Aligning patient values and code status: Choice of Diction's Effect (CODE) study

Decisions regarding resuscitation after cardiac arrest are critical from ethical, patient satisfaction, outcome, and healthcare cost standpoints. Physician-reported discussion barriers include topic discomfort, fear of time commitment, and difficulty articulating end-of-life concepts. The influence of language used in these discussions has not been tested. This study explored whether utilizing the alternate term "allow (a) natural death" changed code status decisions in hospitalized patients versus "do not resuscitate" (DNR).

All patients age 65 and over admitted to a general medicine hospital teaching service were screened (English-speaking, not ICU-level care, no active psychiatric illness, no substance misuse, no active DNR). Participants were randomized to resuscitation discussions with either DNR or "allow natural death" as the "no code" phrasing. Outcomes included patient resuscitation decision, satisfaction with and duration of the conversation, and decision correlation with illness severity and predicted resuscitation success.

102 participants were randomized to the "allow natural death" (N = 49) or DNR (N = 53) arms. The overall "no code" rate for our sample of hospitalized general medicine inpatients age >65 was 16.7%, with 13% in the DNR and 20.4% in the "allow natural death" arms (p = 0.35). Discussion length was similar in the DNR and "allow natural death" arms (3.9 + 3.2 vs. 4.9 + 3.9 minutes), and not significantly different (p = 0.53). Over 90% of participants were highly satisfied with their code status decision, without difference between arms (p = 0.49).

Participants' code status discussions did not differ in "no code" rate between "allow natural death" and DNR arms but were short in length and had high patient satisfaction. Previously reported code status discussion barriers were not encountered. It is appropriate to screen code status in all hospitalized patients regardless of phrasing used.