death—and many patients died after long stays in intensive care units (ICUs) either comatose or with mechanical ventilation.

  • Survival time could be better predicted by a computerized model with appropriate data inputs than by an individual physician.

  • An unexpectedly large percentage of patients experienced substantial pain across all diagnoses.

  • The study participants were younger than anticipated (median age less than 65), which led to HELP, a companion study of patients more than 80 years of age.

The Phase II intervention employed a skilled nurse specialist to interact with patients and their families, staff, and the intervention physicians. Specifically,

  • physicians were given detailed prognostic information for each patient on survival, outcome if cardiopulmonary resuscitation was used, and prospect of severe disability;

  • nurse specialists talked to patients and families about their specific wishes regarding treatment and communicated that information to the physicians and nurses treating the patient; and

  • physicians were given written information regarding each patient’s wishes about treatment, including pain control and the use of technology-intensive measures (e.g., CPR).

All participating physicians also were given feedback on the overall results of the observational phase of the study, characterizing the shortcomings of physician-patient communication, pain, and the timing of DNR orders.

A form of “cluster randomization” (by physician specialty and study site) was used to assign patients to either the intervention or the usual-care groups (see SUPPORT Principal Investigators, 1995, for details). The evidence after enrollment of 4,804 patients in two years was examined for five outcomes:

  1. median time until the DNR order was written,

  2. agreement between patient and physician regarding the DNR order,

  3. number of days spent in an “undesirable state” (e.g., comatose, on mechanical ventilation, in ICUs),

  4. percentage of patients in substantial pain, and

  5. median resource use (in 1993 dollars).

None of the outcomes was better for patients in the intervention group than for those in the control group.

ently. This led RWJF to begin its Last Acts campaign, an effort to improve end-of-life care at the grassroots level that now has more than 400 members (Schroeder, 1999), and funding of demonstration programs to reduce the identified barriers to high-quality care for those who are dying.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement