Date of Award

January 2012

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Mark D. Siegel

Subject Area(s)

Medicine, Ethics


Advance care decision making and standard code status discussions for hospitalized patients have several shortcomings, including a failure to address patient preferences regarding undesired outcomes following CPR. The goal of this study was to characterize the preferences of inpatients regarding code status and withdrawal of life support using a questionnaire based on a model of in-the-moment decision making. Specifically, we sought to assess patient baseline knowledge of interventions involved in CPR and possible outcomes, to determine the extent to which patients communicate their preferences with physicians and possible surrogates, to characterize patient preferences regarding initiation of CPR in light of general prognostic information, and to characterize patient preferences regarding continuation of life support in the event of cognitive impairment. We also sought to characterize demographic trends associated with particular preferences. One hundred two inpatients were enrolled in the study from January to March 2011. Face to face interviews and the electronic medical record were used to gather demographic information and determine subject preferences. Consistent with prior work, we confirmed that subjects had a limited understanding of terminology and outcomes associated with CPR. We found that a minority (38%) of subjects had been asked about code status and that only 50% of subjects with an advance directive were sure that the team was aware of it. We found that preferences regarding the discontinuation of life support in the event of cognitive impairment was heterogeneous. Forty-two percent of subjects were willing to accept a significant degree of cognitive impairment while 8% of subjects wanted to continue life support even with an outcome of permanent unconsciousness. In prior studies, some subjects reversed code status decisions after learning about prognosis, however in our study only one subject did so. We hypothesize that this occurred because we included a discussion of possible actions should a patient experience an unacceptable outcome following CPR. We also identified a small population of subjects who did not accept the provided general prognostic data. This small study suggests that detailed discussions of advance care preferences may be valuable in the inpatient setting, with an in-the-moment model likely to be useful in eliciting patient goals and values. The impact that detailed advance care discussion in the inpatient setting may have on adherence to patient wishes warrants further investigation.


This thesis is restricted to Yale network users only. This thesis is permanently embargoed from public release.