The quality of mercy. Caring for patients with 'do not resuscitate' orders

JAMA. 1992 Feb 5;267(5):682-6. doi: 10.1001/jama.267.5.682.


Objective: To assess (1) the effect of an ethics education intervention for medical house officers on practices surrounding "Do Not Resuscitate" (DNR) orders and (2) the association of DNR care with patient diagnosis and demographic variables.

Design: A 1-year randomized, controlled trial.

Setting: An urban, university teaching hospital.

Participants: Eighty-eight internal medicine house officers.

Intervention: House officers were arbitrarily assigned to four "firms." One firm was randomized to an extensive ethics education intervention (EI), one to a limited intervention, and two served as controls.

Main outcome measures: Charts of patients with DNR orders were reviewed for compliance with the hospital's DNR policy, which instructs that when DNR orders are written there should be (1) an attending signature, (2) documentation of reasons, (3) appropriate consent, and (4) attention to 11 concurrent care concerns (CCCs) (eg, the appropriateness of intubation, tube feedings, hospice).

Results: Thirty-nine charts were reviewed before the intervention and 57 after. The number of CCCs per DNR order fell among patients cared for by controls (1.9 to 1.0, P less than .05) and rose among patients cared for by the EI group (0.9 to 3.8, P less than .05). Compliance with the DNR policy varied among patients with differing diagnoses. "Do Not Resuscitate" orders were signed less frequently (P = .01) for patients with the acquired immunodeficiency syndrome (AIDS) (65%) compared with patients who had other diagnoses (85%) or malignancy (91%). Similarly, appropriate consent was recorded for 59% of patients with AIDS, 83% of others, and 85% of those with malignancy (P less than .05). The number of CCCs per DNR was 0.7 for AIDS, 1.4 for others, and 2.4 for malignancy (P less than .05). In multivariate regression analysis, house officer ethics education and patient diagnosis, but not patient gender, age, race, or insurance status, were predictors of the number of CCCs per DNR.

Conclusions: (1) An extensive ethics education intervention can improve care for DNR patients, especially with respect to CCCs. (2) In this setting, quality of care for DNR patients varied systematically with diagnosis. These results have implications for the design and implementation of ethics education programs.

Publication types

  • Clinical Trial
  • Randomized Controlled Trial
  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Baltimore
  • Curriculum
  • Ethics, Medical / education*
  • Female
  • Health Knowledge, Attitudes, Practice
  • Hospitals, University
  • Humans
  • Internal Medicine
  • Internship and Residency*
  • Linear Models
  • Male
  • Patient Selection*
  • Quality of Health Care
  • Resuscitation Orders*
  • Withholding Treatment