Predicting Discharge Destination in Older People From Acute General Medical Wards: A Systematic Review of the Psychometric Properties of 23 Assessment Tools.

Background: Predicting discharge in older people from general medicine wards is challenging. It requires consideration of function, mobility, and cognitive levels, which vary within the cohort and may fluctuate over a short period. A previous systematic review identified 23 assessment tools associated with discharge destination in this cohort; however, the psychometric properties of these tools have not been explored.

Purpose: To evaluate, synthesize, and compare the psychometric properties of 23 assessment tools used to predict discharge destination from acute general medical wards.

Methods: Four databases were systematically searched: Medline (Ovid), Embase (Ovid), Cumulative Index of Nursing and Allied Health Literature (CINAHL), and Evidence-Based Medicine Review databases. Studies were included if participants were from general medicine or acute geriatric wards and investigated at least one psychometric property (reliability, internal consistency, measurement error, responsiveness, hypothesis testing, and structural or criterion validity) in 23 previously identified assessment tools. Data were extracted and methodological quality were assessed independently by 2 assessors using the COnsensus-based Standards for selection of health Measure INstruments (COSMIN) checklist. As per the COSMIN checklist, results were rated against "sufficient," "insufficient," or "indeterminate."

Results: Forty-one studies were included. The de Morton Mobility Index (DEMMI) was the most rigorously evaluated assessment tool; it scored "sufficient" psychometric properties in 5 of 7 psychometric categories. The Alpha Functional Independence Measure (AlphaFIM), Barthel Index, and Mini-Mental State Examination (MMSE) demonstrated "sufficient" psychometric properties in at least 3 psychometric categories. The remainder of the tools (n = 19, 83%) had "sufficient" psychometric properties in 2 or fewer psychometric categories.

Discussion and conclusion: Based on current evidence, out of 23 assessment tools associated with discharge destination in acute general medicine, the DEMMI has the strongest psychometric properties. Other tools with substantial evidence in this cohort include the AlphaFIM, MMSE, and Barthel Index. Research is required to thoroughly evaluate the psychometric properties of the remaining tools, which have been insufficiently researched to date. Results can be used by physical therapists to guide selection of appropriate tools to assess mobility and predict discharge destination.

Trial registration: A priori, PROSPERO (CRD 42017064209).