An automated, EHR-integrated intervention system for patient-reported insomnia symptoms.

12058 Background: Insomnia affects >20% of patients on cancer therapy. It is consistently reported to be one of the most burdensome sequelae. Systematic screening for insomnia in cancer patients is uncommon. Without identification, patients cannot access care. Unfortunately, even if insomnia is identified, most cancer centers lack on-site sleep specialists. To address both the screening and treatment gaps, we leveraged our institution's EHR-based patient-reported outcomes (PRO) data collection system to develop a fully automated solution to identify patients with insomnia, and share an evidence-based psychoeducational video intervention with them. Methods: Prior to established outpatient appointments at our hospital, all patients are asked to complete eSyM, an electronic PRO-CTCAE symptom assessment. Two insomnia items (symptom severity and interference) are combined to generate an insomnia score (0-3; higher is worse, with present analyses for those reporting a score ≥2). Patients with an elevated insomnia score were sent a patient portal message explaining they were identified as having insomnia, and offered an educational video with strategies to improve their sleep. We tracked patient engagement, as well as report of insomnia symptoms via eSyM response at the patient’s subsequent appointment closest to 60 days post-Baseline (±30 days). Data was analyzed from patients who reported completed eSyM insomnia questions at both timepoints between 02/25 to 09/25. Results: Patients with insomnia (N=307) were an average of 63 years old, predominantly female (63%), White (84%), married (63%), and possessed a college degree or higher (57%). They were diagnosed with a broad range of cancers (breast = 20%, GI = 15%, thoracic = 10%). The vast majority (92%) opened the patient portal message with the video link, with the video watched in entirety 498 times. Overall, insomnia scores significantly decreased from baseline (x̄ = 2.3) to follow-up (x̄ = 1.9; p <.001), with 42% reporting symptomatic improvement at follow-up (48% no change and 9% worse). Specifically, 41% reported either Severe or Very Severe insomnia symptoms at baseline, improving to 30% at follow-up; 37% reported insomnia interfered with their life Quite a Bit or Very Much at baseline, improving to 23% at follow-up. Multivariate analyses revealed sociodemographic characteristics (Black, Medicare, and disabled) associated with less insomnia improvement. Conclusions: We developed a fully automated system that leverages electronic PRO data and our EHR to identify symptomatic cancer patients with insomnia and intervene. Our findings demonstrate that we can successfully connect with patients reporting insomnia and potentially improve outcomes with a low-intensity intervention. This framework can readily be deployed to identify and address other common side effects of cancer treatment (e.g., fatigue, pain) that impair recovery and quality-of-life.