Predictive and discriminant validity of different psychopathology and temperament scales for major psychiatric disorders – 23-year follow-up of the Northern Finland Birth Cohort 1966

Abstract

Introduction

Several psychological and psychiatric instruments have been developed to recognize or predict different psychiatric disorders.

Objectives

We studied the predictive, and discriminant validity of different psychopathology scales and temperament traits for subsequent psychiatric diagnoses due to schizophrenia, bipolar and depressive disorders in a 23-year follow-up.

Methods

Temperament traits, perceptual aberration, physical and social anhedonia, depression and anxiety subscales of Symptom Checklist (SCL-D and SCL-A), Hypomanic Personality Scale (HPS), Schizoidia Scale, and Bipolar II Scale were completed as part of the 31-year follow-up survey of the prospective Northern Finland 1966 Birth Cohort (n = 5006). New onset psychiatric diagnoses were followed until age of 54 years using different nationwide registers.

Results

In the follow-up 28 (0.6%) individuals had diagnosis of schizophrenia, 40 (0.8%) bipolar and 405 (8.1%) depressive disorders. Several of the included scales associated statistically significantly with subsequent diagnoses. High SCL-A and SCL-D scores were strong predictors (Cohen’s d’s between 0.76 and 1.08) for schizophrenia and depressive disorders, whereas high HPS score was best predictor (d=0.67) for bipolar disorders. When comparing patient groups, schizophrenia group had low scores in reward dependence when compared with both bipolar (d=-0.80) and depressive (d=-0.66) disorders. Harm avoidance was the best trait to discriminate depressive and bipolar disorders, with higher scores in depressive disorders (d=0.48).

Conclusions

Interestingly we found that differed psychopathology scales were strong but non-specific predictors for these psychiatric disorders, whereas temperament traits were useful predictors regarding discriminating these disorders. The presented scales can be used in population samples when predicting psychiatric illnesses.

Disclosure

No significant relationships.