Patient is a woman in her mid-30's with a pretty complicated remote and recent history - ESRD, multiple belly surgeries in the past with a very long postoperative hospitalization (family said a year with several months of ICU time). For us, she came up about 2 months ago for distal pancreatectomy for chronic pancreatitis. They did a splenectomy as well, and she ended up requiring something like 60 units of blood over a couple of days postop. Spent a couple of weeks on nimbex with a BIS, and very slowly made progress. Spent about 5 weeks in the ICU and went to the floor with a large abdominal wound and some fistulas.

She had switched over to HD with the transfer but hadn't really been able to tolerate the fluid removal and so she ended up getting ahead several liters over the course of a few days. They finally transferred her back to the unit when she was in significant respiratory distress and was pretty hypotensive (her norm is apparently pretty dang low - 80's/40's - but her primary doc was out of town; not sure how low she was when they transferred). They suspect she's going septic and someone makes the call to put her on xigris. Its a slow night last week when I first had her and she's running pretty stable so I start doing some research on xigris since she doesn't seem sick enough for it to me. I find the second study Eli Lilly did showing xigris wasn't worth the risk for patients not in "severe sepsis", ie, APACHE II < 25. We've got a nifty online tool, so I bust it out and it turns out her APACHE II score would really have only been about 20.

Here are my questions:

Do you, as nurses, ever use the APACHE II tool and do you know much about it? At my hospital, we never calculate it and you hardly ever hear it referenced.

Quite a few studies say APACHE II loses it prognosticative value for death-risk after 7 days, and its ideally used within 24 hours of admission. Does it also lose its accuracy in describing the severity of a patient's condition? When would it be appropriate to use this on a long-term patient? This particular patient made a pretty slow slide over several days into getting readmitted - would it be more appropriate to take the worst values over the course of that several days to control for the treatment she was receiving?

On the tool itself, for someone in chronic renal failure, which serum creatinine box do you fill in - with or without acute renal failure?

Any help or a point in the right direction for resources would be greatly appreciated.