Digital phenotyping for monitoring mental health outcomes in patients with cancer: a scoping review

Patients with cancer often experience anxiety, depression, psychological distress, and fear throughout diagnosis, treatment, and rehabilitation. Conventional mental health assessment relies mainly on periodic clinical evaluations and patient-reported measures, which may increase patient burden and inadequately capture dynamic psychological changes in daily life. Digital phenotyping enables the passive and continuous collection of behavioral and physiological data through smartphones, wearable devices, and other sensors; however, evidence for its application to mental health in patients with cancer remains limited and fragmented. PubMed, Web of Science, CINAHL, Cochrane Library, and Embase were searched from inception to July 2026 using terms related to digital phenotyping and cancer. A supplementary targeted sensitivity search was also conducted. Two reviewers independently screened the records and extracted data on study characteristics, participant characteristics, data collection characteristics, mental health outcomes, assessment tools, and associations found. A total of 17 studies involving 829 participants were included. Wearable devices and smartphones were used for passive data collection, most commonly incorporating accelerometers and GPS. Derived features covered activity, location and mobility, sleep, cardiac physiology, and device-use behavior. Depression and anxiety were the most frequently assessed mental health outcomes. Higher physical activity was generally associated with lower depression, anxiety, sadness, and psychological distress, whereas greater sedentary behavior and disrupted sleep or rest–activity rhythms were associated with higher psychological symptom levels. Digital phenotyping shows potential to support remote mental health monitoring in patients with cancer. However, the current evidence remains limited and heterogeneous and is insufficient to support the routine clinical implementation of digital phenotyping. Future research should prioritize larger-scale, standardized, multicenter prospective studies.