PROJECT SUMMARY Patients with advanced cancers experience distressing and debilitating symptoms that are often undertreated. Moreover, patients and caregivers are frequently expected to manage complex side effects at home, without ready access to real-time support, and use the emergency department to manage poorly-controlled symptoms, braving long wait times for treatment that might be more comfortably delivered in the outpatient setting. Although timely detection and management of symptoms can improve patients' quality of life, functional status, health care use, and disease-related outcomes, remote symptom monitoring is rarely used within oncology. The long-term term goal of this project is to improve advanced cancer patients' symptom burden, physical functioning, and survival, while reducing burdensome ED visits by using smartphones to detect acute clinical deteriorations in cancer patients without requiring their active involvement. Smartphones are an ideal scalable technology for long-term, remote monitoring because they can objectively measure physical activity (e.g. daily activity, time spent in house), behavior (e.g., sleep), and social and cognitive functioning using passive data collected from phone sensors and logs. In this project, we will identify a set of clinically significant patient symptoms and outcomes that are associated with continuously collected passive smartphone data using an existing cohort of patients with gynecologic cancers who are receiving palliative chemotherapy (Aim 1). We will develop and validate a method to identify clinically meaningful anomalies in cancer patients' behaviors (Aim 2). Finally, we will determine the feasibility and acceptability of using our smartphone-based real-time behavioral anomaly detection system as a trigger for timely clinical intervention in a new cohort of 20 gynecologic cancer patients and their cancer care providers (Aim 3). This project is responsive to PA-18-163 “Use of Technology to Enhance Patient Outcomes.” Results from this research will inform the design and conduct of a future randomized controlled trial to test the effectiveness of remote real-time behavioral anomaly detection in improving cancer patients' symptom burden (primary outcome) and physical functioning, health care use, and survival (secondary outcomes). If successful, this tool has the potential to improve outcomes in diverse patient populations, disease conditions, and environments.