Medical education has a long tradition of using various patient representations in teaching and assessment. With this literature review we aim, first, to provide an overview of the most important patient representations used to teach and assess clinical skills, considering in particular “summative exams” that have a pass or fail outcome; second, to provide arguments for choosing certain patient representations; and third, to show the advantages and limitations of different patient representations, especially simulated patients (SPs) and real patients (RPs). Typical patient representations include case narratives, anatomical models, simulators and mannequins, as well as SPs and RPs. The literature indicates that there are multiple ways of using various patient representations in teaching and that the intended didactical purpose informs the choice of representation. Early in the educational programme, even low-fidelity patient representations can be a good fit for assessment purposes if chosen to match the educational level. The use of RPs in summative, high-stakes assessments (exams with particularly important consequences for the examinee) is limited for methodological and ethical reasons. The methodological implementation of summative exams also entails specific challenges, such as ensuring measurement reliability and fairness towards the examinees. Carefully prepared, SPs can perform their roles with a sufficient degree of authenticity, making summative exams more manageable, and imposing no strain or risk on RPs. The ongoing debate concerning the use of SPs and RPs in summative assessment highlights perceived limitations of SPs in relation to RPs that are often not supported by research. Evidence shows that SPs, in combination with additional simulation modalities as needed, represent the first choice for summative clinical assessment. We also consider the strengths and limitations of this review and reflect on the applicability of our findings. We conclude that in order to select the right patient representations in clinical teaching and/or assessment, a number of perspectives must be considered: (i) the learning goals, aligned with the stage of study, (ii) the corresponding requirements of the clinical task itself (e.g., performing a phlebotomy or a communication task), (iii) the level of authenticity required and (iv) the resources needed, taking patient safety and feasibility into consideration.