A novel performance-based balance assessment for community-living older adults

Han Chen

doi:10.17918/etd-7721

Falls are serious and life threatening for community-living older adults (CLOAs). Every 20 minutes an older adult in the United States dies owing to a fall. Falls can also reduce older adults' independence. Balance assessment is critical to detect fall risk and provide early prevention strategies. To systematically assess balance ability, optimal clinical balance measures include tests of at least 5 balance domains; specifically, quiet stance, anticipatory postural adjustments to perturbations caused by self-initiated movements, response to external perturbations, sensory orientation, and stability in gait. However, current clinical balance tests frequently have limitations including lack of comprehensive testing of balance domains or capacity to guide interventions, ceiling effects, high test burden, and fixed-test formats. The purposes of this study were to: 1) construct a new Performance-based Balance Scale (PBS) for use with community-living older adults; 2) determine the practicality, clarity of instructions, and time to administer; 3) identify and validate the non-fixed test format; and 4) determine the psychometric properties of the PBS. We developed a 26-item, dichotomously-scored initial draft of the PBS using an expert panel by adapting test items from existing balance measures and organizing these items by difficulty into the 5 balance domains. We pilot tested the initial draft of the PBS for practicality, clarity of instructions, and administration time with 17 participants (4 men, 13 women; mean age = 81.8 yrs, 54 - 94 yrs). After revision, we created and tested a preliminary PBS with 35 participants (10 men, 25 women; mean age = 75.9, 60 - 91 yrs). Then using the Rasch dichotomous model, we created the non-fixed-item format. We re-tested 32 participants (9 men, 23 women; mean age = 75.5, 60 - 91 yrs) of the same participants 4-8 weeks later using the non-fixed-item PBS to validate the format using Spearman's correlation coefficient (rs = .94) and a paired t-test that showed no difference between the 2 scores, t (31) = -.32 (95% CI: -.70 - .51), p = .75. Further, we eliminated redundant test items and reduced the final PBS to 20 items. To determine the psychometric properties we tested 24 CLOAs (1 man, 23 women; mean age = 71.2 yrs, 62-86 yrs) twice, 7 to 14 days apart. Results showed that the typical time to administer the PBS was 15 to 20 minutes. Test-retest reliability was determined using the intraclass correlation coefficient (ICC(2,1) = .95) and method error (CVME = 9.3%). Internal consistency, validity, minimal detectable change (MDC), cutoff score, sensitivity, specificity, likelihoods ratios and overall diagnostic accuracy were calculated. The PBS showed high internal consistency (Cronbach's [alpha] = .94) and good to excellent convergent validity (r = .81) with the Short Physical Performance Battery (SPPB). The SEM for the PBS = 1.2, MDC90% = 2.8, and MDC95% = 3.3. The cutoff score to identify fallers was 14/20, with sensitivity of .78 and specificity of .47. Overall diagnostic accuracy was 58%. In conclusion, we provided clinicians with the first reliable and valid comprehensive non-fixed-item balance test for CLOAs. The PBS also reduces test burden, ceiling effects, and may have the capacity to guide targeted interventions. However, the diagnostic accuracy suggests that the PBS should not be used alone to determine fall risk. Future prospective studies with larger and more representative samples are needed to further investigate the diagnostic accuracy for predicting falls and to address the need for estimating the minimal clinical important difference (MCID).

A novel performance-based balance assessment for community-living older adults

Files and links (1)

Abstract

Metrics

Details

A novel performance-based balance assessment for community-living older adults

Files and links (1)

Abstract

Metrics

Details

Drexel University Social media