Psychometric properties of a computerized and hand-reaction time tests in older adults using long-term facilities with and without mild cognitive impairment.

Abstract The aim of this study was to determine the absolute and relative reliability and validity of two reaction time (RT) tests in older adults using long-term facility: the ruler drop test (RDT, a simple RT) and the Deary-Liewald RT task (DLRT, simple and four-choice RT). Participants (≥65 years old) using long-term facilities were distributed into a group without cognitive impairment (GWCI, n = 41), and a group with mild cognitive impairment (GCI, n = 32). The tests were administered one week apart, by the same evaluator. Relative reliability was measured by the Intraclass Correlation Coefficient (ICC3.1), absolute reliability by the standard error of measurement (SEM) and minimal detectable change (MDC95) and convergent validity by the Pearson correlation. The results showed that the relative reliability was good for the RDT (GWCI, ICC = 0.84; GCI, ICC = 0.80) and moderate for the Deary Liewald RT simple (DLRT-S) (GWCI, ICC = 0.61; GCI, ICC = 0.65). The relative reliability of the Deary Liewald RT choice (DLRT-C) was good (ICC = 0.89) for the GWCI and excellent (ICC = 0.93) for the GCI. However, the DLRT-C had low feasibility in the GCI, as most participants were unable to complete the test. Measurement precision was acceptable for the RDT and DLRT-C (GWCI) and for the DLRT-C (GCI). There were no significant associations between the RDT and DLRT scores. We conclude that the RDT could be more suitable than the DLRT-S for assessing simple RT in older adults. Future studies should explore if fewer than the four stimulus-response alternatives used in the DLRT-C could be more appropriate for assessing choice RT in older adults with cognitive impairment.