Daily step counts to predict hospitalizations during concurrent chemoradiotherapy for solid tumors.

Abstract

293 Background: Wearable activity trackers could be useful tools in cancer care. Potential gains include better selection of patients fit for aggressive treatment, improved evaluation of patients during treatment, and provision of enhanced supportive care for patients. In a pilot study, low daily step counts were associated with increased risk for hospital admission during concurrent chemoradiotherapy for solid tumors of the head/neck, lung, or gastrointestinal tract. Here we seek to refine those findings using a larger patient cohort. Methods: Study subjects wore a commercial fitness tracker during a course of concurrent chemoradiotherapy. Patient characteristics, daily step counts, and hospital admissions between the initiation and completion of radiotherapy were tabulated. EORTC QLQ-C30 questionnaires were completed weekly, and summary QoL (range: 0-100) for each subject and week were calculated. Cox regression models with time-dependent and time-fixed covariates were used to identify predictors of first hospital admission during the radiotherapy course. Results: 188 subjects from three trials met inclusion criteria. The most common cancer diagnoses were head and neck (31%), lung (28%), and a gastrointestinal (30%, most commonly rectal [10%] or esophageal [6%]). Median age was 61 (range: 27-90). Baseline ECOG performance status was 0 for 46% of subjects, 1 for 47%, and 2 for 7%. The median patient compliance of wearing activity trackers and contributing data was 98%; 38subjects (20%) were hospitalized during the radiotherapy course. Cox regression demonstrated a 36% reduction in the risk of hospitalization for every 1000 daily steps (averaged over the past 3 days; hazard ratio 0.62, 95% confidence interval 0.48-0.81, p=0.0005). Diagnosis, age, baseline performance status, and QoL score were not statistically significant predictors of hospitalization risk. Conclusions: Continuous activity monitoring with wearable devices during concurrent chemoradiotherapy is feasible. We have validated daily step counts as dynamic predictors of hospitalization risk during treatment. Interventional trials should leverage activity data to optimize supportive care during cancer therapy.