Identifying and differentiating melancholic depression in a non-clinical sample

Differentiating melancholic and non-melancholic depressive disorders and evaluating whether they differ categorically or dimensionally has had a lengthy history, but has not previously been evaluated in a non-clinical adolescent sample.We studied a sample of 1579 senior high school students and evaluated the capacity of the Sydney Melancholia Prototype Index (SMPI) to differentiate melancholic from non-melancholic depression, both using a 'top down' strategy of imposing a pre-established cut-off score and a 'bottom up' strategy of employing latent class analyses.The two strategies respectively generated prevalence figures of 3.4% and 8.1% of the students having experienced a melancholic depressive episode and with the difference reflecting the LCA assigning some students who did not reach the pre-established cut-off score for the SMPI in the putative melancholic class. The principal latent class analysis failed to generate pristine melancholic and non-melancholic depressive classes, in that it also generated an 'intermediate' as well as a non-clinical depressive class. Both SMPI strategies identified similar symptoms-such as anhedonia and anergia-and several illness correlates that best differentiated those assigned melancholia status, and both strategies confirmed melancholia assignment being associated with factors indicative of more severe depressive disorders and of likely melancholic depression.Data were assessed by self-report only, only lifetime depression was assessed, and no other depressive diagnostic validating measure was administered.The SMPI appears capable of identifying and differentiating melancholic from non-melancholic depression in a non-clinical adolescent sample.