Operational feasibility and impact of a digital platform for symptom management in oncology: A real-world healthcare provider perspective.

Abstract

e13724 Background: Digital health technologies are emerging as valuable tools for enhancing patient monitoring, improving operational efficiency, and delivering high-quality, patient-centered care. Evaluating their effectiveness in real-world healthcare settings is essential to understanding their impact. We aimed to assess the operational feasibility and impact of a mHealth ePRO platform (ThummiOnco) for telemonitoring cancer patients. Methods: This prospective pilot study included patients (pts) aged > 18 years undergoing systemic treatment for cancer at a major private healthcare provider in Brazil. Eligible patients (pts) were identified through an automated, AI-driven screening process using the provider's database. Telemonitoring was conducted by nursing navigators, following a standard protocol. Medical oncologist consultations were arranged as clinically required. Baseline characteristics were collected. Operational feasibility was assessed using key performance indicators, including: patient activation rate (PAR) – e.g., proportion of pts who completed onboarding; engagement rate (ER) – e.g., proportion of pts actively interacting with the platform in the last month; net promoter score (NPS); and healthcare resource utilization (HRU) – e.g., proportion of pts experiencing critical events (hospitalizations, emergency visits) in the past 30 days, by risk level. HRU metrics were compared to a reference group (identified via AI database screening) and historical cross-institutional benchmarks. Descriptive statistics were performed. Results: Between June 2023 to October 2023, 26 pts were enrolled. The median age was 50 years (range: 24-74 years). Most pts were female (n = 14, 65.4%), had breast cancer (n = 8, 30.8%), presented with advanced disease (stages III-IV; n = 17, 73.7%), and were receiving chemotherapy (n = 13, 50%). A total of 108 symptoms were reported, mostly grade I-II (n = 87, 80.5%), with fatigue (n = 16, 14.8%) and joint pain (n = 12, 11.1%) being the most common. The majority of symptoms (n = 104, 96.3%) were resolved exclusively by the nursing team. The PAR was 68.5%, while the ER and NPS were 65.4% and 66.7%, respectively. Regarding HRU, the distribution across risk levels was 73.7% low, 10.5% moderate, 5.3% high, and 10.5% critical. When compared to the reference group, there was a reduction of 3.7% in the high-risk group (9% to 5.3%) and 6.7% in the critical-risk group (17.2% to 10.5%). The overall hospitalization rate was 8%, a 2.4% reduction compared to the cross-institutional benchmark (10.4%). Conclusions: The platform demonstrated high patient activation and engagement rates, with most symptoms being resolved by the nursing team alone. HRU rates were lower compared to the provider's historical benchmarks. Large-scale clinical trials are needed to validate the platform and further strengthen its evidence base.