Documenting life-support preferences in hospitalized patients

Purpose: The purpose of this article was to determine the extent to which patients at high risk of hospital death who undergo cardiopulmonary resuscitation (CPR) have previously had their life support preferences addressed and documented. Materials and Methods: We conducted a retrospective chart review of all patients older than 18 years of age hospitalized for more than 24 hours who sustained a cardiac arrest with attempted CPR at our tertiary care university teaching hospital during 1994 (n = 71). We searched all hospital charts specifying ICD-9 codes: Cardiac arrest, ventricular fibrillation, ventricular tachycardia, asystole, electromechanical dissociation, defibrillation, or CPR, Patients were selected if (1) they had a true cardiac arrest (abrupt cessation of spontaneous circulation) and (2) had attempted CPR or defibrillation. Patients were classified as "high risk" if they satisfied at least one of the following: modified prearrest morbidity index greater than or equal to 7, moderate/severe dementia, day 1 APACHE II score > 24 or greater than or equal to 4 dysfunctional organ systems. Results: We searched 147 charts; of 71 patients meeting inclusion criteria, 53 were high risk. Of patients at high risk of sustaining a cardiopulmonary arrest during the index hospital admission, 3 (6%) had preferences addressed within the first 24 hours of hospitalization, 7 (13%) had delayed discussion bf preferences before arrest, 23 (43%) had preferences addressed post arrest, and 20 (38%) had no documented discussions. Of the 23 high-risk patients initially surviving cardiac arrest, all were subsequently given "do not resuscitate" orders. Univariate analysis of factors associated with life-support discussion before cardiac arrest were previous cardiac arrest (OR, 5.9) and APACHE II score > 24 (OR, 1.1), although neither reached statistical significance. None of the 32 patients with a modified PAM index greater than or equal to 7 (32 of 71) survived hospitalization. Only 3 patients survived to hospital discharge. Conclusions: Early communication regarding life-support preferences is important in high-risk patients so that inappropriate or unwanted treatment is not implemented. Given that optimal care includes addressing and documenting life-support preferences in high-risk patients early in their hospitalization, this standard was infrequently met.