Symptom alert by type and severity among cancer patients using electronic patient reported outcomes for remote symptom monitoring.

Abstract

340 Background: Remote symptom monitoring (RSM) by electronic patient-reported outcomes (ePRO) data can elicit actionable symptoms from patients with cancer. However, patients with different cancer diagnoses are likely to have differing symptom profiles and variability in symptom alerts. To understand potential workflow needs, this analysis was conducted to determine which types of symptoms and severity of alerts can be expected based on cancer type. Methods: Cancer patients initiating chemotherapy, immunotherapy, or targeted therapy at 2 academic cancer centers in Alabama, UAB O’Neal Comprehensive Cancer Center and USA Health Mitchell Cancer Institute (MCI), were enrolled in ePRO-based RSM. Site rollouts were differential: UAB enrolled by disease group starting May 2021, MCI by provider starting July 2021. Patients received weekly symptom surveys of selected PRO-CTCAE questions through the Carevive ePRO mobile platform (PROmpt), triggering alerts to clinical teams if reported symptoms were determined to be moderate or severe. Demographics, cancer diagnosis, and ePRO data were extracted from electronic health records and Carevive. Descriptive statistics of categorical variables were calculated by frequencies and percentages; Cramer’s V and Cohen’s d were used for associations and effect size. Results: UAB enrolled 598 patients and MCI enrolled 274 patients by April 2023, consistent with the patient volume difference of the centers. 68% of enrollees were White; MCI saw a moderately higher % of Black patients (V: 0.21). 67% of enrolled patients were female. Median age was 61 years (Interquartile range: 51-69); UAB patients were slightly younger (d: 0.15). Among 872 enrolled patients, 9765 symptom alerts were generated. There was a small effect of cancer type on the overall type of symptoms and symptom severity reported (V: 0.11 and 0.09 respectively). Of the total number of symptom alerts, 28.0% of the alerts generated were for pain, followed by nausea/vomiting (14.9%), and constipation (11.7%). When broken down by cancer type, pain was the symptom most frequently reported for each type. The next most frequently reported symptoms differed but were as expected by cancer type: coughing/dyspnea by lung cancer patients (20.6%); urinary complaints in genitourinary cancers (14.1%); and nausea/vomiting in gastrointestinal cancers (18.0%). The frequency of moderate alerts was 62.1%, varying from 34.0% in sarcoma to 66.6% in gastrointestinal cancers. 31.1% of the alerts were severe; sarcoma had the most severe alerts (56.0%) and hematologic had the least (27.1%). Conclusions: Across patients with differing cancer types, pain and gastrointestinal issues were over half the reported symptoms. However, variability by cancer diagnosis in both symptom type and severity was observed, suggesting the remote symptom management workload for providers may vary by cancer type.