Due to the iterative nature of medical device innovation and development, a single once-off assessment does not provide all the answers that decision-makers need over the device’s lifetime. Consequently, a lifecycle approach is frequently recommended. However, there is no lifecycle model recognized internationally for conducting such evaluations, nor is there explicit agreement regarding what is meant by, or evaluated over, the lifecycle. The purpose of this review was to identify and explore the range of models/frameworks used for evaluating medical devices across their lifetime – to determine what people mean by ‘the lifecycle’, what is evaluated, how, and why.

A qualitative evidence synthesis of lifecycle models described in the literature from a wide variety of disciplines was performed. Literature searching and selection of models iterated with analysis. Similarities, differences, and patterns were identified, from which themes became apparent, and explanatory theories were developed.

Fifty-three models are included in the synthesis. The dimensions of difference include, amongst others, the lifecycle scope, level of application, evaluation timepoints and methods, factors included in the models, and the focus of interest. These are each influenced by the purpose of the lifecycle evaluation, which depends on the perspective and the decision or activity the evaluation is intended to inform. Few models provide a lifecycle approach to evaluating safety or efficacy. Theories explaining the differences are postulated.

Lifecycle evaluation means different things to different actors, with varied reasons for evaluation and different variables included in the models. Thus, discussions between different actors on lifecycle evaluation may be inadvertently at cross-purposes. Without first defining what is meant by the lifecycle (including the stages or phases of activity it covers) and the variables included in an evaluation, care must be exercised when discussing a lifecycle evaluation approach – to ensure that the meaning (and intended objective) is not lost in translation. Indeed, promoting lifecycle evaluation may result in necessary evidence not being generated early enough, being deferred instead until later.