Predictors of Latour et al for Early After-Discharge Mortality in ICU Patients

 

Population: 700 patients older than 14 years discharged from ICUs at 3 hospitals in Valencia, Spain, during the period of January 1986 to May 1987.

 

Parameters found predictive of outcome:

(1) age

(2) days in the ICU

(3) number of organ system failures

(4) SAPS score in the first hour after admission to the ICU

 

Parameter	beta	Finding	Value
age	-0.633	<= 65	- 1
		>= 66	+ 1
days in ICU	-0.415	<= 9	- 1
		>= 10	+ 1
organ system failure (OSF)	-0.423	0	- 1
		>= 1	+ 1
SAPS	-0.822	<= 10	- 1
		>= 11	+ 1

 

X =

= 2.08 - (0.822 * (SAPS value)) - (0.423 * (organ system failure value)) - (0.415 * (days in ICU value)) - (0.633 * (age value))

 

probability of early death within 2 months after discharge from the ICU =

= 1 / (1 + EXP(X))

 

Interpretation:

• The minimum mortality rate is about 1% with all favorable parameters.

• The maximum mortality rate is about 56% with all unfavorable parameters.

• The authors found that the equation was not suitable to make predictions in the individual patient since the variables had a low predictive value. However, it did allow identification of a high risk group of patients who might benefit from more intensive care.

• Patients with an initial SAPs score >= 11 had a 6-fold greater risk of dying in the immediate post-ICU period than those with a score <= 10.

 

NOTE: On page 126, ln (p / (1 - p)) = X. If this is rearranged, p = 1 / (1 + EXP(-X)), not 1 / (1 + EXP(X)) as given. Since (1 / (1 + EXP(-X))) + (1 / (1 + EXP(X))) = 1, one value is the probability of death and the other is the probability of survival. I have elected to go with the example given at the bottom of the page.