Lu et al reported a model for predicting surgery-related pressure ulcers in cardiovascular surgical patients. This can help to identify patients to target for pressure ulcer prevention. The authors are from Nantong University in Jiangsu, China.
Patient selection: cardiovascular surgical patient
Parameters from nomogram:
(1) therapy with corticosteroids
(2) surgical duration in minutes
(3) body weight in kilograms
(4) cardiovascular disease category
points for surgical duration =
= (0.01478 * (duration)) + 0.005726
points for weight =
= (0.08921 * (weight)) - 0.004465
Parameter
|
Finding
|
Points
|
therapy with steroids
|
no
|
0
|
|
yes
|
3.8
|
disease category
|
congenital heart disease
|
0
|
|
thoracic aortic aneurysm
|
5.3
|
|
valvular disease
|
5.6
|
|
coronary heart disease
|
6.6
|
total score for nomogram =
= SUM(points for all of the parameters)
value of X =
= (0.3297 * (score)) - 5.247
probability of pressure ulcer from nomogram =
= 1 / (1 + EXP((-1) * X))
Alternatively, the logistic regression model in Table 3 is:
Parameter
|
Finding
|
Points
|
therapy with steroids
|
no
|
0
|
|
yes
|
1.241
|
disease category
|
congenital heart disease
|
0
|
|
thoracic aortic aneurysm
|
1.747
|
|
valvular disease
|
1.861
|
|
coronary heart disease
|
2.174
|
duration
|
|
0.005 * (duration)
|
weight
|
|
-0.029 * (weight)
|
Y =
= SUM(points for parameters) - 2.235
probability of pressure ulcer =
= 1 / (1 + EXP((-1) * Y))
However, the logistic regression model gives results discrepant from the nomogram. The coefficient may need review.