Dementia-specific adaptations to physical performance tests of balance, mobility, and lower limb strength and...

Valid and reliable physical performance tests are crucial for accurately assessing the physical performance of people with dementia (PwD) and for evaluating the effects of interventions. However, existing physical performance tests for PwD often show insufficient reliability. This study aims to investigate the reliability of physical performance tests of balance, mobility and lower limb strength and function that were specifically adapted for PwD.

We conducted a reliability study with test-retest design and a one-week gap between tests among PwD living in nursing homes. Adaptations were made to either the instruction and administration, or the scoring of the three physical performance tests as follows: The Frailty and Injuries: Cooperative Studies of Intervention Techniques (FICSIT) for balance (adaptations to administration and scoring), the Timed-Up and Go Test (TUG) and its five phases (standing up, walking, turning around, walking, sitting down) for mobility (adaptations to instruction, administration and scoring in terms of analysing the five phases separately, and the Sit-to-Stand test (STS) for lower limb strength and function (adaptations to instruction and administration). The tests were standardized in terms of cues used to administer and guide them. We assessed absolute reliability (Standard Error of Measurement; Coefficient of Variance, Minimal Detectable Change) and relative reliability (Intraclass correlation coefficient; weighted Cohen's Kappa). The absolute and relative test-retest reliability of the PP tests was assessed.

We examined relative and absolute reliability values of PP tests in a sample of 26 PwD (mean age, 88 years; mean Mini Mental State Examination (MMSE) score, 14). No statistically significant differences were found between baseline and retest. Relative reliability values ranged from 0.258 to 0.505 for balance (FICSIT), 0.011 to 0.860 for mobility (TUG), and 0.506 to 0.678 for lower limb strength and function (STS). Absolute reliability values as indicated by the coefficient of variation (CV) ranged from 23.5 to 92.8.

Adaptations regarding test administration and/ or scoring did not improve reliability values as compared to the original test versions. TUG test phases showed the highest reliability values for the gait phases. Future adaptations should focus on reducing the cognitive component of demand during physical performance tests. Assistive technologies such as augmented reality could improve test reliability by providing more consistent and controlled test environment.

German Clinical Trials Register DRKS00019205, retrospectively registered, registration date: 11 November 2019.