Multiple predictive model for clinical and subclinical depression in adolescence

Abstract

Abstract

Introduction: Identification of potential groups of differences and similarities between clinical and subclinical depression in adolescence, compared to the risk factors of cognitive vulnerability and psychosocial risk factors, contributes to shedding light on the etiological picture of clinical and subclinical depression.

Objectives: The present study investigated whether we can build multiple predictive model for clinical and subclinical depression in adolescence, based on risk factors of cognitive vulnerability and psychosocial risk factors.

Material and Methods: Multiple predictive model for group membership in three groups: clinical, subclinical and control group, has been built by using canonical discriminant analysis. The final survey sample consists of 412 adolescents (61.7% female and 38.3% male) aged 13-17 years. The data were collected from Psychiatric words and Centers for Mental Health for children and adolescents from three Clinics in Macedonia (for clinical sample) and schools (for subclinical and control sample) during last four years, including one year of pilot study for psychological instruments. Cognitive vulnerability factors for depression (dysfunctional attitudes, negative inferential style, ruminative response style) and psychosocial risk factors (negative life events and perceived social support) were measured by a set of psychological instruments.

Results: The discriminant analysis showed that it was possible to statistically significantly distinguish groups on the basis of the mentioned predictors and it allocated two discriminant functions. The overall test of the first discriminant function was significant (p < .001) indicating that scores discriminated between the three groups accounting for 99.3% of the total variance of depression. The first discriminant function consists of all listed predictors, ruminative response style, dysfunctional attitudes, negative inferential style, and negative life events in a positive direction and perceived social support in a negative direction, and in this function the highest scores are for the clinical and the lowest for the control subjects. Subclinical respondents were in-between. The overall test of the second discriminant function was significant (p < .001) indicating that scores discriminated between the three groups accounting for 0.7% of the total variance of depression. Only ruminative response style makes the second discriminant function and the subclinical sample has the highest scores, while the clinical sample has the lowest. The control group is in-between.

Conclusion: We have confirmed our expectations that we can build multi predictive model of clinical and subclinical depression in adolescence. The clinical, subclinical group and control group differ significantly with respect to the factors of cognitive vulnerability and psychosocial risk factors. We believe that this research has scientific and applicative value and that it has not only theoretical and empirical contributions, but has contributions to clinical practice as well.

Key words: adolescence, clinical depression, predictive model, risk factors, subclinical depression.