Post-hoc comparisons showed these significant
improvements existed in the exercise and
education compared with the usual physical therapy
group for each variable (SF-36 bodily pain,
P = 0.029; SF-36 physical function, P = 0.034),
except for SF-36 role physical (P = 0.125). No significant
differences in postintervention scores were
found between the exercise and education and education
only groups, or between the education only
and usual physical therapy groups (all, P > 0.05).