Activity Monitoring for Early Detection and Management of Toxicity in Patients Undergoing Chemoradiation for Gastrointestinal Malignancy

Abstract

<h3>Purpose/Objective(s)</h3> Higher levels of physical activity assessed by step count have been associated with lower risk of hospitalization during cancer treatment. We hypothesize that it is feasible to use activity data to identify patients undergoing concurrent chemoradiation for gastrointestinal (GI) malignancy who are at high risk for emergency department (ED) visits and hospitalizations, and to successfully trigger and execute triage visits for symptom management. <h3>Materials/Methods</h3> This prospective study randomized patients to activity monitoring versus observation. Each group was provided an activity monitor. If a patient in the intervention arm had 20% decreased activity or 20% increase in heart rate from their baseline, a triage visit was triggered to evaluate and treat symptoms. In the observation group, activity data was recorded but no triage visit was triggered. Baseline step count and heart rate were established during a one-week period between radiation simulation and treatment start. The primary endpoint was to demonstrate an increased rate of triage visits in the activity monitoring group compared to observation. Secondary outcomes included rates of ED visits and hospitalizations. Crude and adjusted odds ratios (OR) were computed using logistic regression modeling. <h3>Results</h3> A total of 40 patients were enrolled on the study: 22 in the intervention group and 18 in the observation group. Primary disease sites included anus (n = 4), gastric/esophagus (n = 14), hepatobiliary (n = 4), pancreas (n = 7), and rectum (n = 11). Median age was 60 years in the intervention arm and 62.5 years in the observation arm (<i>P</i> = 0.69). The median radiation dose and fractionation were similar among the two groups. Average baseline daily step counts were similar in the two groups (5,103 in intervention group vs. 5,668 in observation group, <i>P</i> = 0.43). Average daily step counts decreased from week 1 to week 5 in both groups (-960 steps in observation group and -1,164 steps in the intervention group). There was an increased rate of triage visits in the intervention arm compared to the observation arm (86.4% v 38.9%, OR 9.95, 95% CI 2.12-46.56, <i>P</i> = 0.015). Rates of ED visits and hospitalizations were numerically lower in the intervention group compared to the observation group (9.1% vs 22.2%, <i>P</i> = 0.31; 4.5% vs 16.7%, <i>P</i> = 0.31, respectively). Patients with anal (-1456 steps) cancer showed the largest decrease in mean daily step count over the treatment course. Medical intervention was more common in the intervention group compared to observation (<i>P</i> < 0.001). <h3>Conclusion</h3> This study supports the feasibility of actively monitoring patient's daily step and heart rate data to successfully trigger triage visits for patients at high risk for toxicity. Further studies are ongoing and may support the use of automated activity monitoring to decrease the rates of ED visits and hospitalizations.