Morning interprofessional rounds (MIRs) are used in critical care medicine to improve team-based care and patient outcomes. Given existing evidence of conflict between and dissatisfaction among rounds participants, this study sought to better understand how the operational realities of care delivery in the intensive care unit (ICU) impact the success of MIRs. We conducted a year-long comparative ethnographic study of interprofessional collaboration and patient and family involvement in four ICUs in tertiary academic hospitals in two American cities. The study included 576 h of observation of team interactions, 47 shadowing sessions and 40 clinician interviews. In line with best practices in ethnographic research, data collection and analysis were done iteratively using the constant comparative method. Member check was conducted regularly throughout the project. MIRs were implemented on all units with the explicit goals of improving team-based and patient-centered care. Operational conditions on the units, despite interprofessional commitment and engagement, appeared to thwart ICU teams from achieving these goals. Specifically, time constraints, struggles over space, and conflicts between MIRs' educational and care-plan-development functions all prevented teams from achieving collaboration and patient-involvement. Moreover, physicians' de facto control of rounds often meant that they resembled medical rounds (their historical predecessors), and sidelined other providers' contributions. This study suggests that the MIRs model, as presently practiced, might not be well suited to the provision of team-based, patient-centered care. In the interest of interprofessional collaboration, of the optimization of clinicians' time, of high-quality medical education and of patient-centered care, further research on interprofessional rounds models is needed.