Not all sensors are created equal: a framework for evaluating human performance measurement technologies

Recent years have witnessed an explosion in the number of wearable sensing devices and associated apps that target a wide range of biomedical metrics, from actigraphy to glucose monitoring to lung function. This offers big opportunities for achieving scale in the use of such devices in application contexts such as telehealth, human performance and behaviour research and digitally enabled clinical trials. However, this increased availability and choice of sensors also brings with it a great challenge in optimising the match between the sensor and a specific application context. There is a need for a structured approach to first refining the requirements for a specific application, and then evaluating the available devices against those requirements. In this paper we will outline the main features of such an evaluation framework that has been developed with input from stakeholders in academic, clinical and industry settings.