Ackroyd et al reported a score for predicting the need for red blood cell transfusion in a woman undergoing hysterectomy for ovarian cancer. The authors are from Temple University and Fox Chase Cancer Center in Philadelphia.
Patient selection: hysterectomy for ovarian cancer
Outcome: red blood cell transfusion
Parameters:
(1) age in years
(2) race
(3) ascites at presentation
(4) preoperative hematocrit in percent
(5) preoperative platelet count
(6) disseminated cancer
(7) type of hysterectomy
(8) anticipated advanced procedure (large bowel resection, exenteration)
|
Parameter
|
Finding
|
Points
|
|
age in years
|
< 65 years of age
|
0
|
|
|
>= 65 years of age
|
1
|
|
race
|
White
|
0
|
|
|
African American
|
1.25
|
|
|
Other
|
1
|
|
ascites at presentation
|
no
|
0
|
|
|
yes
|
1.25
|
|
preoperative hematocrit
|
>= 37%
|
0
|
|
|
33 to 36.9%
|
1.5
|
|
|
30 to 32.9%
|
3
|
|
|
< 30%
|
4.5
|
|
preoperative platelet count
|
>= 400,000 per µL
|
1.25
|
|
|
< 400,000 per µL
|
0
|
|
disseminated cancer
|
no
|
0
|
|
|
yes
|
1
|
|
type of hysterectomy
|
laparoscopic
|
0
|
|
|
abdominal
|
4
|
|
advanced procedure
|
no
|
0
|
|
|
yes
|
3
|
total score == SUM(points for all 8 parameters)
Interpretation:
• minimum score: 0
• maximum score: 17.25
• The higher the score the greater the likelihood of blood transfusion.
An equation based on the data in Figure 2:
probability of blood transfusion =
= (-0.04041 * ((score)^3)) + (1.17 * ((score)^2)) - (2.384 * (score)) + 1.039
Performance:
• The area under the ROC curve was 0.80 for derivation and 0.69 for validation cohorts.
