A review of self-rating instruments to identify young people at risk of bipolar spectrum disorders

Background: Despite 15-25 years being the peak age at onset for bipolar disorder (BD), recognition and introduction of appropriate treatment are typically delayed by 601 more years. A major reason for these delays is the failure to identify prior mood episodes or risk factors for bipolarity, suggesting that there may be a particular benefit in using screening tools in this age group. Methods: We identified papers published between 1980 and 2012 that assessed the validity of measures specifically used to screen for BD, risk factors for bipolarity, or a prior history of depressive, hypomanic or manic episodes. From these, we identified those studies that included sufficient information about participants aged 15-25 years. Results: Searches yielded only eleven independent studies with available data on the target age group. The studies employed seven separate scales that demonstrated quite different screening properties. The cut-off scores for optimal sensitivity and specificity often differed from those reported for adult samples and varied according to sample characteristics. Limitations: Our findings may be undermined by the small sample sizes and low number of individuals in the specified age range included in the studies identified. Although we explored all available methods for detecting studies and data sets, we cannot exclude the likelihood that other relevant studies on this age range exist. Conclusions: Available screening tools for BD have mainly been validated in middle-aged adult samples with established mood disorders, and the instruments reviewed show sub-optimal screening properties when applied to adolescents and young adults. Overall, those measures that examine personality traits or temperament appear to perform better than those assessing discrete symptoms or episodes. However, all the measures need further testing and/or modification. In the interim, to improve screening for BD in those in the peak age range for onset of this disorder, it may be appropriate to combine a measure with high sensitivity with one that demonstrates high specificity for a given sampling frame. (C) 2013 Elsevier B.V. All rights reserved.