Electronic patient-reported outcomes (ePRO)-based alerts deployed in clinical practice to inform treatment burden and care management in pancreatic cancer.

Abstract

11130 Background: Pancreatic ductal adenocarcinoma (PDAC) is characterized by aggressive growth and late-stage diagnosis. About 20% of patients are eligible for surgery at diagnosis and are usually accompanied by adjuvant/neoadjuvant chemotherapy and/or radiation, with high recurrence rates. The use of ePROs to inform patient symptom (SX) experience and guide clinical care is increasing. Objectives for this study were to assess treatment burden and management of PDAC and explore care team engagement in the clinical practice using ePRO-based alerts data. Methods: PDAC patients were enrolled in an ePRO platform called PROmpt from 9/2020 to 11/2024. Patients received weekly surveys to report PRO-CTCAE-derived SX experienced during treatment. When a SX was reported as moderate or severe, an algorithm-based system would generate an “alert” notification per SX to the care team who reviewed the alert, interacted with the patient, and recorded the response on the platform. Treatment burden was measured by the symptom prevalence and number of alerts per week. Results were explored by stage (I-II or III-IV), frailty status (fit vs. frail), and age (<65 vs. 65+). Time to alert resolution and clinical actions were described. Results: A total of 67 patients were included, of which 58 (86.6%) reported a moderate/severe SX at least once. Median age was 67 (range: 40-87), 56.7% male, 62.7% late stage, and 50.7% frail. Most patients (64.2%) received FOLFOX and 35.8% were on Gemcitabine-based regimen. No baseline difference in patients who reported an alert versus not. Patients reported, on average, 2.8 symptoms per week with median follow up of 9.3 weeks. Of 513 total alerts generated, 54.6% occurred in the first 8 weeks of treatment. The top SX triggering an alert were pain (24.4%), nausea/vomiting (15%), and decreased appetite (14%). Average number of alerts per patient per week was 1.7 (SD=1.2, Median=1). No significant difference was observed by frailty status or age, but patients in the late-stage group generated higher alert/week than early stage (1.8 vs. 1.6). Average time to alert resolution was 2.2 days with a median of 1 day. Most alerts (72.3%) were addressed in less than 2 days. The most common clinical action taken to resolve the alerts was monitoring (86%), which included contacting oncologist support as needed, increased patient education, or follow-up at next scheduled visit. Conclusions: The treatment burden in PDAC was high with patients experiencing about 3 symptoms per week, of which nearly 2 of them were moderate/severe. Care team engagement was high as evidenced by prompt SX mitigation. Data collected from an ePRO-based alerts system can be used to characterize treatment burden and care team response in PDAC. Future study should include broader sample size and assess the impact of alerts on health resource utilization in PDAC.