The development and validation of an algorithm to predict future depression onset in unselected youth

Abstract

Abstract Background Universal depression screening in youth typically focuses on strategies for identifying current distress and impairment. However, these protocols also play a critical role in primary prevention initiatives that depend on correctly estimating future depression risk. Thus, the present study aimed to identify the best screening approach for predicting depression onset in youth. Methods Two multi-wave longitudinal studies (N = 591, AgeM = 11.74; N = 348, AgeM = 12.56) were used as the ‘test’ and ‘validation’ datasets among youth who did not present with a history of clinical depression. Youth and caregivers completed inventories for depressive symptoms, adversity exposure (including maternal depression), social/academic impairment, cognitive vulnerabilities (rumination, dysfunctional attitudes, and negative cognitive style), and emotional predispositions (negative and positive affect) at baseline. Subsequently, multi-informant diagnostic interviews were completed every 6 months for 2 years. Results Self-reported rumination, social/academic impairment, and negative affect best predicted first depression onsets in youth across both samples. Self- and parent-reported depressive symptoms did not consistently predict depression onset after controlling for other predictors. Youth with high scores on the three inventories were approximately twice as likely to experience a future first depressive episode compared to the sample average. Results suggested that one s likelihood of developing depression could be estimated based on subthreshold and threshold risk scores. Conclusions Most pediatric depression screening protocols assess current manifestations of depressive symptoms. Screening for prospective first onsets of depressive episodes can be better accomplished via an algorithm incorporating rumination, negative affect, and impairment.