Methods
This retrospective cohort study examined 336 consecutive patients who died in a university tertiary care medical center: 242 without advance directives, 66 with a previously completed advance directive, 13 admitted for the express purpose of terminal care, and 15 who signed an advance directive during their terminal hospitalization. Total charges (hospital and physician) were calculated for all patients and were adjusted using both physician and hospital diagnosis-related group weights. Patient participation in end-of-life decisions was determined by chart review.

Results
The group without advance directives had dramatically higher mean total ($49 900 vs $31 200) terminal hospitalization charges than the group with previously completed advance directives, producing a charge ratio of 1.6. After diagnosis-related group adjustment, the charge ratio was 1.35 (95% confidence interval, 1.07 to 1.72) for physician charge, 1.36 (95% confidence interval, 1.06 to 1.74) for hospital charge, and 1.35 (95% confidence interval, 1.08 to 1.73) for total charge. Multiple regression analysis controlling for age, sex, and cancer diagnosis confirmed these findings. Patients with advance directives were significantly more likely to limit treatment and to participate in end-of-life decisions.

Conclusion
Patients without advance directives have significantly higher terminal hospitalization charges than those with advance directives. Our investigation suggests that the preferences of patients with advance directives are to limit care and these preferences influence the cost of terminal hospitalization.

(Arch Intern Med. 1994;154:2077-2083)