4. Discussion
Our study had advantages of analysing data on a large number of
participants diagnosed with unipolar depression following independent
clinical and research interviews, in analysing data from both
development and validation sample sub-sets, and in analysing the
utility of a model that does not assume clinical features as having
equal diagnostic value. Instead it sought to identify those having the
highest signal capacities, a model that is consistent with much
clinical practice in medicine and in psychiatry. As noted, only
relatively few items were required to differentiate melancholic and
non-melancholic depression with a high level of precision.