Using Standard Deviation Of Individual Responses To Assess Exercise Treatment Response Heterogeneity In Postmenopausal Women

PURPOSE: Treatment response heterogeneity (TRH) to exercise training is widely assumed and often depicted using simple waterfall plots of individual pre-post change scores within exercise treatment arms. However, these plots do not account for exercise-independent sources of variation and are infrequently compared to no-exercise control groups. The standard deviation of individual responses (SDir) compares the variances of change scores between treatment and control groups and has been suggested as a more robust method for assessing TRH in randomized controlled trials. METHODS: Data were obtained from a 6-month randomized exercise trial in postmenopausal women with overweight or obesity. Study conditions were a no-exercise control group and 3 exercise groups w/ energy expenditures of 4, 8, and 12 kcal·kg−1·wk−1 (KKW). The study recruited 411 female participants (mean age 57.2 y, 63.5% Caucasian). This study used SDir to compare the interindividual variability of weight change, relative VO2, and absolute VO2 between the control and exercise groups. RESULTS: SDir analysis revealed no evidence of significant TRH for weight change and relative VO2 between the exercise and control groups (p > 0.05). In contrast, for absolute VO2, a negative SDir was observed in the 4 KKW (SDir -0.057, CI: -0.063, -0.051) & 8 KKW (SDir -0.019, CI: -0.026, -0.012) exercise groups, indicating greater variability in the control group. The 12 KKW group exhibited a positive SDir (SDir 0.065, CI: 0.057, 0.072), suggesting potential TRH with higher exercise volumes. CONCLUSIONS: The original study used a simple waterfall plot to show the distribution of pre-post weight change scores, but there were not any analyses on TRH. SDir is considered to be a reliable method for assessing TRH and this study has shown no statistically significant individual response variation for weight change and relative VO2 in the exercise and control groups. Interestingly, our analysis also demonstrated a negative SDir in absolute VO2 for the 4 and 8 KKW exercise groups suggesting greater variability in the control group compared to certain exercise groups. Further investigation regarding this is warranted.