Factors associated with discharge home in older patients admitted to emergency department observation units: Looking for a predictive scale.

Background: The selection of patients who are going to be admitted to an emergency department observation unit (EDOU) is essential for the good management of these units, intended fundamentally to avoid unnecessary hospitalization of patients. This is especially important when dealing with older patients. It would be important to know what factors are associated with discharge home and to have a clinical predictive scale that appropriately selects older patients who are going to be admitted to an EDOU.

Methods: A retrospective cross-sectional study was conducted of all patients ≥65 years of age assisted in 48 Spanish Emergency Departments for 7 consecutive days and were admitted to the EDOU. Demographics-functional, vital signs data and initial laboratory results were analyzed to investigate its association with discharge home and develop and validate a prediction model for discharge home from EDOU. Multivariable logistic regression was performed to develop a prediction model, and a scoring system was created.

Results: Among 5457 patients admitted to the EDOU from the emergency room, 2508 (46%) patients were discharged home, and 2949 (54%) were admitted to the hospital. Five variables were strongly associated with discharge home: the absence of fever (adjusted OR: 3.61, 95% CI:1.53-8.54), Glasgow Coma Scale score of 15 points (2.80, 1.63-4.82), absence of tachypnea (2.51, 1.74-3.64) or leukocytosis (2.07, 1.70-2.52) and oxygen saturation >94% (2.00, 1.64-2.43). The final model achieved an area under the receiver operating characteristic curve of 0.648 (IC95% = 0.627-0.668) in the development cohort and 0.635 (0.614-0.656) in the validation cohort.

Conclusions: There are factors associated with a greater probability of discharge home of older patients admitted to EDOUs. Prediction at the individual level remains elusive, as the best model obtained in this study did not have sufficient validity to be applied in the clinical setting.