[The alexithymia questionnaire for children (AQC): Psychometric properties in a sample of 105 adolescents hospitalized in an adolescent psychiatric department].

OBJECTIVE: The 20-item Toronto Alexithymia Scale (TAS-20) is the most widely used measure of alexithymia in non-clinical or clinical populations. The TAS-20 evaluates three dimensions of the alexithymia construct: the difficulty identifying feelings (DIF), the difficulty describing feelings (DDF) and externally oriented thinking (EOT). The TAS-20 is also used in adolescents or children, and the psychometric properties of the scale have not been systematically evaluated in these populations. Recently several studies have shown systematic age differences in the factor structure and a decrease of the quality of the measurement with age. Notably, low reliability measured by the Cronbach α coefficient has been found for the EOT factor. Taking into account the limitations of the TAS-20 in pre-adult populations the Alexithymia Questionnaire for Children (AQC), an adaptation of the TAS-20, has been proposed by a reformulation of the TAS-20 items (Rieffe et al., 2006). Two studies in healthy children found satisfactory psychometric properties with the three-factor structure demonstrating adequate parameters in the confirmatory factorial analyses (CFA). In the two studies low reliabilities of the EOT factor were reported, and recent studies in adolescents using the TAS-20 found that a two-factor model (DDF, DIF) had a better fit than the original three-factor model. Thus, the aim of the present study was firstly to verify the psychometric properties of the AQC in a sample of adolescents presenting various psychiatric disorders and secondly to test the adequacy of the bi- or tridimensional model of the scale.

METHOD: One hundred and five adolescents (27 males, 78 females) with a mean age of 15.06 years (sd=1.55, range: 12-18 years) were hospitalized in the adolescent psychiatric department of the Erasme Hospital (Bruxelles, Belgium) for various psychiatric disorders. The main diagnoses were: adjustment disorder (n=56, 53.5%), mood or anxiety disorders (n=17, 16.2%), impulse control disorder (n=11, 10.5%). The subjects completed the French version of the AQC. CFA were done testing the adequacy of the three or two-factor structure of the scale. Two indices were considered: the normed χ2 (ratio of χ2 and degrees of freedom) and the root mean square error of approximation (RMSEA). The values for acceptable fit were normed χ2 <3 and RMSEA<0.05. Cronbach α were also calculated.

RESULTS: Fit indices for the three-factor model were respectively 1.165 and 0.0303 for the normed χ2 and RMSEA. For the two-factor model, the values were respectively 0.819 and 0. Thus, these two well-fitting models were compared using the χ2 -difference test, which indicated a significantly better fit for the two-factor model over the three-factor model (χ2 -difference=151.447, delta df=114, P<0.05). The values of the Cronbach α coefficients were respectively 0.72, 0.75, 0.18 for DIF, DDF and EOT subscales. Moreover, the values of the Cronbach α coefficients were respectively 0.71 and 0.83 for the full scale and the scale without the EOT items.

CONCLUSION: The Alexithymia Questionnaire for Children had satisfactory psychometric properties found in a sample of adolescent psychiatric inpatients. Taking into account firstly the superiority of the two-factor model over the classical three-factor model and secondly the low value of the Cronbach α for the EOT factor, it is proposed to use only the twelve-item scale by excluding the EOT items.