Background: Depression is common in patients with cancer and is associated with worse cancer treatment outcomes. Oncology nurses and patient navigators (ONPNs) can help patients manage depression and distress associated with a cancer diagnosis and during the cancer treatment journey. However, this benefit is only realized to the extent that the depression diagnosis is made, and proper intervention is implemented. Understanding that depression is often underdiagnosed or undertreated as cancer clinicians are focused on the more complex aspects of therapy and care coordination, the staff at Rainier Hematology/Oncology of Northwest Medical Specialties (NWMS) sought a solution. Augmented intelligence (AI) has a potential application in identifying patients at high risk for depression. Jvion Inc has developed a prescriptive analytics solution (the Jvion CORE) that uses AI algorithms and machine learning techniques applied to combined clinical and exogenous data sets to identify patients with a tendency for poor clinical outcomes. Based on its analyses, the Jvion CORE generates patient-specific, dynamic, and actionable insights without the need for additional documentation within the electronic health record (EHR). The insights generated can be integrated back into any EHR to help inform the care plan.
Objectives: To apply the Jvion CORE insights to patients’ risk for depression within the next 6 months to increase depression screenings and depression care delivery to cancer patients at this single oncology practice.
Methods: Patients are scored weekly using the Jvion CORE depression product. The AI tool risk-stratifies patients and generates recommendations for the care team. Patients identified as “at risk” in 5 or more vectors were flagged for the care coordinator to schedule a supportive care visit with an advanced practice provider. Patients are screened with the Patient Health Questionnaire Depression Module (PHQ-9) and a National Comprehensive Cancer Network Distress Thermometer before the visit. ONPNs assess the patient’s wellness screening irrespective of previous results and refer positive screenings to social work for further assessment. Patients with a confirmed depression diagnosis were managed appropriately (medication and mental health services). The rate per 1000 unique patients per month (PPM) of depression screenings, case management evaluations, and antidepressant prescriptions were calculated for the 5 months before and 17 months after deployment of the Jvion CORE at NWMS.
Results: NWMS has 21 providers managing 4329 unique PPM on average. The rate of depression screenings increased 270% since implementing the Jvion CORE, rising from 6.0 to 16.2 per 1000 PPM. The downstream workflow outcomes of case management evaluations increased 184% (rising from 11.6 to 21.4 per 1000 PPM), and antidepressant prescriptions increased 168% (from 9.2 to 15.5 per 1000 PPM). Providers reported high satisfaction with the use of the AI solution in depression screening.
Conclusions: Insights for depression risk generated by the Jvion CORE were actionable, could be incorporated into workflow, and increased the number of patients identified. With AI-improved (quicker and effectual) selection of at-risk patients, ONPNs could intervene sooner and deliver the care needed to manage depression. If confirmed in more extensive studies, AI-driven insights may improve the identification and management of depression in patients with cancer.