Rationale: Despite multiple trials of interventions to improve end-of-life care of the critically ill, there is a persistent lack of understanding of factors associated with barriers to decision-making at the end of life.
Objective: To apply the principles of complexity science in examining the extent to which transitions to end-of-life care can be predicted by physician, family, or patient characteristics; outcome expectations; and the evaluation of treatment effectiveness.
Methods: A descriptive, longitudinal study was conducted in 3 adult intensive care units (ICUs). Two hundred sixty-four family surrogates of patients lacking decisional capacity and the physicians caring for the patients were interviewed every 5 days until ICU discharge or patient death.
Measurements: Characteristics of patients, physicians, and family members; values and preferences of physicians and family; and evaluation of treatment effectiveness, expectations for patient outcomes, and relative priorities in treatment (comfort vs survival). The primary outcome, focus of care, was categorized as (1) maintaining a survival orientation (no treatment limitations), (2) transitioning to a stronger palliative focus (eg, some treatment limitations), or (3) transitioning to an explicit end-of-life, comfort-oriented care plan.
Main results: Physician expectations for survival and future cognitive status were the only variables consistently and significantly related to the focus of care. Neither physician or family evaluations of treatment effectiveness nor what was most important to physicians or family members was influential.
Conclusion: Lack of influence of family and physician views, in comparison to the consistent effect of survival probabilities, suggests barriers to incorporation of individual values in treatment decisions.
Keywords: complexity science; critically ill; end of life.