Data pooling comes up routinely in the device trials. Sometimes questions of whether pooling is appropriate arise, and such questions can be referred to as those of poolability. As a result of the generality of the term, poolability comprises a heterogeneous set of problems that need to be addressed on a case by case basis, depending, among other things, on whether the groups in question are investigational sites, geographical regions, clinical studies (or substudies), patient populations, device models, and so on. FDA and the industry will share their perspectives on the following issues:

(1) How to treat sites in the model: fixed or random effects? Why?

(2) Usually, the site-by-treatment interaction tests are underpowered. Would graphical analysis of site by treatment interaction be more appropriate than a formal statistical test of the interaction hypothesis?

(3) If site heterogeneity depends on the site-specific covariates, how do we model it?

(4) In what circumstances, pooling of feasibility and pivotal studies may be allowed?

(5) If pooling across sites is not justified, what can be done?