This study was not designed to examine fall-related
injuries as an endpoint. Nevertheless, the figures are listed
in Table 2. Three patients in the intervention group had
hip fractures and one a fracture of the forearm, and two
patients in the control group had a hip fracture and one a
vertebral fracture. Ten subjects in the intervention and 13
subjects in the control group reported that at least one of
their falls resulted in an injury. The reported rate of out-
door falls was too low to make statistical comparisons (9
falls away from home in the intervention group and 11 in
the control group).
The intervention program had a different effect on
subjects with and without a history of falls (see Table 3).
The proportion of frequent fallers among those with a his-
tory of frequent falls could be reduced but not among
those who did not report frequent falls in the past year
(Breslow-Day test for interaction; P .009).
27 In this sub-
group of subjects with a history of frequent falls, there was
also a 37% lower fall rate for the intervention group than
for the control group (IRR 0.63, 95% CI 0.43–0.94;
P .028).