Validity, reliability and feasibility of tools to identify frail older patients in inpatient hospital care: a systematic review

Background: The objective of this study is to identify and review screening tools for frailty in older adults admitted to inpatient hospital care with respect to their validity, reliability and feasibility. Methods: Studies were identified through systematically searching PubMed, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Embase and PsycINFO and screening reference lists till June 2014. Papers dealing with screening tools aimed at identifying frail older patients in in-hospital care, and including information about validity, reliability or feasibility, were included in the review. The quality of the included studies was critically appraised via the Quality Assessment of Diagnostic Accuracy Studies (QUADAS). Results: From the originally identified 2001 studies 32 studies met the inclusion criteria, in which 16 screening tools were presented. The screening tools showed different characteristics with respect to the number of items, the method of administration and the domains included. The most frequently studied tools with respect to predictive validity were the Identification Seniors At Risk (ISAR) and Triage Risk Stratification Tool (TRST). Studies retrieved poorer information about reliability and feasibility. Overall sensitivity was fairly good. The ISAR, ISAR-HP (Identification Seniors At Risk Hospitalized Patients) and Multidimensional Prognostic Index (MPI) generally had the best sensitivity. Conclusions: Many screening tools are available for daily practice. These tools to identify frail older patients in inpatient hospital care could be useful. For no tool, however, is clear evidence available yet regarding validity, reliability and feasibility. The overall sensitivity of the included screening tools was fairly good, whereas information on reliability and feasibility was lacking for most tools. In future research more attention should be given to the latter items.