4 GEDI WISE: Evaluation of the Transitional Care Nurse Intervention Impact on Hospital Admissions from the Emergency Department

Abstract

With the recognition that inpatient costs continue to rise and the population is aging and will require greater health care services, the emergency department (ED) sits at the crossroads of multiple health care settings and plays a critical role in decisionmaking toward either inpatient or outpatient care and the transitions surrounding these decisions. GEDI WISE (Geriatric ED Innovations in care through Workforce, Informatics, and Structural Enhancements) is a CMMI Health Care Innovation Award program (1C1CMS331055-01). We will examine the impact of the GEDI WISE ED-based transitional care nurse (TCN) program on the outcome of hospital admission. Three US hospitals located in the Midwest and Northeast received CMMI support to implement interdisciplinary models of care and interventions that follow nationally endorsed Geriatric ED Guidelines. One of these interventions, the TCN program, consisted of emergency nurses at each site trained to facilitate care transitions for older ED patients with the goal of safely avoiding hospital admission. Patients targeted for assessment by the TCN included those with Identification of Senior At Risk score (ISAR) scores 2+, Emergency Severity Index (ESI) 2+, and those referred by ED clinicians for consultation by the TCN. Within each site, patients seen by a TCN will be compared to those not seen by a TCN using either coarsened exact matching (CEM) or entropy balancing. We will use these methods to account for selection bias and ensure patients in the TCN and comparison groups are similar on observed characteristics thought to be associated with both TCN use and hospital admission while preserving sample size. Patients will be matched on age, sex, night or weekend ED visit, acuity (ESI), ISAR, Charlson comorbidity score, and whether they had a hospital admission and discharge in the prior 30 days. Within each matched sample, we will use logistic regression to examine relationships between TCN use and hospital admission from the ED and subsequent hospital readmission within 30 days. Regression models will be adjusted for chief complaint, comorbidities, and variables used in matching. We will evaluate unique patient (age 65+) ED visits during the GEDI WISE implementation (1/1/13-6/30/15) at the 3 participating hospitals. Projected sample sizes: site 1 16,591; site 2 20,895; site 3 22,599. Of the older patients who visited the ED during the first 18 months of the study (combined sample 37,419), approximately 10% were seen by a TCN. Preliminary use of CEM, at least 90% of the TCN patients and 80% of the comparison group patients were found suitable for matched comparison. By site, early findings in adjusted analyses indicate the TCN intervention is associated with reduced likelihood of hospital admission: Odds Ratio (OR) ranging from 0.32 to 0.61 when compared to matched controls (ref=1). Early evaluation of a preliminary set of data for older ED patients evaluated by transitional care nurses demonstrate a reduced likelihood of hospital admission. We anticipate completed analyses with matched intervention to control subjects on a 30-month dataset (combined total sample 60,085) for 3 sites that also include readmission outcomes. The TCN intervention may be a promising and effective way to not only improve care transitions, but also reduce likelihood of hospital admission and readmissions.