Recent failures of newly designed, poorly evaluated and widely released implants have become a major concern demanding changes in the decision-making process of implant approval and implementation. Moreover, a systematic review summarizing the current state of research on comparative studies of total hip arthroplasties (THA) concludes that (1) there is a lack of studies and evidence on the comparative effectiveness of THAs and (2) the overall quality of reporting in the few available studies is less than adequate. In this context, the objective of this work was to highlight the importance of a hospital-based registry for comparative effectiveness research in the field of THA. The work includes (1) an overview of different types of registries in THA; (2) an assessment of methodological particularities in registry (observational) research; (3) an illustration of the practical usefulness of a registry by research conducted in the Geneva University Hospitals Hip Arthroplasty Registry; and (4) a reflection on future development needs. The research conducted has been useful in identifying both well-functioning and unsafe implants, in assessing ways to decrease the risk of prosthesis dislocation, in advancing the epidemiology of less common causes of arthroplasty infections, and in elucidating the influence of obesity, very old age, and intensive physical activity on THA outcomes. It has been innovative by combining classical orthopaedic research with other disciplines such as patient education, rheumatology, and obesity research; in systematically evaluating a variety of outcomes including those that are patient-reported; and in using advanced epidemiological and statistical methods. As for the future, we need a more widespread implementation of registries of all sizes focusing on a combination of clinically relevant data, patient-centered outcomes and cost evaluation. We urgently need to work on methodological improvement and innovation, and to adopt valuable approaches from other fields. The data gathered have to efficiently assess implant safety, be useful and accessible for the orthopaedic surgeon, and relevant to the patient and the society.