We are pleased that our paper has started a discussion about the ethics of autonomy for severely mentally ill patients. In our view, this has been long overdue. Both authors replying to our editorial ^{Reference Lepping, Palmstierna and Raveesh1} have reasonable reflections, deepening our deliberations about the impact of any immediate reduction of autonomy on severely mentally ill patients and the balance with other ethical pillars that we all rely on in psychiatry.

Wilkinson raises the question of how attachment styles of the doctor could affect his or her communication style towards the patient, possibly increasing paternalism. It is an interesting point. It emphasises how paternalism could occur by the doctor being unaware of a ‘paternalistic’ communication and decision style. This is a relevant comment regarding how we as doctors interact with our patients, creating a more or less ‘coercive’ style.

Crichton, on the other hand, elaborates on the issue of how autonomy is in fact already restricted for patients. We acknowledge this aspect as relevant; however, we would equally like to stress that autonomy is not automatically more important than other ethical pillars. In our opinion, there is a danger in over-emphasising immediately expressed autonomy in every situation, as it risks compromising both future autonomy and other pillars of medical ethics. We merely discuss the balance between autonomy and the other central pillars of medical ethics in medicine, and particularly in psychiatry. Crichton's call for consideration of the already limited autonomy is justified, but this should be a starting point for a more detailed discussion. Patients may understand their situation and choices, but are not autonomous unless they are able to form value judgements about their reasons for choosing treatment. So stating that autonomy is limited is a judgement which needs to be carefully examined from an ethical point of view. In addition, autonomy will be interpreted differently in various social, religious, judicial, political, philosophical and medical contexts.

We are aware that autonomy is restricted for all of us by several components, and that action should be taken to increase it. But we would like to argue that, in order to increase patients' autonomy over time, there is a need to act upon all pillars of medical ethics. We argue that we should consider that the immediate choices expressed by the patient may occasionally have to be balanced with best interests decisions, both to preserve the integrity of the other pillars of medical ethics (providing safety, protection, treatment) and to promote future autonomy of the patient. Furthermore, we strongly believe that the immediate request for autonomous decision-making expressed by a severely disordered patient should not be a simple excuse to neglect other ethical considerations, just because it is the easiest way to proceed. In our view, this would be a dangerous road to follow, although anecdotal evidence suggests that it is already occasionally happening. It demands nothing of us as psychiatrists, but could have devastating consequences for patients in the end. It could undermine all the ethical pillars we regard as important, not only for the well-being of the patient but also for the patient's future ability to make true autonomous decisions about his or her life. We argue that taking a stand to evaluate all the ethical pillars of medical ethics is the right way to go, but it is also a demanding way along ‘a long and winding road’.