Sanger et al reported a model for predicting surgical site infection based on daily clinical wound assessment. The authors are from the University of Washington and VU Medical Center Amsterdam.
Patient: status post surgery
Evaluation: daily clinical wound assessment, serial
Parameters:
(1) nasogastric tube
(2) degree exudate (presumed exudate amount), maximum
(3) pulse in beats per minute, maximum
(4) slough type, maximum
(5) wound length in cm, maximum
|
Parameter
|
Finding
|
Points
|
|
nasogastric tube
|
no
|
0
|
|
|
yes
|
0.645
|
|
degree exudate
|
none
|
0
|
|
|
scant
|
0.213
|
|
|
small
|
0.426
|
|
|
moderate
|
0.639
|
|
|
large
|
0.852
|
|
pulse
|
|
0.017 * (pulse)
|
|
slough type
|
none
|
0
|
|
|
white/grey nonviable
|
0.165
|
|
|
loosely adherent, yellow
|
0.33
|
|
|
adherent, soft black eschar
|
0.495
|
|
|
firmly adherent, hard black eschar
|
0.66
|
|
wound length
|
|
0.016 * (length in cm)
|
value of X =
= SUM(points for all of the parameters) - 3.087
probability of surgical site infection =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.76.
• The positive predictive value is 0.35 and negative predictive value 0.93.
