Construct Validity, Test-Retest Reliability, Sensitivity to Change, and Feasibility of the Patient-Specific Functional Scale in Acutely Hospitalized Older Patients With and Without Cognitive Impairment.

Abstract

Background and purpose: The Patient-Specific Functional Scale (PSFS) is an individualized patient-reported outcome measure designed to assess the limitations and changes in self-determined functional activities most important to an older patient in the acute care setting. However, its clinimetric properties have not yet been evaluated in these patients. The study aimed to investigate the construct validity, test-retest reliability, sensitivity to change, and feasibility of the PSFS in acutely hospitalized older patients with and without cognitive impairment (CI).

Methods: The clinimetric properties of the PSFS were investigated by secondary data analysis from a prospective observational cohort study examining physical activity and mobility in acutely hospitalized older patients. In this analysis, 120 older patients-83.0 (6.4) years-with and without CI (Mini-Mental State Examination [MMSE] score 18-23, n = 52, and MMSE ≥24, n = 68, respectively) receiving early multidisciplinary geriatric rehabilitation in acute care were included. Construct validity was assessed by Spearman correlations (rs) with the Activity-specific Balance Confidence Scale (ABC-6), Short Falls Efficacy Scale-International (Short FES-I), EuroQoL-5 Dimensions (EQ-5D), Short Physical Performance Battery (SPPB), de Morton Mobility Index (DEMMI), and Barthel Index (BI); test-retest reliability within 24 hours by intraclass correlation coefficients (ICCs); sensitivity to change by standardized response means (SRMs) calculated for treatment effects, and feasibility by completion rates/times and floor/ceiling effects.

Results: The PSFS showed fair to moderate correlations with all construct variables in patients with CI (rs = 0.31 to 0.53). In patients without CI, correlations were fair for the ABC-6, FES-I, EQ-5D, and BI (rs = |0.27 to 0.36|), but low for the SPPB and DEMMI (rs =-0.04 to 0.14). Test-retest reliability (both: ICC = 0.76) and sensitivity to change (CI: SRM = 1.10, non-CI: SRM = 0.89) were excellent in both subgroups. Excellent feasibility was documented by high completion rates (>94%), brief completion times (<8 min), and no floor/ceiling effects in either subgroup.

Conclusions: The PSFS has adequate clinimetric properties for assessing patient-specific functional limitations and changes in acutely hospitalized older patients with and without CI. It might be an appropriate complement to traditional functional scales to enhance patient-centeredness in clinical geriatric assessment.