The issue of coercion has been one of the major challenges in psychiatry since its earliest days. Coercion is a complex phenomenon, encompassing a wide range of practices, all of which have in common that they conflict with patients’ autonomy. There are different dimensions to restraint, which need to be considered within a global model. It is therefore possible to distinguish at least three central dimensions: coercive measures in the strict sense of the term, which include measures restricting freedom of movement and involuntary pharmacological treatment; informal coercion, which consists of the use of pressures and incentives to make patients accept therapeutic proposals; and subjective or perceived coercion, which encompasses all coercive experiences of patients in the psychiatric system. This thesis is structured around five original articles on three areas related to coercive measures: the negative consequences of coercion, risk factors for coercive measures, and interventions to reduce their use and consequences. The first article looks at the effects of coercion on the severity of psychiatric symptoms during hospitalisation, showing that the use of coercive measures is likely to worsen patients' mental state. The second and third articles are devoted to the study of risk factors for coercive measures in general adult psychiatric populations admitted via emergency departments, and in psycho-geriatric populations. Regarding emergency admissions, the results show that police intervention, aggressive behaviour and involuntary admission are all risk factors for the occurrence of coercive measures during hospitalisation. These measures mainly take place during the first 24 hours following admission. As for the elderly population, the study shows the determining role of cognitive disorders and associated aggressive behaviour in the use of coercive measures. Finally, the fourth and fifth articles present the results of a randomised clinical trial studying the effects of a debriefing intervention for coercive measures. They show that this intervention is effective in limiting the onset of post-traumatic stress symptoms and the level of perceived coercion. Future research should focus on better defining informal coercive practices, studying the effects of new models of hospital care on the use of coercive measures, and finally on the use of coercion in outpatient setting, as well as in other areas of medicine.