Abstract
RATIONALE: Compared with their Full Code counterparts, patients with do not resuscitate/do not intubate (DNR/DNI) status receive fewer interventions and have higher mortality than predicted by clinical characteristics. OBJECTIVES: To assess whether internal medicine residents, the front-line providers for many hospitalized patients, would manage hypothetical patients differently based on code status. We hypothesized respondents would be less likely to provide a variety of interventions to DNR/DNI patients than to Full Code patients. METHODS: Cross-sectional, randomized survey of U.S. internal medicine residents. We created two versions of an internet survey, each containing four clinical vignettes followed by questions regarding possible interventions; the versions were identical except for varying code status of the vignettes. Residency programs were randomly allocated between the two versions. RESULTS: Five hundred thirty-three residents responded to the survey. As determined by Chi-squared and Fisher's exact test, decisions to intubate or perform cardiopulmonary resuscitation were largely dictated by patient code status (>94% if Full Code, <5% if DNR/DNI; P < 0.0001 for all scenarios). Resident proclivity to deliver noninvasive interventions (e.g., blood cultures, medications, imaging) was uniformly high (>90%) and unaffected by code status. However, decisions to pursue other aggressive or invasive options (e.g., dialysis, bronchoscopy, surgical consultation, transfer to intensive care unit) differed significantly based on code status in most vignettes. CONCLUSIONS: Residents appear to assume that patients who would refuse cardiopulmonary resuscitation would prefer not to receive other interventions. Without explicit clarification of the patient's goals of care, potentially beneficial care may be withheld against the patient's wishes.