Sample size
Power calculations are based on expected progress in use of communication strategies (the Communication Strategies Scale of the Hearing Handicap and Disability Inventory), which has been the primary outcome of previous studies in persons with hearing loss [41]. In a previous RCT by Kramer et al. (2005) on the effectiveness of a home education program for older adults with hearing impairment only, the mean difference in communication skills between the intervention and control group was about 0.5 (SD=0.8) [41]. Sample size calculations of this RCT are based on a linear mixed models and confirmed by some simulations (data not shown). Sixty-two participants per arm, with adjustment for clustering by 8 therapists in the intervention condition, provide a power of 0.80 (1-β) with alpha 0.05 (two-sided significance level), to detect a 0.5 difference between trial arms after 3 months (corrected for the differences at baseline), after taking into account a 20% dropout rate.