[To analyse the transferability of an intervention : applying the functions/implementation/context model on a diabetes prevention programme].

INTRODUCTION: A type 2 diabetes primary prevention programme that was shown to be feasible and effective in 2003 was transferred to another vulnerable district of Reunion Island in 2004, but its short-term effectiveness could not be reproduced. Based on this example, this article analyses whether the key functions/implementation/context model can be useful to : 1- more accurately describe an evaluated intervention ; and 2- identify the factors involved in the transferability of this intervention.

METHODS: The causality model of the initial programme is described. We then identified the key functions (or theoretical processes) of this initial programme, implementation of these functions, and the context in which the intervention took place. Transfer was analysed by focusing on the differences between the initial programme and the transferred programme in terms of key functions, implementation and context.

RESULTS: The causality model involves individual, meso-social, and environmental health determinants. Our analysis of programme transfer highlights differences in : key functions (two key functions dropped, one key function added, one key function modified), implementation (failure to implement, loss/improvement of quality), and context (population and socioeconomic level of the district concerned).

CONCLUSION: This work supports the hypothesis that the key functions/implementation/context model can be useful to improve the description of an intervention, and analyse the factors involved in its transferability.