Evaluating the feasibility of complex interventions in mental health services : standardised measure and reporting guidelines.

Abstract
Background
The feasibility of implementation is insufficiently considered in clinical guideline development, leading to human and financial resource wastage.
Aims
To develop (a) an empirically based standardised measure of the feasibility of complex interventions for use within mental health services and (b) reporting guidelines to facilitate feasibility assessment.
Method
A focused narrative review of studies assessing implementation blocks and enablers was conducted with thematic analysis and vote counting used to determine candidate items for the measure. Twenty purposively sampled studies (15 trial reports, 5 protocols) were included in the psychometric evaluation, spanning different interventions types. Cohen’s kappa (κ) was calculated for interrater reliability and test-retest reliability.
Results
In total, 95 influences on implementation were identified from 299 references. The final measure - Structured Assessment of FEasibility (SAFE) - comprises 16 items rated on a Likert scale. There was excellent interrater (κ = 0.84, 95% CI 0.79-0.89) and test-retest reliability (κ = 0.89, 95% CI 0.85-0.93). Cost information and training time were the two influences least likely to be reported in intervention papers. The SAFE reporting guidelines include 16 items organised into three categories (intervention, resource consequences, evaluation).
Conclusions
A novel approach to evaluating interventions, SAFE, supplements efficacy and health economic evidence. The SAFE reporting guidelines will allow feasibility of an intervention to be systematically assessed.