ABSTRACT
Most of the published papers on assessment of consistency of treatment effect in multiregional clinical trials (MRCTs) are focused on a few predefined regions. Usually, countries in an MRCT are grouped into regions based on geographical as well as intrinsic/extrinsic factors (Tanaka et al., 2011), and the number of regions is suggested to be small in order to have a reasonable probability to detect a true heterogeneity (Chen et al., 2011). Countrylevel analysis is known to have many limitations because data from each individual country are generally limited and the large number of countries can increase the chance of false-positive findings due to multiplicity (Gallo et al., 2011). However, there is strong interest from the local regulatory agencies in assessing consistency of treatment effect to ensure that the overall conclusions can apply to their own judiciary. Such interest is illustrated by the recent debate on the US findings from the Metoprolol CR/XL Randomized Intervention Trial in Congestive Heart Failure (MERIT-HF) (Wedel et al., 2001) and Platelet Inhibition and Patient Outcomes (PLATO) (Wallentin et al., 2009) trials. In fact, country-specific analyses are requested by health authorities, including Japan Pharmaceuticals and Medical Devices Agency (2007) and U.S. Food and Drug Administration (FDA) (2011). In the Japanese product
CONTENTS
17.1 Introduction .............................................................................................. 215 17.2 Probabilistic Assessment and Expected Results ................................. 216 17.3 Graphic Tools ............................................................................................ 220
17.3.1 Funnel/Galbraith Plot ................................................................ 220 17.3.2 Quantile-Quantile Plot ............................................................. 221
17.4 Country-Specific Estimate of Treatment Effect ....................................222 17.5 Concluding Remarks ...............................................................................223 References ............................................................................................................. 224
label, data from the Japanese cohort are usually analyzed and reported separately from the rest of the world. Other local health authorities, including Korea FDA, require ethnicity sensitivity report that compares data from local to the global data. This section is to address country-specific analysis due to the fact that some local health authorities will be looking at their own countries. In some places, the terms “region” and “country” are exchangeable if there are a large number of regions/countries.
