Endogenous depression is known to be associated with good outcome following electroconvulsive therapy (ECT). In a double-blind, prospective study, we applied three clinical predictive indices and one diagnostic index to a cohort of 29 endogenous depressed patients, to obtain better predictors of outcome following ECT. The Newcastle Prognostic Index identified ECT responders with high specificity but low sensitivity; other indices, such as those described by Hobson (1953) and by Mendels (1967), were neither sensitive nor specific in predictive standards. If ECT-treated depressed patients are preselected for endogenous symptomatology, fresh clinical predictive indices need to be developed.