To analyze the quality of life of critically ill patients before their intensive care admission and its relation to age, variables measured in the intensive care unit (ICU; severity of illness, therapeutic effort, resource utilization, and length of stay), and in-hospital mortality rate. Observational prospective multicenter study. Eighty-six medical-surgical ICUs in Spain, including coronary patients. We studied 8,685 patients between 1992 and 1993. Patients <16 yrs old and those dying within the first 6 hrs were excluded. Data collection included age, gender, admission diagnosis, severity level by Acute Physiology and Chronic Health Evaluation (APACHE) III, quality of life survey score, therapeutic activity level by Therapeutic Intervention Scoring System (TISS), and ICU and hospital mortality rate. Pre-ICU quality-of-life score was 3.74 +/- 4.42 points; 33.24% of patients had a normal quality of life (0 points), and numbers of patients declined logarithmically in relationship to increasing quality-of-life scores, with only 189 patients having a score >15 points. Pre-ICU quality-of-life score correlated with age (r =.289, p <.001), with severity level by APACHE III score (r =.217, p <.001), and weakly with TISS (r =.067, p <.001). There was no correlation between quality of life and length of ICU stay. Patients dying in hospital after ICU discharge (n = 429) had worse quality of life (5.88 +/- 5.38 points) than those dying in the ICU (n = 1,453, 4.8 +/- 4.94), who themselves had a worse quality of life than hospital survivors (n = 6,803, 5.05 +/- 5.07; p <.0001 by analysis of variance), with significant differences between all three groups. In the multivariate analysis, pre-ICU quality-of-life was related to age, APACHE III score, and hospital mortality rate but not to TISS or ICU length of stay. Pre-ICU quality of life was introduced as a variable in the APACHE III prediction model and entered the model after acute physiology score, diagnosis, and age and before prior patient location and comorbidities. The area under the receiver operating characteristics curve was 0.834 when quality-of-life was included and 0.83 when not. In Spain, the quality of life of critically ill patients before their ICU admission is good, and only a small proportion of patients have a low quality of life before admission. Previous quality of life is related to hospital mortality rate but contributes very little to the discriminatory ability of the APACHE III prediction model and has little influence on ICU resource utilization as measured by length of stay and therapeutic activity.