November 27, 1995

Promoting Inpatient Directives About Life-Sustaining Treatments in a Community HospitalResults of a 3-Year Time-Series Intervention Trial

Author Affiliations

From the Department of Medicine, St Mary's Hospital, University of Rochester (NY) School of Medicine and Dentistry. Dr Reilly is now affiliated with the Department of Medicine, Cook County Hospital, Chicago, Ill.

Arch Intern Med. 1995;155(21):2317-2323. doi:10.1001/archinte.1995.00430210067010

Background:  Hospitalization presents an opportunity for physicians to discuss advance directives with patients and to encourage completion of health care proxies.

Objective:  To prospectively promote discussion and documentation of treatment-specific directives about life-sustaining interventions (cardiopulmonary resuscitation, admission to critical care units, mechanical ventilation, electrical cardioversion, and vasopressor therapy) among unselected medical inpatients in a community teaching hospital.

Methods:  We conducted a time-series intervention trial from January 1,1991, through June 30 1993, divided into three phases. During the education phase, we provided reminders, education, and feedback to attending physicians; during the intervention phase, we promoted a new documentation form for directives to be used by attending physicians; during the control phase, no interventions occurred. We studied consecutive patients (N=1780) admitted to the hospital acute medical service in each of the following 10 periods: three in the education phase (n=598), three in the intervention phase (n=826), and four in the control phase (n=356). The primary outcome measures were the frequency and content of directives documented by attending physicians in their patients' hospital charts. Secondary outcome measures included physicians' and patients' attitudes about directives, surveyed repeatedly.

Results:  The proportion of inpatients with directives increased significantly during the intervention phase (62.5% vs 23.6% during the education phase and 25.3% during the control period, P<.001, Pearson χ2 test). During the final intervention phase, 227 (83.2%) of 273 inpatients had directives documented in the hospital chart. Increases in clinically important ("impact") directives usually involved intensive care, not do-not-resuscitate status. Overall, 366 (86.7%) of 422 physician-attested directives agreed with the treatment preferences of interviewed patients (κ ranges, 0.53 to 0.79). Physicians' attitudes about and interest in directives improved.

Conclusions:  Institutional interventions can facilitate attending physicians' documentation of treatment-specific directives about life-sustaining care for most medical inpatients. More research is needed to confirm the effect of these efforts on quality and cost of hospital care, patients' autonomy, and their eventual execution of durable directives and proxies.(Arch Intern Med. 1995;155:2317-2323)