First, attributing causation is challenging because of the multiple variables that impact patient outcomes in ICUs beyond the Intensivist. Second, for patients in the entire cohort that had more than one Intensivist involved in their care, we attributed patient outcomes only to the admitting ICU attending physician. While we justified selleck this decision based upon the evidence that the first 24 to 48 hours of a patient’s care often sets the trajectory for both short and long-term patient outcomes [17-21], this did not allow us to account for management decisions made later in the ICU course by other Intensivists. We specifically performed the subgroup analysis to attempt to address this issue; however, since these patients represent less than half of the patients admitted to our ICUs, it is difficult to generalize the results.

A third limitation was our inability to control for the contribution of residents, ICU fellows and members of the multidisciplinary team to patient care; however, it has previously been shown that outcomes do not change depending on whether a critical care fellow is involved in patient care [22]. Additionally, the generalizability of our results may be limited because it was performed in three Canadian teaching hospitals and because the mix of base specialties was heavily weighted towards Internal Medicine and Pulmonary Medicine. Finally, given the size of our cohort, while some of our results are considered statistically significant, one could argue whether the differences should be considered clinically significant; the difference in the number of invasive procedures performed would be an example of this.

Taken in the context of these limitations, we believe that the results of our study still have potential implications for the way in which critical care medicine is administered. First, training programs need to be aware that a trainee’s base specialty may substantially influence the knowledge and skills with which they enter a fellowship program and hence their training needs and fellowship experience. Second, given predictions of a significant shortage of Intensivists in years to come [23,24], there may be pressure to preferentially recruit trainees from base specialties with a shorter AV-951 duration of training, for example Internal Medicine. While this could deliver more Intensivists to the workforce more quickly, prospective multicentre investigations are first warranted to further assess the potential impact of physicians’ base specialty of training on patient outcomes. Third, although care in most ICUs is provided by a highly trained multidisciplinary team with the assistance of evidence-based protocols, individual physician leadership clearly influences patient care.