Benefit assessments according to Sect. 35a of the Social Code Book V are frequently divided into subsets of populations. This division of therapeutic indications is evident in almost every other of the completed appraisals by the Federal Joint Committee (G-BA). The subsets can be deduced from the wording of the approved therapeutic indication (subpopulations) or from potential effect modification (subgroups). Methodological challenges are illustrated on the basis of current scientific knowledge of how to tackle subgroups. Practical assessment examples are used to explain the methods implemented by IQWiG and the G-BA and reveal major problem areas concerning subgroup analyses, subpopulations as well as mismatches between the requirements of early benefit assessment and approval. Overall, it appears there is a need for establishing a more comprehensive discussion regarding the validity of subgroup analyses and their associated limitations.