Implementing a clinical risk prediction tool for patients undergoing active cancer treatment.

Abstract

TPS2086 Background: Acute care utilization (ACU), encompassing both emergency department visits and hospitalizations, is common in patients with cancer, with nearly three quarters of patients with advanced disease hospitalized at least once in the year after their diagnosis. Efforts to prospectively identify these patients prior to ACU have led to the development of a variety of scoring systems for specific cancer patient populations, including the elderly and those initiating palliative infusional chemotherapy. Prospectively identifying patients may enable early interventions to reduce ACU. However, few studies have demonstrated effective implementation of such prediction tools in clinical practice. We developed an oncology risk score (ORS) for active oncology patients (defined as patients with an active cancer diagnosis in the last 12 months who had a Medical Oncology encounter in a 180-day period ) to prospectively determine risk of ACU. Patients are defined as high risk (18% of patients, accounting for 57% of historical ACU), intermediate risk (25% of patients, accounting for 25% of ACU), or low risk (56% of patients, accounting for 18% of ACU) by the ORS. We are currently deploying a pragmatic implementation initiative to evaluate the impact of targeted nurse navigator (NN) outreach to patients defined as high risk for ACU by the ORS. Methods: The ORS is embedded within the health system electronic medical record. The ORS will be queried on a weekly basis. NNs will contact identified patients, prioritizing patients not yet identified by the navigation team by other means. Following chart review, NNs will either meet patients in person (if a visit is already planned within 24 hours) or complete standard navigation outreach and documentation (consisting of phone call and barrier assessment, as well as appropriate nursing intervention) if no visit is planned. NNs will determine follow up cadence based on clinical judgement. Efficacy will be determined using a case-control method. Case patients will be OCM patients defined as high risk by the ORS (historical n = 289); control patients will be non-OCM high risk patients (historical n = 388). The total number of patients in the case and control groups, as well as the proportion of patients in the group utilizing acute care, will be monitored over time. Proportion of high risk patients known to navigation will be tracked. ACU in medium and low risk groups will also be monitored. Targeted outreach to high risk patients using the ORS began on 2/5/2019.