Intensive care triage in the elderly

We read the editorial by Dr. Levy (1) regarding our manuscript (2) with interest but disagree with several of his points. We used the Acute Physiology Score II, part of the Simplified Acute Physiology Score (SAPS II), without age points to adjust for severity within age groups. Dr. Levy believes that we should have added the age points in calculating the score as was performed in the original SAPS II study. A serious drawback in the interpretation of multivariate analyses occurs when collinear variables are entered. In order to examine, the effect of age per se and to avoid collinearity, age points were excluded from the SAPS II scores. We believe that SAPS without age is still a measure of the risk for 28-day mortality. When the study population is stratified by age and the comparisons are between accepted and rejected patients, the impact of age on the SAPS score is minor as the age points within each stratum are similar for the groups being compared. In addition, other studies have adjusted for severity between groups by using acute physiological scores without age points (3), including one that evaluated illness severity in the elderly (4). One cannot really compare different age group severities if one adds additional points just for the patient’s age. Our previously reported propensity score (5) demonstrated that although patients refused intensive care unit admission were older and had higher SAPS II scores than those accepted, they had lower Acute Physiology Score II scores, suggesting that the observed association of admission refusal with higher severity of illness could be due to a confounding effect of old age in increasing SAPS II scores. The rationale for ascribing a “mortality benefit” to elderly patients was based on a bivariate analysis comparing refused and admitted patients within age groups <65 and >=65, demonstrating a greater reduction in the survival among the refused elderly compared to younger patients. The odds of survival, refused relative to accepted patients controlling for SAPS II, were 0.74 for the younger group and 0.65 in the elder one. Dr. Levy states that the manuscript appears to be written with a bias toward the belief that elderly patients are being denied access to intensive care unit care. The Eldicus study was performed without any bias toward the belief that the elderly are or are not being denied intensive care unit access. Prior to the results of the study we did not think that the elderly would have greater mortality differences between accepted and rejected patients, but these were the findings of the study. We believe the fact that the triage literature is replete with studies showing that physicians reject the elderly more than younger patients together with our new findings compels physicians to relook at their triage policies for the elderly. This is not bias, but rather an evaluation of the facts and recommendations based on them.