Electronic Health Record-Enabled Identification and Targeted Multipronged Intervention to Reduce Readmission and Length of Stay Among Patients With Metastatic Cancer.

Abstract

Patients with metastatic cancer experience increased rates of unplanned acute care, hospitalizations, and mortality. Risk-stratification techniques have been developed, but research studying the sustainable implementation of targeted interventions are lacking. We aimed to demonstrate an integrated program through the early identification of patients with high-risk cancer and multipronged intervention to reduce acute care utilization and mortality. We conducted a prospective quality improvement project within a health system, including the main tertiary, specialty, and community hospital settings. We used the electronic health record (EHR) to identify admitted patients with metastatic cancer. Targeted interventions included (1) coordination of care, (2) inpatient consultation, (3) patient education and symptom management, and (4) post-discharge follow-up. Six-hundred intervention and 103 baseline patients were included. Primary outcome variables were 30-day readmission, hospital length of stay (LOS), and 30-day mortality. We used quality control run charts to identify special cause variation and multivariable Cox regressions to identify associations between groups while adjusting for demographic factors. A <i>P</i> < .10 was considered significant. There was special cause variation with a 9.4% decrease in 30-day readmission (41.2% at baseline to 31.8% postintervention). LOS decreased by 1.14 days (5.65 days at baseline to 4.51 days). Survival analyses confirmed a reduction in 30-day readmission (hazard ratio [HR], 0.69 [0.46-1.01]; <i>P</i> = .057) and a significant increase in the hazard rate of discharge in the intervention group compared with the baseline group (HR, 1.26 [1.01-1.56]; <i>P</i> = .034), translating to a 1-day decrease in LOS. There was no statistically significant change in 30-day mortality. EHR-enabled risk stratification and identification of patients with high-risk metastatic cancer enabled delivery of targeted intervention in a real-world setting, resulting in decreased 30-day readmission and LOS.