0 was considered very large (Batterham and Hopkins 2006). Fifty-eight people expressed an interest in participating in the study during the recruitment period, and 40 were included. All 40 participants (20 experimental and 20 control) completed the measurement and intervention AZD8055 chemical structure period (Figure 1). The baseline characteristics of the participants are presented in Table 2 and in the first two columns of data in Table 3. The groups were comparable with respect to their

demographic characteristics and their baseline values of the outcome measures. All experimental participants attended all balance training sessions and no participants in the control group attended any of the sessions. One participant from the experimental group became dizzy during training. The participant was checked by medical staff and found to have sustained no problems. The participant then completed the training session and continued with all other sessions. Complete data sets were obtained from all participants. Forskolin supplier Group data for all outcomes are presented in Table 3. Individual participant data are presented in Table 4 (see eAddenda

for Table 4). Fear of falling measured by the Falls Efficacy Scale International questionnaire improved 7 points (SD 7) in the experimental group but deteriorated by 1 point (SD 4) in the control group during the intervention period. The between-group difference in change in the Falls Efficacy Scale International questionnaire scores was a mean of 8 points (95% CI 4 to 12), which equated to a moderate effect size of 0.96. Dynamic balance improved by 2.1° (95% CI 1.3 to 3.0) more on the Falls Risk Test in the exercise group participants after the balance training than in the control group participants over the same period (Table 3, individual patient data in Table aminophylline 4). This equated to a moderate effect size of 0.86. The effect of the balance training on isometric strength in the knee is also presented in Table 3 (individual patient data in Table 4). The exercise group had substantial improvements while the control

group had minor deteriorations in strength. On average, the effect of the training was to increase knee flexor strength by 7 Nm (95% CI 3 to 11), which equated to a moderate effect size of 0.81. The increase in knee extensor strength of 7 Nm (95% CI 1 to 12) equated to a small effect size of 0.24. The regression analysis indicated that the initial Falls Efficacy Scale International and Falls Risk Test scores predicted improvements after training in fear of falling (Table 5). The regression model predicted 64% of the observed changes in the Falls Efficacy Scale International scores (Table 5). These improvements in fear of falling can also be explained (26%) by the improvement in dynamic balance after treatment (Table 6). Improvements in dynamic balance (29%) can be partly explained by the improvement in knee extensor isometric strength after treatment (Table 7).