Wibring et al reported models for evaluating patients with chest pain. One model predicts high-risk chest pain while a second predicts low-risk chest pain. The authors are from multiple institutions in Sweden.
Patient selection: chest pain
Parameters:
(1) age in years
(2) sex
(3) previous history of heart failure
(4) previous history of kidney failure
(5) prehospital serum troponin T in ng/L
(6) ST depression in ECG
(7) pallor
(8) pain onset during activity
(9) constant pain
(10) pain in right arm
(11) pressuring pain
Parameter
Finding
OR
beta-coefficient (LN OR)
age in years
0.98
-0.02* (age)
sex
female
1
0
male
0.73
-0.315
history heart failure
no
1
0
yes
1.57
0.451
history kidney disease
no
1
0
yes
1.91
0.647
serum troponin
< 40 ng/L
1
0
40 to 100 ng/L
0.25
-1.386
101-1000 ng/L
0.09
-2.408
> 1,000 ng/L
0.11
-2.207
ST depression
no
1
0
yes
0.24
-1.427
pallor
no
1
0
yes
0.41
-0.892
pain during activity
no
1
0
yes
0.69
-0.371
constant pain
no
1
0
yes
0.67
-0.4
pain right arm
no
1
0
yes
0.44
-0.821
pressuring pain
no
1
0
yes
0.64
-0.446
constant
52.56
3.962
value of X =
= SUM(beta-coefficients)
probability of high-risk chest pain =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.78.
To read more or access our algorithms and calculators, please log in or register.
