OBJECTIVE: To establish the psychometric properties of a simple 'low-tech' choice stepping reaction time test (CSRT-M) by investigating its validity and test-retest reliability. DESIGN: Cross-sectional. SETTING: Community. SUBJECTS: A total of 169 older people from the control arm of a clinical trial and a convenience sample of 30 older people. MAIN MEASURES: Demographic, physical, cognitive and prospective falls data were collected in addition to CSRT-M. The CSRT-M time was taken as the total time to complete 20 steps onto four targets printed on a portable rubber mat. Assessment of the original electronic version (CSRT-E) and re-administration of the CSRT-M the next day was done in 30 participants. RESULTS: Multivariate regression analysis showed that the CSRT-M time was best explained by leaning balance control, quadriceps strength and cognitive functioning (R2 = 0.44). Performance on the CSRT-M was worse in older participants and participants with a presence of fall risk factors, supporting good discriminant validity. The odds of suffering multiple future falls increased by 74% (odds ratio (OR) = 1.74, 95% CI (confidence interval) = 1.14-2.65, p = 0.010) for each standard deviation increase in CSRT-M, supporting good predictive validity. Criterion validity was confirmed by a strong bivariate correlation between CSRT-M and CSRT-E (0.81, p < 0.001). Test-retest reliability for the CSRT-M was good (intraclass correlation coefficient = 0.74, 95% CI = 0.45-0.88, p < 0.001). CONCLUSIONS: A simple test of unplanned volitional stepping (CSRT-M) has excellent predictive validity for future falls, good inter-day test-retest reliability and excellent criterion validity with respect to the well-validated CSRT-E. The CSRT-M, therefore, may be a useful fall risk screening tool for older people.
a1477-0873 (Electronic)