### Description

Moskowitz et al developed a model for predicting the probability that a patient will receive an allogeneic blood transfusion associated with cardiac surgery at a hospital practicing blood conservation method. The authors are from Englewood Hospital (New Jersey) and Mount Sinai Medical Center (New York City).

Parameters (available preoperatively):

(1) red cell mass in mL (should be red cell volume in mL)

(2) serum creatinine in mg/dL

(3) preoperative prothrombin time in seconds

(4) timing for the operation

(5) type of cardiac surgery performed

(6) number of diseased coronary arteries

 Parameter Finding Points timing for the operation elective 0 urgent or emergent 1 type of cardiac surgery CABG and valve 1 CABG or valve 0 number of diseased arteries 1 0 >= 2 1 serum creatinine < 1.3 mg/dL 0 >= 1.3 mg/dL 1

where:

• The parameter red cell mass is given in mL (see Figure 3, page 630). Red cell mass should be in grams. Red cell mass divided by 1.094 gives red cell volume in mL.

• The method for determining RBC mass is not stated. I assume that it is estimated blood volume times the hematocrit but this can be directly measured.

• CABG = coronary artery bypass graft

X =

= (0.39 * (preoperative PT in seconds)) - (0.0016 * (red cell volume in mL)) + (1.81 * (points for serum creatinine)) + 1.76 * (points for number of diseased arteries)) + (1.37 * (points for type of cardiac surgery)) + (1.25 * (points for operation timing)) + 7.82

probability of allogeneic red blood cell transfusion =

= 1 / (1 + EXP((-1) * X))

NOTE: If the equation is used as written, the estimates do not match expectations. In many logistic regression models the constant is a negative number. More believable answers are returned if (-7.82) is used rather than 7.82.

Limitations:

• The method for determining the prothrombin time and its normal reference range is not stated.

• The volume of blood transfused is not predicted.