Almost all pivotal medical device trials are conducted in multiple centers and/or regions. The device effectiveness and/or safety profiles might be different across centers/regions. Therefore, directly pooling the data together without evaluation of data poolability might lead to biased treatment effect estimate. In this presentation, we will discuss from a reviewer’s perspective the challenges and possible approaches of evaluation of the site/region poolability and briefly touch on what to do if the data turn out to be non-poolable.