Biases in Digital Health Measures

The field of digital measures, in particular digital biomarkers, has gone through remarkable progress in recent years. This is due to the ability of digital measures to provide frequent, objective, and convenient measures of disease symptoms and to the emergence of new frameworks for development, evaluation, and qualification of such measures. Nevertheless, biases that can be inadvertently incorporated during several stages of their development, evaluation, and deployment represent a threat to their effectiveness. In this work, we provide an overview of the young field of digital measures with a particular emphasis in neurological and mental health disorders. We include conceptual definitions and historical milestones. Furthermore, we describe the potential sex and gender biases that can arise during key steps of their development and provide recommendations to address them.