Real-world implementation of an EHR-integrated, SMS text remote patient monitoring platform at an academic cancer center.

e13669 Background: Remote monitoring studies in oncology have shown improved quality of life, decreased healthcare utilization, and even improved overall survival in an oncologic population, however applications in a real-world setting have been limited. Barriers to successful deployment include low patient engagement rates with smartphone apps, lack of integration with the EHR and burdensome clinical workflows for staff. Methods: We sought to assess the feasibility of an EHR-integrated, SMS text conversational AI agent to automate the collection of ePRO data as well as the assessment and triage of common patient symptoms for patients with GI and GU tumors receiving anti-cancer therapy. This program was developed collaboratively by a health system digital health team, nursing informatics, IT support resources, oncology nurses, medical oncologists, and representatives from Memora Health. Core components included 1) weekly ePRO collection based on the NIH’s CTCAE, 2) NLP activated symptom triage for 16 common symptoms and toxicities 3) weekly symptom check-in 4) weekly medication adherence assessment for patients on oral therapy. Patients received all messages through SMS on their phones, and all data collected were routed into the EHR. In-EHR visualizations allowed nurse coordinators to track ePRO symptoms over time and symptoms that met pre-specified criteria triggered EHR-based alerts. Results: Of 99 patients ordered into this SMS program, 83 enrolled (84% enrollment rate), with 55 Uro Onc patients; 44 GI Onc patients. A total of 2,583 surveys were sent to patients. 98% of patients completed at least 1 survey. Overall patient survey completion rate was 62%. Medication adherence surveys had the highest completion rate at 81%. ePRO completion rates were higher for GU (65%) vs GI (39%) patients (p<0.01). 71% of ePRO surveys generated a symptom alert to the nursing staff. Patient satisfaction was high with 93% rating this SMS program as “excellent” or “good” (n=78). Care team surveys found that 3 of the 6 Oncologists believe ePRO collection is valuable and 3 of 6 used ePRO data to adjust a care plan. Nurse coordinators had a mixed view of the program with a majority reporting that there were too many symptom alerts. Conclusions: We successfully deployed an EHR-integrated SMS program to collect ePRO data, track patient symptoms, and monitor adherence to oral therapy. We believe having a physician champion influenced patient engagement in the GU patient population. Patients had a favorable impression of this program and survey completion rates demonstrated ease of use for patients. Clinical staff, particularly nurses, felt that there were too many alerts. Additional workflow changes may reduce alert burden and better incorporate ePRO learnings into clinical practice.