The Walking Trail Making Test is more accurate than a Dual-Task Walking Test for screening the level of fall risk among community-dwelling older people

Background Preventing falls is a major issue for older people. Aging leads to a decline in executive functions and functional capacities, increasing the risk of falling. Most cognitive-motor assessments are additional Dual-Tasks (DTs), open to prioritization strategies, leading to heterogeneous results in the screening of the level of fall risk. The Walking Trail Making Test (W-TMT) is an incorporated cognitive-motor task sensitive to detecting cognitive impairments. A high level of fall risk is generally associated with higher cognitive impairments. The objective of this study is to evaluate the detection performance of the most complex condition of the W-TMT (WTMT-B) in the screening of people at a high level of fall risk, and to compare it to an additional DT, the 6-Meter Walking Test in the DT condition (6MWT-DT). Methodology 101 community-dwelling older people participated in the study. They were classified into three levels of fall risk (i.e., low, moderate, high) by the French Health Examination Center Fall Risk screening tool. They performed the three conditions of the W-TMT (N, A, and B) and the 6MWT in single-task and DT conditions. For both tasks, execution times were measured and compared between the levels of fall risk. Concordance index ( c -index) analysis was performed to evaluate and compare the overall classification performance of the W-TMT-B and the 6MWT-DT for the three levels of fall risk. Receiver operating characteristic analyses (providing the Area Under the Curve, AUC) were conducted on the same variables to evaluate and compare the capacity of both tasks to detect people at a high level of fall risk. Results For the W-TMT, ANCOVA (corrected by spontaneous walking speed) showed significant group*condition interaction ( F 4,194 = 8.27, p < .001). Post-hoc analyses showed that, only in the W-TMT-B condition, execution time was longer for people at a high level of fall risk vs moderate ( M = 30.69s, SD = 15.45s vs M = 21.54s, SD = 7.17s , p = .048) and low levels ( M = 20.16s, SD = 7.90s , p = .031). No differences were found for the 6MWT-DT execution times. The W-TMT-B better classified the levels of fall risk than the 6MWT-DT ( c -index of respectively 0.63, 95%CI = .52−.74 vs 0.48, CI%95 = .38−.59, p = .03). The W-TMT-B also detected people at a high level of fall risk better than the 6MWT-DT (AUC of respectively 0.72, 95%CI = .53−.90 vs 0.46, 95%CI = .30−.62; p = .03) with a sensitivity of 61.54% and a specificity of 85.23%. Conclusions Screening the level of fall risk with the W-TMT-B condition is more accurate than with the 6MWT-DT. The detection performance of the W-TMT-B condition could be explained by a high cognitive-motor interference mobilizing cognitive flexibility in a complex walking context. In contrast to additional DT, the incorporated feature of the W-TMT-B condition prevents any task prioritization strategy. Consequently, the W-TMT-B could be a simple and low-cost tool using only the execution time to detect people at a high level of fall risk in community-dwelling older people.