Rahmanian et al developed a model for predicting mortality in patients undergoing cardiac surgery. This includes preoperative risk factors and postoperative complications. The authors are from University Hospital Cologne in Germany.
Patient selection: cardiac surgery
Outcome: perioperative mortality
Parameters:
(1) emergency procedure
(2) preoperative pulmonary hypertension
(3) age in years
(4) peripheral vascular disease
(5) preoperative atrial fibrillation
(6) type of procedure
(7) renal failure after surgery requiring dialysis
(8) sepsis after surgery
(9) respiratory failure after surgery
(10) gastrointestinal complications after surgery
|
Parameter
|
Finding
|
Points (beta coefficients)
|
|
emergency procedure
|
no
|
0
|
|
|
yes
|
1.6
|
|
preoperative pulmonary hypertension
|
no
|
0
|
|
|
yes
|
1
|
|
age in years
|
<= 70 years
|
0
|
|
|
> 70 years
|
0.6
|
|
peripheral vascular disease
|
no
|
0
|
|
|
yes
|
0.9
|
|
preoperative atrial fibrillation
|
no
|
0
|
|
|
yes
|
0.4
|
|
procedure
|
CABG
|
0
|
|
|
other than CABG
|
0.4
|
|
renal failure after surgeyr
|
no
|
0
|
|
|
yes
|
1.4
|
|
sepsis after surgery
|
no
|
0
|
|
|
yes
|
1.2
|
|
respiratory failure after surgery
|
no
|
0
|
|
|
yes
|
1.2
|
|
GI complications after surgery
|
no
|
0
|
|
|
yes
|
1.2
|
where:
• The procedure with risk was "other than CABG". The relation to a person who had a combined CABG and other procedure is unclear. In the implementation it is scored as CABG only or not.
X =
= SUM(points for all 10 parameters) - 4.9
probability of mortality =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve was 0.89 in the derivation cohort and 0.92 in the validation cohort.
