Confiabilidad y validez del WAI-SF en adolescentes argentinos

Abstract

La alianza terapéutica (AT) es un concepto central tanto en la investigación como en la práctica en psicoterapia. Existen diversos instrumentos que evalúan este constructo pero la mayoría de ellos han sido desarrollados para adultos. En Argentina, no existen instrumentos para estudiar la AT con adolescentes. Este estudio tuvo como objetivo evaluar las propiedades psicométricas, en términos de confiabilidad y validez del Inventario de Alianza de Trabajo versión corta para pacientes (WAI-SF) en adolescentes de Argentina. La muestra incluyó a 206 pacientes argentinos (entre 12 y 20 años) que se encontraban realizando psicoterapia. Para estudiar la fiabilidad del instrumento se analizaron medidas de consistencia interna (alfa de Cronbach y ordinal) y de homogeneidad de los ítems (correlaciones ítem-total ajustadas). Para el estudio de evidencias de validez del instrumento se analizó la validez de constructo (análisis factorial confirmatorio) y la validez concurrente (correlaciones con el Alliance Negotiation Scale). Se encontraron evidencias de una adecuada consistencia interna (alfa de Cronbach= 0.92) y homogeneidad de los ítems (todos los ítems con correlaciones entre 0.30 y 0.85), así como de validez de constructo (CFI = 0.97; SRMR = 0.03 y RMSEA = 0.08 )y validez concurrente (r = 0.73). El análisis factorial mostró un buen ajuste del modelo original de tres subescalas. Este desarrollo pretende contribuir a los psicoterapeutas, a los pacientes y a los investigadores en Argentina, ofreciendo un instrumento válido y confiable para el análisis de la alianza terapéutica en adolescentes.

Therapeutic alliance (TA) represents a central concept for research and clinical practice. Bordin's definition has historically been one of the most widely accepted when understanding the construct and designing assessment instruments. Despite the importance of TA in the psychotherapeutic process, most studies on this topic have been conducted in the adult population. This lack of information implies a significant challenge when interpreting the construct in clinical practice in other age groups such as adolescence. Several studies suggest the importance of studying this construct in adolescence. Adolescents tend to attend psychotherapy unaware of their problems, in conflict with their parents regarding their motivation to consult, and resistant to change because they do not recognize their problems. The lack of a unified approach for measuring TA in young people represents a methodological problem for several reasons. Instruments for measuring TA in adolescents may lack a consistent basis due to a lack of knowledge. It is necessary to understand what aspects of the therapeutic relationship each scale measures and how these factors relate to one another. Internationally, the most widely used instrument is the Working Alliance Inventory (WAI), based on Bordin's tripartite model Bond-Task-Goals. In the original study, the correlation between the subscales showed to be high but factor analysis methods were not used in the development of the subscales. Subsequent studies found high correlations among the three WAI dimensions. This has led many researchers to use the mean of the total WAI score. Few studies have evaluated the factor structure of different versions of the WAI with young people and none of them supported the hypothesized Bond-Task-Goals alliance model. This study aimed to evaluate the reliability and validity of the Working Alliance Inventory short version (WAI-SF) in adolescents from Argentina. Therapists from different psychotherapy centers working with adolescents were invited to take part in the study. Once they accepted, they shared an online link with their patients who completed the WAI-SF anonymously. The sample of this study was non-probabilistic and included 206 Argentinean patients undergoing psychotherapy. Of the 206 participants 79.6 % indicated to be female and 17.9 % indicated to be male. The age ranged from 12 to 20 years. Of the participants, 65.9 % resided in the Buenos Aires Metropolitan Area, 23.2 % resided in the Province of Buenos Aires, and the remaining 10.9 % lived in other provinces of Argentina. Most of the participants in the sample were in advanced treatment. When asked about the number of sessions, the adolescents could not indicate the number. The minimum number of sessions was 5. To study the reliability of the instrument, measures of internal consistency (Cronbach's alpha) and item homogeneity (adjusted item-total correlations) were analyzed. For studying the instrument's validity, construct validity (confirmatory factor analysis) and concurrent validity (correlations with the Alliance Negotiation Scale) were analyzed. Construct validity was studied using a bifactor model with a general alliance factor at one level, and the three subfactors of Bonding, Tasks, and Goals. All correlations between latent variables were constrained to zero. We found evidence of adequate internal consistency (Cronbach’s alpha = .92) and item homogeneity (all correlations between .30 - .85), as well as construct validity (CFI = .97; SRMR = .03 y RMSEA = .08) and concurrent validity (r = .73). The factor analysis showed a good fit for the three subscales model. This development aims to contribute to psychotherapists, patients, and researchers in Argentina by offering a valid and reliable instrument for the analysis of the therapeutic alliance in adolescents. In the future, it would be relevant to analyze the properties of the WAI-SF in Argentina, using samples collected with different strategies to observe whether the sampling method influences its psychometric properties.