Development and Validation of the Emergency Department Geriatric Readmission Assessment at Yale (ED GRAY): Part 2, Prognostic Accuracy
Lori A. Post *
Yale University, USA
Thomas L. Conner
Michigan State University, USA.
James F. Oehmke
USAID, USA
Fuad Abujarad
Yale University, USA
Leo M. Cooney
Yale University, USA
Cynthia Brandt
Yale University, USA
Connie Page
Michigan State University, USA
Sarah Swierenga
Michigan State University, USA
Brian Biroscak
Yale University, USA.
Christal Esposito
Yale University, USA
Jim Dziura
Yale University, USA
*Author to whom correspondence should be addressed.
Abstract
Objective: The Emergency Department (ED) Geriatric Readmission Assessment (ED-GRAY) is a 16-item questionnaire we developed that can be used to evaluate the prognostic ability to determine if geriatric patients will recidivate to the ED, be admitted or die within 30 days.
Methods: We conducted an observational prospective cohort study at a single academic urban university-affiliated hospital. Subjects (n=250) were consenting, non-critically ill, English-speaking adults older than 65 years and receiving care in the ED. Multivariable logistic regression and receiver operating characteristic (ROC) curves were used to evaluate the ability to accurately predict the likelihood of a 30-day event: subsequent ED visit, hospitalization, or death.
Results: 56 participants (22%) experienced at least one 30-day return visit or death. Greater disability as measured by the ED GRAY global disability was associated with an increased likelihood of an event (OR=1.7 for each 1-point worsening in severity; 95% CI 1.2, 2.5). In the multivariable model, prior ED visits (OR=2.7, 95% CI=1.4, 5.2), ED GRAY global score (OR=1.4, 95% CI=1.0, 2.1), and age (OR=1.03, 95% CI=0.99, 1.07) were associated with a greater likelihood of a 30-day event. The fit of the multivariable model was good (Hosmer-Lemeshow Goodness of Fit test, p=0.85) and it provided good discrimination between those having and not having a 30-day event (AUROC=0.70). The predicted probability of a return visit ranged from 3–56%.
Conclusions: ED GRAY demonstrated good predictive validity when combined with prior ED visits, poly-pharmacy and age.
Keywords: Geriatric disability, emergency department visits, disability measurement, prognosis