ABSTRACT
There are extensive discussions about the relative merits of performing
individual-patient level data (IPD) analysis versus meta-analysis (MA) in
those cases where IPD are accessible. Some favor IPD and others favor MA.
In this chapter we use placebo controlled clinical studies of lamotrigine in the
treatment of bipolar depression to illustrate the pros and cons of IPD and MA.
Two clinical outcome measures, the Hamilton Rating Scale for Depression
(HAMD) and the Montgomery-Asberg Depression Rating Scale (MADRS),
are used. This chapter is organized as follows. In Section 8.1 we introduce
this data as well as the descriptive statistics for the data. We then present
data analysis from both IPD and MA on treatment comparison for changes in
HAMD in Section 8.2 and changes in MADRS in Section 8.3 - both of which
led to the same conclusions as summarized in Section 8.4. Based on these
conclusions, we then formulate a simulation to compare the efficiency of IPD
to MA in Section 8.5 followed by discussion in Section 8.6.