Author Affiliation: Departments of Pediatrics and Anesthesia, Harvard Medical School and Children's Hospital, Boston, Mass.
The task of clearly
demonstrating that the type of care delivered in an intensive care unit
(ICU) ultimately affects patient outcomes is a challenge. It rarely is
feasible to manipulate already-established ICU characteristics.
Randomly assigning patients to ICUs that do or do not have certain
organizational features, such as ICU subspecialists performing daily
patient rounds or nurse-to-patient ratios of less than 1:2, is
logistically complex. Therefore, investigators who wish to explore the
relationship between patient outcome and the structure and function of
ICUs need to extract data from large databases that contain a
representative sample of ICUs, identify their organizational
characteristics, and use risk adjustment methods1 to
evaluate the association of these characteristics with the health
outcomes of interest. Risk adjustment ideally takes into account
underlying differences in the pre-ICU care patient population among
ICUs, the individual performance of each ICU, and differences in
post-ICU care for outcomes measured beyond the ICU stay.
Randolph AG. Reorganizing the Delivery of Intensive Care May Improve Patient Outcomes. JAMA. 1999;281(14):1330-1331. doi:10.1001/jama.281.14.1330