Background: Prognostic stratification of older patients with complex medical problems among those who access the emergency department (ED) may improve the effectiveness of geriatric interventions. Whether such targeting can be performed through simple administrative data is unknown.
Methods: We examined the discharge records for 10,913 patients aged 75 years or older admitted during 2005 to the ED of all public hospitals in Florence, Italy. Using information on demographics, drug treatment, previous hospital admissions, and discharge diagnoses, we developed a 1-year mortality prognostic index. The predictive validity of this index was tested in a subsample of patients independent of the subsample used for its original development. Finally, we tested whether patients stratified by the prognostic index had different mortality when admitted to a geriatrics compared with an internal medicine ward.
Results: In the validation subsample, patients with scores of 4-6, 7-10, and 11+ compared with those with scores less than 4 had hazard ratios (95% confidence interval) for 1-year mortality of, respectively, 1.5 (1.3-1.7), 2.2 (1.3-1.7), and 3.0 (2.6-3.4). Patients in the worse prognostic stratum experienced 33% higher mortality when admitted to an internal medicine compared with a geriatrics ward, although mortality was not significantly affected by the type of ward of admission in all other risk strata.
Conclusions: Simple administrative data provide prognostic information on long-term mortality in older patients hospitalized via ED. Patients with worse prognostic index scores appear to benefit from admission in a geriatrics compared with an internal medicine ward.