The Potential of Wearable Devices in Cancer Care Delivery

Abstract

Wearable devices, or devices worn on the bodies of patients such as the Apple Watch, FitBit, or other monitoring devices, offer opportunities to innovate and improve cancer care delivery, from prevention and screening through survivorship.Fueled partly by the COVID-19 pandemic and growing need for remote and virtual care, the number of studies of technology-supported health interventions in cancer-including studies on wearables in cancer care-has increased dramatically in the past 3 to 5 years. 1,2 These studies underscore the potential offered by wearable devices across the landscape of cancer care delivery. Wearable devices may be used to support a range of activities including real-time monitoring of patient data, such as heart rate, blood pressure, or sleep quality, as well as allowing patients to directly report symptoms and mood using the wearable device. This allows for more accurate and ongoing healthrelated data collection for patients that can be used to manage and adjust treatment and to predict the potential for negative outcomes such as hospitalization. Integrating wearable devices into cancer care delivery holds promise for improving outcomes for patients across the care continuum.This issue of JAMA Oncology includes 2 studies that evaluated the use of wearable devices in different contexts. 3,4 Ito et al 3 assessed use of a wearable device (specifically the amuelink [Sony], which is worn around one's waist for the entire day) to measure physical activity in 119 patients with nonsmall cell lung cancer as an alternative to the traditionally used and patient-reported Eastern Cooperative Oncology Group Performance Scale (ECOG PS) to assess outcomes, including survival. They found that mean distanced walked obtained from the wearable was the best possible predictor of 6-month survival. The authors concluded that wearables, such as the one studied, provide more objective data than the subjectively assessed ECOG PS.Friesner et al 4 used data from 3 prospective trials in which a total of 214 patients with a variety of solid tumors (primary sites including the lung and head/neck) used wearable devices during chemoradiation. The authors' objective was to develop and validate a machine learning approach to predict unplanned hospitalizations associated with chemoradiation toxic effects based on daily step count from the wearable devices. The elastic net-regularized logistic regression (EN) with step count performed similarly to the EN with step count and clinical features. The authors concluded that EN models with step count plus clinical features or step count alone have potential to predict hospitalization during chemoradiation. The models developed from this work are being implemented in a larger study within NRG Oncology. 5 Opinion EDITORIAL jamaoncology.com (Reprinted)