However, the difference in change of the PANSS total in the PP
population was 3.98 points (p = 0.0566). This is close to significant.
Other comparisons showed weaker but consistent trends in favor of risperidone
on efficacy. For example, CGI were near significant. In the FAS
sample, 18.7% of patients were normal or borderline ill on aripiprazole
at endpoint compared with 32.1% on risperidone. 21.6% of aripiprazole
patients were still markedly ill at the end of the study, vs. only 11.4% on
risperidone. These seem numerically important though p was 0.0839. In
the PP sample, 21.2% were very much improved on aripiprazole vs
35.3% on risperidone, and this time p = 0.0351. These data implicated
that there may be better efficacy with risperidone over aripiprazole for
overall improvement. However, a recent meta-analysis article found
that there were no significant differences on global state (n = 6381,
80 RCTs, low quality evidence) between aripiprazole and risperidone(Khanna et al., 2014). So it is necessary to design a further study with a
larger sample to demonstrate the conclusion.