Saturday, January 14, 2006
ICU outcomes for the oldest old
The current Journal of the American Geriatrics Society has a study on ICU mortality for those 90 years old & older. It is a prospective cohort study in a single Greek ICU comparing mortality for those over 90 with other patients. ICU mortality for the oldest was 20%, and 30 day in-hospital mortality was 40% (compared to 7% and 9%, respectively, for those under 90 years old). Interestingly, ICU mortality for those in their 80's was 11%, half of that for those in the next decade. Functional status for those who left the hospital alive was not measured, although most were discharged home. APACHE II score was the only independent predictor of mortality for the oldest old.
The authors set this data against what they suggest is a movement to a priori deny ICU care to the oldest old as a cost/resource savings measure. If this movement, in any organized or serious form, actually exists anywhere I'd be curious to see. They conclude that age alone can't be used to deny people ICU care because these outcomes aren't all that bad. Well, yes & no. Advanced age does carry with it a much higher risk of mortality ( 4.5 times higher in this study!), and if one were to ration ICU services based on likelihood of benefit then age would need be part of this mix. But if one were to do this then APACHE II scores would probably need to be a part of this mix as well, as the APACHE II strongly predicts mortality. This would then lead to an argument of denying ICU care to those who are sickest--who need it the most--which is ridiculous. The issue then, which strikes me as a false one, is the idea of denying ICU services to any broad group of people a priori. Use of the ICU--or any medical intervention--should be decided individually, based on an individual's comorbidities (including age), severity of illness, goals of care, and values. This is harder, messier, and more labor-intensive than the use of demographics as a clinical decision-making tool, but that's a good thing in my opinion.
