As college counseling centers struggle to meet the growing demands of behavioral health services, digital mental health tools like smartphone apps offer a scalable solution to increase access to care. However, clinicians report greater time demands and uncertainty over how to act upon digital data. In this paper, by using established statistical techniques, we condense complex smartphone data into results that are quickly understood and clinically meaningful. Specifically, we show how smartphone digital phenotyping data collected by college students can be used to predict an individual’s anxiety and depression level on a daily or weekly basis with an error of less than 10%. These predictions are then condensed into a 1 to 5 scale with a 1 representing patients with the lowest risk of presenting high anxiety or depression, and a 5 representing the patients with the highest risk. If used in a clinical setting, these risk scores have the potential to help college counseling centers monitor symptom severity in real-time via students’ own smartphones, allocate resources more efficiently, and ensure that students are receiving the appropriate level of treatment.