Clinical trials usually enroll a general patient population expected to benefit from the investigational therapy. However, the benefit might vary by the subject’s background characteristics. There has been growing interest in designing trials with assessment of a targeted subgroup as a primary objective of the trial. However, as observed treatment effect in subgroups might be due to chance, there are challenges in interpreting completed study findings. This presentation discusses issues related to subgroup analysis including chance findings and testing for treatment by subgroup interaction. For trials designed for a targeted subgroup, we discuss issues related to identification of the targeted subgroup and impact of errors in the subgroup identifier on subgroup findings, power considerations and subgroup enrichment, multiplicity considerations, and the intended population for labeling.