A "Hospital-Day-1" Model to Predict the Risk of Discharge to a Skilled Nursing Facility.

Autor:	Oseran AS; Division of General Internal Medicine, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, MA., Lage DE; Division of General Internal Medicine, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, MA., Jernigan MC; Division of General Internal Medicine, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, MA., Metlay JP; Division of General Internal Medicine, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, MA., Shah SJ; Division of Hospital Medicine, Department of Medicine, University of California, San Francisco, CA. Electronic address: sachin.shah@ucsf.edu.
Jazyk:	angličtina
Zdroj:	Journal of the American Medical Directors Association [J Am Med Dir Assoc] 2019 Jun; Vol. 20 (6), pp. 689-695.e5.
DOI:	10.1016/j.jamda.2019.03.035
Abstrakt:	Objective: To derive and validate a model to predict a patient's probability of skilled nursing facility (SNF) discharge using data available from day 1 of hospitalization. Design: Using a retrospective cohort of 11,380 hospitalized patients, we obtained administrative and electronic medical data to identify predictors of SNF discharge. Setting and Participants: Single, urban academic medical center. Patients older than 50 years admitted to the medical service from July 2014 to August 2015. Methods: Primary outcome defined as SNF discharge. We split the cohort into derivation and validation sets (80/20). We created 1000 bootstrapped samples of the derivation set and used backward selection logistic regression on each bootstrapped sample. The final model included variables selected in ≥60% of the samples. To create a point-based index, a point value was assigned to each predictor variable relative to the logistic regression coefficient. The model's discrimination, calibration, positive predictive value, and negative predictive value tested in the validation set. Results: The overall frequency of SNF discharge was 12%. The final model included 11 variables. Significant demographic variables included age, marital status, insurance type, living alone, residence, and distance from hospital. The final model included 2 significant functional variables (mobility, bathing) and 3 significant clinical variables (admission mode, admission diagnosis, admission day of week). Impairment in mobility [odds ratio (OR) 1.8, 95% confidence interval (CI) 1.4-2.2] and impairment in bathing (OR 1.9, 95% CI 1.6-2.4) were both significant predictors of SNF discharge. The final model discriminated well in the validation cohort (c-statistic = 0.82) and was well calibrated. Conclusions/implications: It is possible to predict the day of admission with good accuracy and clinical usability a patient's risk of SNF discharge. The ability to identify early in the hospitalization patients likely to use post-acute services has implications for clinicians, administrators, and policy makers working to improve discharge planning and care transitions. (Copyright © 2019 AMDA – The Society for Post-Acute and Long-Term Care Medicine. Published by Elsevier Inc. All rights reserved.)
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu