskinen M, Veijola J. Epidemiology of alexithymia among adolescents. J Psychosom Res 2007;63:373-6. Riggs ML, Haviland MG, Berthoz S. The relationship between core alexithymic deficits and their expressions. Poster session presented at the meeting of the Association for Psychological Science, Washington, DC, May; 2007.