This is the first site level economic evaluation of the Improving Access to Psychological Therapies programme for severe mental illness (IAPT-SMI) that is funded by NHS England. It also aims to illustrate the challenges involved in evaluations based on routine data with low internal validity. Six IAPT-SMI pilot sites treated 1 of 2 clinical groups: (i) psychosis or bipolar disorder; (ii) personality disorder. A decision analytical model nested in a before-after framework- the same patients 12 months after treatment versus 12 months before treatment-was used to compare the cost-effectiveness of IAPT-SMI with treatment as usual (TAU). IAPT-SMI appears to be more costly overall but save non-psychological treatment costs. There is evidence it may improve function and lower incidence of harmful behaviour. However, there is a need for evaluations with a more conventional study design that measure a more comprehensive array of resource use and clinical outcomes.
Keywords: Cost-effectiveness; Economic evaluation; IAPT; Improving access to psychological therapies; Severe mental illness.