Facial expression recognition (FER) is a technology designed to track and classify facial expressive behaviours and produce some form of meaningful knowledge about people based on those behaviours. This report investigates the impact of FER systems in healthcare contexts. Here, these tools are designed to detect and assess mental health conditions (depression and anxiety), neurodevelopmental disorders (autism, bipolar disorders, and attention deficit hyperactivity disorder or ADHD), and assess pain levels on the basis of facial expressions. Many have pointed to the limitations of these systems, as well as the privacy risks associated with their use in high-stakes contexts. This report highlights the broader societal impact of these technologies and offers recommendations to identify and address harm.