OBJECTIVE: To quantify the extent of 'avoidable mortality' in those with schizophrenia or bipolar disorder and to quantify the effect a reduction in these might have on the mortality gap associated with severe mental illness. METHOD: A cohort was studied of people aged