How Reliable and Valid are Dual-Task Cost Metrics? A Meta-analysis of Locomotor-Cognitive Dual-Task Paradigms.

OBJECTIVE: To assess the retest reliability, predictive validity, and concurrent validity of locomotor and cognitive dual-task cost (DTC) metrics derived from locomotor-cognitive dual-task paradigms. DATA SOURCES: A literature search of electronic databases (PubMed, PsycINFO, MEDLINE, CINAHL, and Scopus) was conducted on May 29th, 2021, without time restriction. STUDY SELECTION: For 1559 search results, titles and abstracts were screened by a single reviewer and full text of potentially eligible papers was considered by 2 independent reviewers. 25 studies that evaluated retest reliability, predictive validity, and concurrent validity of locomotor-cognitive DTC in healthy and clinical groups met inclusion criteria. DATA EXTRACTION: Study quality was assessed using the Consensus-Based Standards for the Selection of Health Measurement Instrument checklist. Data relating to the retest reliability, predictive validity, and concurrent validity of DTC were extracted. DATA SYNTHESIS: Meta-analysis showed that locomotor DTC metrics (intraclass correlation coefficient [ICC]=0.61, 95% confidence interval [CI; 0.53.0.70]) had better retest reliability than cognitive DTC metrics (ICC=0.27, 95% CI [0.17.0.36]). Larger retest reliability estimates were found for temporal gait outcomes (ICC=0.67-0.72) compared with spatial (ICC=0.34-0.53). Motor DTC metrics showed weak predictive validity for the incidence of future falls (r=0.14, 95% CI [-0.03.0.31]). Motor DTC metrics had weak concurrent validity with other clinical and performance assessments (r=0.11, 95% CI [0.07.0.16]), as did cognitive DTC metrics (r=0.19, 95% CI [0.08.0.30]). CONCLUSIONS: Gait-related temporal DTC metrics achieve adequate retest reliability, while predictive and concurrent validity of DTC needs to be improved before being used widely in clinical practice and other applied settings. Future research should ensure the reliability and validity of DTC outcomes before being used to assess dual-task interference.