Detecting And Capitalizing On Physiological Dimensions Of Psychiatric Illness

Serious mental illnesses, including bipolar disorders (BD), account for a large share of the worldwide healthcare burden-estimated at $62.7B in the U.S. alone. Bipolar disorders represent a family of common, lifelong illnesses associated with poor functional and clinical outcomes, high suicide rates, and huge societal costs. Interpersonal and Social Rhythm Therapy (IPSRT), a validated treatment for BD, helps patients lead lives characterized by greater stability of daily rhythms, using a 5 item paper-and-pencil self-monitoring instrument called the Social Rhythm Metric (SRM). IPSRT has been shown to improve patient outcomes, yet many patients struggle to monitor their daily routine or even access the treatment. In this paper we describe how biological characteristics of bipolar disorder can be taken into consideration when developing systems to detect and stabilize mood episodes. We describe the co-design of MoodRhythm, a smartphone and web app, with patients and therapists. It is designed to support patients in tracking their health passively and actively over a long period of time. MoodRhythm uses the phone's onboard sensors to automatically track sleep and social activity patterns. We report results of a small clinical pilot with experienced IPSRT clinicians and patients with bipolar disorder and finish by describing the role physiological computing could have not just in monitoring psychiatric illnesses according to existing broad categories of diagnosis but in helping radically tailor diagnoses to each individual patient and develop interventions that take advantage of idiosyncratic characteristics of each person's illness in order to increase patient engagement in and adherence to treatment.