TY - JOUR T1 - Implementation science at the crossroads JF - BMJ Quality & Safety JO - BMJ Qual Saf DO - 10.1136/bmjqs-2017-007502 SP - bmjqs-2017-007502 AU - Richard J Lilford Y1 - 2017/11/28 UR - http://qualitysafety.bmj.com/content/early/2017/11/28/bmjqs-2017-007502.abstract N2 - Martin Marshall and colleagues1 take themselves to task for the suboptimal design of a complex (multicomponent) intervention to improve safety of services for people in care homes. The authors make much of the complexity of the intervention—service interventions are ‘not like a pill’. Interventions must be adapted—when first promulgated the intervention in question had nine components, and this inflated to 15 over the course of the project. Contrast all of these with a financial incentive promulgated by the Specialist Services Commissioning authority for the West Midlands, England. Hospitals were simply given a financial incentive to promote a switch from facility to home haemodialysis.2So here we have accounts of what appear to be two very different types of interventions; Marshall’s intervention encapsulates 15 components, while the commissioning agent’s intervention was of one component only. One might think that Marshall’s intervention was complex and the commissioning agent’s was simple. But this is an artefact of how the intervention is conceptualised. The commissioning agent’s incentive could not, by itself, achieve any change. It was necessary for each service to respond to it, to make and then implement a plan. These are described in the paper evaluating the incentive. … ER -