A patients’ decision regarding “Do Not Resuscitate” (DNR) code status, which is a legal order to withhold cardiopulmonary resuscitation (CPR) or advanced cardiac life support in respect of a patient’s wish in case of a cardiac arrest or respiratory failure, has important medical and socioeconomic consequences. Yet, hospitalized patients often have insufficient information about expected outcomes from resuscitation measures. Research shows that <20% of in-hospital patients, who require CPR survive and subsequently many have permanent brain damage or disability. Although challenging in clinical routine, sharing information about expected prognosis with patients is a prerequisite for informed decision-making. In a recent hospital-based survey we found that DNR discussions are often insufficient with physicians not actively involving patients in the decision-making process, but relying decisions on the presumed preferences without in-depth discussions of patients‘ choices and consequences. General aim: This trial is designed to investigate whether checklist-guided shared decision-making including decision aids and communication of expected outcome influences patients’ decision regarding DNR code status, and at the same time, improves decision-making quality as judged by patient’s decisional comfort, patient knowledge and involvement in decision-making and patient satisfaction. Patient population: Consecutive adult medical patients admitted for in-hospital care, independent of medical diagnosis are eligible. Patients unable to complete questionnaires or unable to follow code status discussions (e.g., due to cognitive impairment such as dementia or delirium) are excluded. Methods: multicenter, cluster randomized controlled trial involving five Swiss university hospitals.Intervention: Medical residents conducting code status discussions will be randomized by an electronic, web-based system to the intervention group with shared decision-making facilitated by a checklist and a decision aid or to the usual care group. The decision aid was developed during a consensus conference, in which feedback from patients and clinicians was sought and integrated, and later field-tested until thematic saturation was achieved. Intervention group residents will receive detailed instructions about the shared decision-making checklist with specific communication teachings. To reduce bias, control group residents will also receive information about the importance of code status discussions and a general communication training, but no teaching regarding shared decision-making and no checklist will be used.Endpoints: The primary outcome is the frequency of DNR code status among patients per resident (i.e., patients choosing that CPR and intubation measures should not be performed in case of acute deterioration). Our key secondary endpoint is the quality of decision-making as judged by patient’s decisional comfort assessed through the validated German translation of the Decision Conflict Scale. Additionally, we will investigate patient`s knowledge about resuscitation measures and expected outcome, patients’ involvement in the shared decision-making process assessed through the validated German translation of the SDM-q-9 questionnaire, patients’ concerns and fears and overall satisfaction with the code status discussion, as well as physicians perceived comfort with patient’s choice and satisfaction with code status discussion. At hospital discharge, we will assess length of hospital stay, ICU-admissions, in-hospital resuscitations, code status violations and in-hospital deaths through medical chart review. Sample size and statistics: Based on pilot data in this patient population, we expect that 30% of patients in the usual care control will have a DNR code status, which will increase to 45% in the intervention group. We plan to include 174 residents (cluster) with a mean of 5 eligible patients per week over a 3-week period (total of 15 patients per cluster) in the five participating centers over a total study time of 2 years. Based on these assumptions, we aim to include 2610 patients among 174 residents, which will give this study an 80% power at a 0.05 alpha error with an inter-cluster correlation of 0.5.Discussion: Although code status discussions are a cornerstone of patient-centered care, there is a lack of trials investigating the most appropriate approach for physicians to communicate with patients about their preference. This multicenter randomized-controlled trial will close this important knowledge gap by systematically investigating effects of checklist-guided shared decision-making on patients’ decisions regarding DNR status and the quality of decision-making.