Whilst there are now controlled studies supporting the effectiveness of cog
nitive behaviour therapy in psychosis, the effective components of the ther
apy are unclear. It is possible that certain core factors may have a centra
l role in effectiveness. Integrative models of therapy typically consist of
such factors in the form of a list of clinical strategies. A similar forma
t is used here to develop an empirically derived, integrative approach to t
herapy with clients who have been diagnosed as psychotic. The approach is i
llustrated by three case studies of individuals with a long history of schi
zophrenia and poor response to psychotropic medication. These also provide
a preliminary evaluation. Using the SCL-90-R as the outcome measure, the cl
ients each reported a symptom reduction of around 50% after 50-80 sessions
of therapy, which was maintained at 10 month follow-up. Factor scores of th
e SCL-90-R are used to examine differences in the change processes of the t
hree people. The generalizability and limitations of this study are discuss
ed. Copyright (C) 1999 John Wiley & Sans, Ltd.