Two Easy-to-Use Scores to Identify Acute Geriatric Patients With High Rehabilitation Success.

OBJECTIVES Many geriatric patients who would benefit from a stay in an acute geriatric ward (AGW) are not identified because of the lack of objective and standardized predictive measures for rehabilitation success. We thus aimed to develop 2 easy-to-use, standardized predictive scores, Discharged Home Score (DHS) and Barthel Index Improvement Score (BIS), to identify such geriatric patients by predicting the rehabilitation success of a stay in an AGW. DESIGN Retrospective cohort study. SETTING AND PARTICIPANTS This study analyzed data from 78,654 patient stays at 24 Austrian AGWs, incorporating comprehensive geriatric assessment. METHODS DHS (predicting care situation after a stay in an AGW) and BIS (predicting Barthel Index improvement after a stay in an AGW) were developed using the largest multicenter dataset to date. First, feature selection was performed based on clinical discussion and permutation feature importance. Subsequently, both scores were developed using regression models (DHS: logistic regression, BIS: linear regression). The scores were evaluated using a randomly sampled dataset and one that consisted of the most recent data. RESULTS Area under the receiver operating characteristic curve ranged from 0.775 to 0.827 for DHS and from 0.692 to 0.711 for BIS. BIS could explain between 17.7% and 22.1% of the variances. Score performances varied considerably across the AGWs. CONCLUSIONS AND IMPLICATIONS The developed standardized scores DHS and BIS showed acceptable performance and, although they provide independent predictions, together they offer a nuanced assessment of rehabilitation success. DHS is easier to apply, as it does not require the Mini Mental State Examination, whereas BIS gives a more precise assessment of the expected rehabilitation success. The predictive power of BIS is limited because only about 20% of the Barthel Index improvement can be explained. Both scores are not validated and are therefore not allowed to be used in the clinical setting. After further refinement and validation, both scores have great potential to be integrated into the clinical practice to support effective identification of geriatric patients who benefit from a stay in an AGW, thereby ensuring timely therapy and preventing complications.