tabelas_results_responsiveness.xlsx

Introduction: Mobility significantly depends on the strength in the ankle muscles, which are particularly relevant in mobility. There are few assessment tools of ankle strength validated for all psychometric properties. The main purpose of this study was to test the responsiveness of Calf-Raise Senior Test, in order to detecting improvements in plantar-flexor strength/power of elderly participants after a 24 weeks exercise intervention. Methods: 82 older adults participated in an exercise program and were assessed with CRS Test and 30-seconds chair stand test at baseline and at follow-up. Effect size, standardized response mean and minimal detectable change measures of changes were determined using Calf-Raise Senior and chair stand tests scores. ROC curves analysis was used to define a cut-off representing the minimally important difference of Calf-Raise Senior test.Descriptive statistics analyses were performed to allow the sample characterization. Central tendency parameters were determined for continuous variables (mean, standard deviation and median) and relative frequency was calculated for categorical and ordinal variables. The normal distribution of continuous variables was checked with the Kolmogorov-Smirnov Test. Differences in scores between groups of change (stables vs improved) at BL and FU were tested by the Independent samples t-tests (or non-parametric Mann-Whitney U test when appropriate), and Paired t-tests (or non-parametric Wilcoxon tests) were used to compare data at BL and FU within groups.The responsiveness of the CRS test was determined using two different methods of sensitivity to change analysis aiming at evaluating the capacity of this tool to identify important and meaningful changes occurred in the plantar-flexor’s strength of participants after the intervention. A distribution-based approach was employed to assess the magnitude of changes in the scores, using statistical parameters, which are related to sample variability and measurement precision (Crosby, Kolotkin, & Williams, 2003; Lauridsen, Hartvigsen, Manniche, Korsholm, & Grunnet-Nilsson, 2006; Middel & Van Sonderen, 2002; Revicki, Hays, Cella, & Sloan, 2008). In order to analyze the interpretability of the results (i.e., to determine what would be the qualitative meaning of the CRS test quantitative scores), an anchor-based analysis was also applied in this study (Terwee et al., 2007; Terwee et al., 2010). These approaches have advantages and limitations, and some authors advise both, in order to provide a most complete and comprehensive information of changes (Crosby et al., 2003; Lauridsen et al., 2006; Revicki et al., 2008).The change scores of CRS and CS tests were used in both methods of analysis, and calculated from the results of the two assessment phases (FU scores - BL scores). The distribution-based analysis employed the following statistical parameters and calculations: 1) Effect size (ES = Mean Change Score / SD of BL Scores) - provides information about the magnitude of change over time in before-after study period (Cohen, 2013; van Iersel, Munneke, Esselink, Benraad, & Rikkert, 2008). To interpret the effect-size data, the cutoff points proposed by Hopkins (2000) (ES < 0.20 = trivial effect; 0.20 ≥ ES < 0.60 = small effect; 0.60 ≥ ES < 1.20 = moderate effect; 1.20 ≥ ES < 2.0 = large effect; 2.0 ≥ ES < 4.0 = very large; and ES ≥ 4.0 = nearly perfect); 2) Standardized response mean (SRM = Mean Change Score / SD of the Change Score) - indicates if the change was large relatively to the variability in the measurements. SRM values of 0.20, 0.50, and 0.80 are considered as small, moderate and large change, respectively (Crosby et al., 2003); 3) Minimal Detectable Change (MDC = 1.96 (√2 x SEM)) - reflects the smallest change in score that can be interpreted as a ‘true’ change (DeVet et al., 2006). An ICC of 0.90 (95% CI = 0.82-0.95), determined in a previous test-retest reliability study (André et al., 2016), was used for calculation the Standard Error of Measure (SEM = SD of BL scores (√1-ICC)) of CRS.The proportion of participants achieving a degree of improvement that was beyond the MDC was then determined in order to verify if the changes were important to be considered as “true changes” (Haley & Fragala-Pinkham, 2006).The anchor-based approach was performed using a Receiver Operating Characteristic (ROC) curves analysis in order to verify whether the CRS test would be able to discriminate between groups of participants with positive change (improved) versus no change (stable) (de Yébenes Prous, Salvanés, & Ortells, 2008; Lauridsen et al., 2006). The score change (SC) of CS test was used to dichotomize sample considering a cut-off of 3.01 (ICC = 0.92) as the minimal detectable change (MDC) determined in a previous test-retest reliability for this study. An area under the curve (AUC) above 0.5 for specificity and sensitivity (Deyo, Diehr, & Patrick, 1991) was considered, and the cut-off corresponding to the point closer to the upper left corner was defined as the score that best classifies participants who had improved or maintained their state. This cut-off represents the “minimally important difference” (MID) of this test (DeVet et al., 2006), that is, the smallest change in the CRS scores considered as clinically relevant, or worthwhile to the participants (Haley & Fragala-Pinkham, 2006), also frequently called in literature as “minimal clinically important difference” (MCID) (Lauridsen et al., 2006; Terwee et al., 2010). The Spearman rank correlation coefficient was also used to verify the association between change scores on CRS and CS, assuming that these variables were not normally distributed. A correlation of ≥ 0.40 was considered as minimum important change, according to Akoglu (2018). Results: Results revealed a small (ES = 0.42) to moderate (SRM = 0.51) sensitivity to change in plantar-flexion strength and power across time, which was lower than that of chair stand test (ES = 0.64, SRM = 0.67). The sensitivity to change of Calf-Raise Senior test was more evident in groups of subjects with lower initial scores. Results revealed a Calf-Raise Senior cutoff score of 3.5 repetitions for minimally important difference and 4.6 for minimal detectable change as change scores related to beneficial changes in ankle strength and power after an exercise intervention. Conclusion: Calf-Raise Senior Test can be considered an excellent test to assess elderly ankle function in intervention studies. The cutoff scores of minimal detectable change and minimally important difference presented in this study can be useful in determining the success of interventions aiming at improving mobility in senior participants.