Psychometric properties of the Bipolar Spectrum Diagnostic Scale in depressed youth

Background: Youth with depression may be at a higher risk of developing bipolar disorder (BD). Self-reported, dimensional measures, like the Bipolar Spectrum Diagnostic Scale (BSDS), aim to assess for BD in these groups. We explored properties of this instrument within a cohort of depressed, help-seeking youth.

Methods: We used baseline data from two randomized controlled trials for depressed youth (aged 15-25 years) who had no history of BD or psychosis and who completed the BSDS (n = 240; mean Age = 19.9 years, SD = 2.7; Female = 57 %). Structured diagnostic assessments were repeated at 26-weeks to detect new-onset BD. We examined false-positive rates for concurrent BD using established thresholds on the BSDS, utilised factor analyses to determine its underlying structure, and explored associations between the BSDS and demographic, clinical, and personality variables using linear regressions.

Results: False-positives rates were high. Most (60 %) participants scored above BSDS thresholds at baseline, though none developed BD over the 26-week study period. A three-factor model best fit BSDS items, representing depression, mania, and lability factors. BSDS total, mania and lability sub-scale scores were associated with similar characteristics, with a different pattern of association for the BSDS depression subscale.

Limitations: With no long-term follow-up of the current sample nor a separate bipolar youth sample, we were unable to determine the overall discriminant validity, sensitivity, or longer-term predictive validity of the BSDS.

Conclusion: At recommended thresholds, BSDS has high false positive rates for detecting current BD in youth with moderate to severe depression, especially with mental state or personality disorder comorbidities.