Wearable activity monitors to assess performance status and predict clinical outcomes in advanced cancer patients

• DOI: 10.1038/s41746-018-0032-6
• Publication Date: 2018-06-18T11:03:16Z
• Authors: Gillian Gresham; Andrew E. Hendifar; Brennan Spiegel; Elad Neeman; Richard Tuli; B. J. Rimel; Robert A. Figlin; Curtis L. Meinert; Steven Piantadosi; Arvind M. Shinde
• Keywords: Odds; Activity monitor

An objective evaluation of patient performance status (PS) is difficult because patients spend the majority their time outside clinic, self-report to providers, and undergo dynamic changes throughout treatment experience. Real-time, activity data may allow for a more accurate assessment PS physical function, while reducing subjectivity bias associated with current assessments. Consenting advanced cancer wore wearble monitor three consecutive visits in prospective, single-cohort clinical trial. Provider-assessed (ECOG/Karnofsky) NIH PROMIS® patient-reported outcomes (PROs) were assessed at each visit. Associations between wearable metrics (steps, distance, stairs) PS, (adverse events, hospitalizations, survival), PROs using correlation statistics multivariable logistic regression models. Thirty-seven evaluated (54% male, median 62 years). Patients averaged 3700 steps, 1.7 miles, 3 flights stairs per day. Highest correlations observed average daily steps ECOG-PS KPS (r = 0.63 r 0.69, respectively p < 0.01). Each 1000 steps/day increase was reduced odds adverse events (OR: 0.34, 95% CI 0.13, 0.94), hospitalizations 0.21 0.56, 0.79), hazard death (HR: 0.48 0.28-0.83). Significant also PROs. Our trial demonstrates feasibility monitors assess suggests potential use predict outcomes. These findings should be validated larger, randomized trials.