718

Abstract

Introduction: There is limited literature regarding the specific question of how frequently patients’ Do Not Resuscitate (DNR) wishes are not followed. An electronic provider order entry (POE) system may contain prompts for documentation of code status on admission, but documentation in the paper record may coexist. As hospitals move toward completely electronic medical records, the patient’s advanced directives must be faithfully transferred to the electronic system. We measured how often the electronic and paper records were discordant in a major teaching hospital that is migrating to a fully electronic system. Hypothesis: Discrepancies in code status documentation will exist across various components of the inpatient medical record. Methods: We conducted a single-day cross-sectional study of 227 patients in the medical intensive care unit, medical, oncology, and neurology wards at an academic tertiary care center. From the paper medical record, the online POE, and the electronic sign-out, we recorded code status as Full Code, DNR/DNI, comfort-measures only (CMO), and “other” to capture variations on limitations of life-sustaining treatment (LLST). We reviewed the three most recent days of physician and nursing notes for an updated code status. If it was not mentioned in these notes, we used the admission note at the beginning of the medical record. We defined a significant discrepancy as an instance in which a patient was listed as having a LLST in one source, but Full Code or not recorded in another. The primary outcome was the frequency of significant discrepancies in code status documentation. Results: Significant discrepancies were found in 11 of 227 patients, representing a prevalence of 4.85%. These discrepancies occurred on all services, although this study was not powered to detect significant differences across services. There was not any association between the presence of a discrepancy and hospital day or age (HD 9.33 +/- 5.02 vs. 8.13 +/- 10.39 days, p=0.48 and age 71.09 +/- 17.66 vs. 61.91 +/- 16.51 years, p=0.33, respectively). Patients for whom the POE order reflected a “presumed full code” status had a shorter LOS that those for whom the order reflected a confirmed code status discussion (5.85 +/- 5.77 vs. 10.86 +/- 13.19, p=0.0005). Conclusions: Among the medical population at an academic tertiary care center with an electronic POE system and paper medical chart, discrepancies exist in the documentation of patients’ code status that could lead to inappropriate resuscitation or withholding of resuscitation.