We estimated the required sample size needed for an alpha level of
0.05 (two-tailed) on the basis of data from similar trials.
According
to this estimate, a sample of 200 persons in each group was sufficient
(with a statistical power of 0.8) to detect a 40 percent reduction in the
number of persons with disability and a 25 percent change in the
number of acute hospital admissions, with a marginal ability (statistical power of 0.5) to detect a 50 percent reduction in nursing home admissions.