Alexithymia and psychiatric symptoms in a population of nursery workers: A study using the 20-item Toronto alexithymia scale

Social Behavior and Personality, 2001 by Posse, M, Backenroth-Ohsako, G, Hakanson, C-E, Hallstrom, T

Alexithymia was assessed with the 20-item Toronto Alexithymia Scale (TAS20), (Bagby, Parker & Taylor, 1994). The TAS-20 has three factors that assess three salient components of the multifaceted alexithymia construct: difficulty in identifying feelings, difficulty in communicating feelings to others, and externally oriented thinking. In addition to demonstrating stability and replicability of its threefactor structure, the TAS-20 has shown high internal consistency, good test-retest reliability, and convergent, discriminant and concurrent validity, (Bagby, Parker & Taylor 1994; Taylor, Bagby & Parker, 1997).

Consistent with the definition of the alexithymia construct, the TAS-20 also correlates negatively with a measure of fantasy and other imaginative activity, (Taylor, Bagby & Parker 1997). The TAS-20 used in this study has been translated into Swedish by Gillberg, Wahlstrom and Kossler with the permission of the Toronto group (RAstam, Gillberg, Gillberg & Johansson, 1997). It is a self-report questionnaire containing 20 items that are rated on a 5-point scale, yielding total possible scores ranging from 20 to 100. On the basis of previous studies the score that maximises the diagnostic validity of the scale has been set to a cut-off point of 56, (Bagby, Taylor & Parker, 1994; Loas, et al., 1996). In this study the cut-off point of =/> 56 was set as corresponding to "clinical alexithymia". In a previous study using the same translated version the cut-off point for abnormality was set on the basis of distribution of scores in the control group (the 5% of highest scores), corresponding to =/> 54 (Rastam, Gillberg, Gillberg & Johansson, 1997). As the present authors had no control group they followed the recommendation for a cut off at =/> 56 as indication of abnormality.

STATISTICAL METHODS

In the comparison between the alexithymic group and the nonalexithymic one the authors used the non-parametric Wilcoxon Rank Sum test (Kruskal-Wallis) see Gonover (1980).

To be able to study the associations between the symptom-dimensions and TAS20 global as well as the TAS-20 subfactors the authors used Pearson correlation. They also looked for modes of explanation for TAS-20 global and the TAS-- subfactors emanating from the symptom variables by applying multiple regression analysis.

RESULTS

All subjects were women (N= 1032) and 864 subjects delivered complete tests (83.6%). The mean age of the subjects was 42.4 years, age was normally distributed. Age was negatively related to lower feelings of well-being, (p

The mean score for well-being was in the medium rank (2.4 with min 1.3 and max 3.7), otherwise all the rest of the dimensions showed low average scores. Scores on depressive symptoms, social dysfunction, vegetative symptoms of anxiety and gastro-intestinal symptoms were all found to be positively interrelated and had a negative relationship with feelings of well-being.

There was a prevalence of alexithymia of 7.9% found in this all-female population. The total TAS-20 score was highly associated to a lower level of feelings of well-being, more symptoms of somatic anxiety, depressive symptoms and gastrointestinal symptoms as well as a lower level of social functioning. Alexithymic and non-alexithymic subjects differed significantly on all items assessing social disability apart from days off work (sick leave) due to mental health problems, and non-performance of daily activities (Table 1). Although alexithymic subjects showed more somatic anxiety (p

The scores on the three factors of TAS-20 were further investigated and it was found that factor 1, difficulty identifying feelings, was positively significantly related to higher age (p

In an attempt to find explanatory models for TAS and its subfactors as functions of the symptom-dimensions, multiple regression was performed. It was found that the explanatory levels were fairly low. R-square was 0.14 for the global TAS (p

DISCUSSION

The prevalence of alexithymia found in this study is somewhat higher than the 6.7 % found in a group of healthy women matched by age and education to a patient group with anorexia nervosa (Bourke, Taylor, Parker & Bagby, 1992) but below the level (10%) as found in a group of women representing a general population in Finland (Salminen, Saarijarvi, Arela, Toikka & Kauhanen, 1999). This could probably, to some extent, be explained by the higher mean age in the present sample and the sample in the Finnish study compared with the college population, as alexithymia has been found to increase with advancing age in other studies as well (Pasini, Delle Chiaie, Seripa & Ciani, 1992; Posse & Hallstrom, 1999). It is considerably lower than the 38.3% found in a female psychiatric out-patient population (Parker, Taylor & Bagby, 1989). This result is not surprising, because the present sample contained only female subjects of a "healthy" population who have chosen to work with children where especially high demands on ability for openness, empathy and fantasising are required.