The association between wearable device metrics and clinical outcomes in oncology: A systematic review with evidence synthesis and meta-analysis
Milan Kos, Calvin G. Brouwer, Hanneke W.M. van Laarhoven, Maria T. E. Hopman, Martijn G.H. van Oijen, Laurien M. Buffart
Abstract
The emerging study of wearable devices (WDs) in patients with cancer provides opportunities to harness real-time patient data for predicting clinical outcomes. We conducted a systematic review with best evidence synthesis to examine the association between WD metrics and clinical outcomes in patients with cancer. MEDLINE and Embase were searched from inception until June 2022. Risk of bias assessment and best evidence synthesis were performed and, If possible, meta-analysis was conducted. A total of 34 studies was included. We found moderate-to-strong evidence for associations between circadian rest-activity metrics and OS. Disrupted I<O was associated with increased hazard for death (HR 2.08; 95 %CI: 1.50–2.88). For most associations there was insufficient evidence due to lack of studies (n = 32) or inconsistent results (n = 14). Meta-analysis was greatly hampered due to heterogeneity and different methodology used between studies. Studies primarily designed to investigate the association between WD metrics and clinical outcomes are warranted. • There is evidence for positive associations between circadian rhythm metrics and overall survival in patients with cancer. • Most data on associations between wearable device metrics and clinical outcomes come from small pilot or feasibility trials. • Heterogeneity between studies hampers comparison of results and quantification of associations between metrics and outcomes. • There are inconsistent findings regarding associations between the metric daily steps and clinical outcomes.