Abstract
Internalizing disorders have a high prevalence in the population, and represent a major cost for the society and severe suffering for the individuals affected if not diagnosed and treated correctly. Traditionally, diagnostic interviews have been the main methodology for screening and diagnostic assessment. Different types of questionnaires intended to screen for symptoms of mental distress have also been developed to save time and resources, and it is necessary to have knowledge regarding these instruments' correspondence to psychiatric diagnoses. A host of studies have investigated the phenotypic correspondence between questionnaires and interview-based diagnoses. Typically, a moderate correspondence is found, which implies that the questionnaires do a mediocre job in detecting psychiatric cases. Twin modelling renders it possible to investigate the correspondence at a genotypic level, and hence is able to reveal more information about the correspondence than phenotypic studies. By decomposing the variance inherent in the instruments, twin modelling can delineate how much of the phenotypic correspondence that is due to genetic and environmental sources. In addition, it is possible to investigate the genetic and environmental correlation between the variables. To date, only one study has investigated the correspondence by the use of twin modelling (Foley, Neale & Kendler, 2001).
The present study makes use of both phenotypic analyses and twin modelling in order to investigate the correspondence between the Symptom Checklist-5 (SCL-5), and six internalizing disorders from the Computerized International Diagnostic Interview (CIDI). A sample of 7992 twins from the Norwegian Institute of Public Health Twin Panel (NIPHTP) replied to the questionnaire, and 2793 twins were interviewed. By including data from both males and females, and extending the analyses to including several internalizing disorders, the study represents an extensive replication of Foley, Neale and Kendler (2001).
The phenotypic correspondence found was moderate, with a tetrachoric correlation of .48, odds ratio of 4.75, and a relative risk of 2.95. The phenotypic results are in the area of what have been found earlier. A confirmatory factor analysis justified collapsing of the six internalizing disorders into one factor. The twin modelling revealed that 81% of the phenotypic correspondence was due to genetic factors, and 19% was due to environmental factors. The main result was the finding of a strong genetic correlation of .82 (.61 - 1.0, 95% CI) between the genetic component in SCL-5 and internalizing disorder, which suggests that the genetic liability indexed by these instruments are practically the same. The environmental correlation was found to be .16 (.00 - .34, 95% CI), which implies mainly non-shared environmental factors between the SCL-5 and internalizing disorder.
The findings from the twin modelling cast light upon the findings from phenotypic studies and suggests that the results of Foley et al. may be generalized from the single disorder depression to internalizing disorder in general. In addition, the present study finds that the results hold for both males and females. The strong genetic correspondence in the instruments has implications for research and clinical utility of questionnaires measuring symptoms of anxiety and depression. In essence, the SCL-5 and CIDI tap onto the same genetic factor, and hence the finding implies that the SCL-5 may be used as a screener for genetic risk for internalizing disorder.