Determination of the reference value and systematic bias of the functional reach test in Japanese elderly people by meta-analysis

Background/PurposeThe functional reach test (FRT), which was developed as a clinical balance assessment tool, has been widely used as a fall risk assessment tool in elderly people. The aim of the present study was to investigate the reference value and the presence of systematic bias in the FRT using the methodology of meta-analysis in community-dwelling elderly people.MethodsRelevant research articles were sought from electronic databases: MEDLINE, EMBASE, Cumulative Index to Nursing and Allied Health Literature (CINAHL) and Igakucyuouzasshi. The search was conducted from January 1990 to August 2011, and the terms “functional reach” and “elderly” were used in combination in the search. The searches were limited to peer-reviewed research articles involving Japanese elderly people with good functioning, aged 60 years and older. Weighted means were calculated for the reference value of FRT by a fixed effect model and a random effect model. Furthermore, weighted least squares regression was performed to determine the presence of systematic bias in the reference value of FRT.ResultsA total of 19 articles fulfilled the inclusion criteria, including 4274 participants whose mean age ranged from 69.0 to 81.4 years. The reference value of FRT was 29.44 cm (95% confidence interval: 27.60–31.27 cm) using the random-effect model, since the reference value using the fixed-effect model was found to have significant heterogeneity. Furthermore, multivariate weighted least squares regression was performed, and sex, age, height, and measurement method (one-arm or two-arm reach) were all independently associated with the FRT value (multiple R2 = 0.295, χ2 = 76.6, p < 0.001).ConclusionsSince participants' characteristics (sex, age, and height) and measurement method are probably related to systematic error in the FRT, judgment of physical function in elderly people using only the reference value determined in this study may have limitations.