24-Hour Intensivist Staffing Is Not Beneficial for Patients
In 1988, the first study to evaluate the role of intensivists in patient outcomes was published (2). The investigators performed a before-after study in an academic medical ICU that changed attending physician staffing models, from internists of different subspecialties with additional non-ICU clinical responsibilities to critical care specialists dedicated to the ICU. They found that mortality was decreased with the intensivist model across all categories of patient age and severity of illness. Then followed several more studies that suggested that “high-intensity” ICU staffing—that is, where intensivists assume primary responsibility or provide mandatory consultation for all ICU patients—had better patient outcomes (3), leading to a movement toward greater intensivist presence in ICUs (4, 5). Meanwhile, a separate but related literature was growing to suggest that critically ill patients fared more poorly when admitted during weekends and nighttime hours (6)—times when intensivists typically were not immediately available. Therefore, it was a reasonable leap to consider that 24-hour intensivist staffing may further improve patient outcomes.
The earliest study to examine the relationship of 24- intensivist staffing was a before-after study in an academic medical ICU that underwent a transition in staffing in 2006 from remote, on-demand availability of intensivists to on-site intensivists 24 hours per day (7). Investigators evaluated several patient outcomes: mortality, adherence to evidence-based practices, ICU complications, and patient and family satisfaction. In a study population of over 4,000 patients, they found small improvements in processes of care (primarily improvements in administration of prophylaxis for stress ulcers and venous thromboembolism) without any differences in mortality or patient and family satisfaction, after adjustment for patient factors. Thereafter, several more before-after studies in single or small numbers of ICUs were performed with inconsistent results. Two found decreased mortality with 24-hour intensivist staffing but both had other organizational differences between the study groups (8, 9); two found increased mortality (10, 11) and four found no association (12–15).
Two large multicenter, retrospective cohort studies attempted to overcome the biases of these earlier studies. Both used sophisticated statistical methods to account for both patient- and ICU-level confounding. The first study found no association between nighttime intensivist staffing with improved patient mortality among all ICUs. But, it created a buzz in the critical care community because it did find an association in a prespecified subgroup analysis of ICUs without high-intensity intensivist staffing during daytime hours (16). Said another way, this study suggested a benefit to adding nighttime intensivist staffing to ICUs without readily available daytime intensivists. It specifically did “not” demonstrate a benefit to “around-the-clock” intensivist staffing. We performed a subsequent study, using a larger database with a higher percentage of community hospitals, and found no such association (17). Two additional multicenter international studies focused on ICU organizational characteristics more broadly also found no association between nighttime or 24-hour intensivist staffing with mortality (18, 19).