The limits of decision analysis for rapid decision making in ICU nursing

In a study examining rapid decisions in ICU nursing in the context of decision analysis, 40 nurses in 2 ICUs were asked to describe their decisions for six pretested vignettes, suggest alternatives and answer questions about case characteristics. Later, ICU charge nurses and educators were asked to rate the nurses' responses using a modified Q-sort. The nurses did not perceive a finite set of alternatives; the 40 nurses gave 36 to 40 unique sequences that did not cluster. The experts could not agree on which alternatives were good (no "gold standard"). The assumptions of decision analysis appear to be violated when (a) there is not a small finite set of mutually exclusive alternatives; (b) the merit of an intervention cannot be agreed on; or (c) the link between intervention and outcome is unclear. Not only nursing but other process-oriented areas of medical care as well may not fit the decision analysis model.