Usability and feasibility of the open-source GeRI platform for remote assessments in older adults with prostate cancer.

Abstract

e13664 Background: The Geriatric Remote Initiative (GeRI) is an open-source tool to collect patient-generated health data for older adults with cancer. It comprises a WiFi-enabled Raspberry Pi tablet, a Bangle.js 2 wrist accelerometer, and Mi Body Composition Scale 2 for data collection. Following the initial phase where feedback from older cancer survivors and caregivers refined a demonstration GeRI platform, this study deployed a live version of GeRI in advanced prostate cancer patients to evaluate its feasibility and usability. Methods: In this single-arm pilot feasibility and usability study, we evaluated GeRI among ten advanced prostate cancer patients aged ≥65 years undergoing androgen deprivation (ADT) across varied frailty risk categories. Participants interacted with GeRI for ≥12 weeks. Data collection comprised demographics, clinical information, and vulnerability assessment (Cancer and Aging Research Group [CARG] score) from in-clinic surveys. Biometric (wearable/scale) and electronic patient-reported outcomes data from the tablet were securely transmitted to the institution’s HIPAA-compliant cloud. Usability was evaluated using the System Usability Scale (SUS), aiming for an average score of 68 or higher. Feasibility was achieved if ≥6 participants completed at least 50% of all prompted interactions with the system components (surveys, wearable, and weight/bioimpedance) throughout a 12-week study period. Results: Ten men with metastatic hormone-sensitive prostate cancer [median age: 73 years; 40% minority; median of 3 months of ADT exposure (IQR: 2-4.5); 80% had intensified ADT with 60% getting concurrent prostate-directed radiation] engaged with GeRI for a median of 104 days (IQR: 93-131). 40% were vulnerable with a CARG>5. System usability, evaluated by SUS, yielded a mean score of 84 (SD: 12.5). All participants met the feasibility criterion. Deficiencies in obtaining complete 48-hour accelerometry data from 7 participants were noted with only 1 being due to persistent non-wear of the wearable and the remainder due to transmission failure. Conclusions: The GeRI platform exhibited exceptional usability and met all feasibility benchmarks in a geriatric oncology setting, highlighting its potential despite technical hurdles in accelerometry data collection. Further deployment of GeRI is underway [or planned] through an expanded program focusing on other cancer types, including head and neck, and more robust biometric data collection.