Allogeneic Transfusion - Essay Example

a) There were 121 patients in total undergoing primary replacement that required an allogeneic transfusion
b) No, because in Table 3 of the paper for the ABD group, there was no procedure in which there were patients requiring both allogeneic and autologous transfusion. 
c) To determine whether the difference was statistically significant or not; a two-tail t-test is conducted below.
Given, n1 = 2042; ; s1 = 14.8
n2 = 188; ; s1 = 14.6
Degree of Freedom df = n1 + n2 – 2 = 2228
Significance level  = 5% = 0.05
Null Hypothesis Ho:  
Alternate Hypothesis H1:  
critical = t0.05, 2028 = 1.96
Decision Rule:
If that> is critical; the Null Hypothesis will be rejected else it will be accepted.
stat for this hypothesis test will be
 
Therefore, the Null Hypothesis is rejected. This means the difference in the pre-operative haemoglobin was significant between those who participated in ABD and those who had not.
The significance level is < 0.001
d) The standard deviation for mean units is not given in this table for both groups No ABD and ABD. From statistical consideration, this is a clear error as a mean value hardly conveys anything without standard deviation. 
e) This is not reported in this paper i.e. the researcher has not reported the number of subjects according to the predictive values used in this study and therefore, no comments can be made on whether or not there were a sufficient number of predictors to ensure the validity of the reported regression model.
f) Stepwise procedure
It is a procedure for building up a multivariate regression model. It can be either additive or subtractive. In the case of additive procedure, the significant predictive variables are included in the regression model one after another. In the case of the subtractive approach, all the expected variables are included in the model and then those which do not show statistically significant association with the predicted variable are dropped from the equation. 
In this article “stepwise procedure” refers to the procedure, which was used to create the multivariate regression equation for the occurrence of allogeneic blood transfusion.
“Significantly associated in univariate analyses ……..”
In univariate analyses, each variable is analysed separately for its statistics like mean, median, mode, standard deviation, range, skewness etc. Association on the other hand comes by looking at the trend between two variables. If the values of two variables show a tandem movement, then there exists a significant association between the two variables. Though it looks very similar to correlation, it is completely different as in the case of an association, there may not be any causal relationship between two variables, which is there in the case of correlation.
What has been stated in the quoted sentence is that univariate analysis resulted in a significant association between ‘allogeneic transfusion’ and ‘older age’; ‘allogeneic transfusion’ and ‘female sex’; ‘allogeneic transfusion’ and ‘hip procedure’ and so on. It means that instances of allogeneic transfusion were more in older people, females and so on. But it does not necessarily mean that this relationship was causal as well, it may be or may not be. 
g) ‘Revision hip’ was associated with the highest probability of transfusion as the ‘Risk ratio’ is highest for this procedure. This means that those going for “Revision hip’ are at the greatest risk for allogeneic transfusion than those going for other procedures.
h) The older person (aged 77) is at greater risk of requiring a transfusion after an operation than the younger one. This comes from the fact that the multivariate regression model throws up a risk ratio of 1.77 for those aged 70 – 80 as compared to those aged below 70. This means that the risk of requiring a blood transfusion after knee replacement is 1.77 times more for a person aged 77 than that for a person aged 67.
i) The risk ratio is increasing with increasing coefficients in the multivariate regression model.
j)   The “risk of transfusion” for a patient with a pre-operative haemoglobin between 121 and 130g/L is 2.2 times more as compared to a patient with a pre-operative haemoglobin >130g/L
k) Positive Predictive value = 34.4 Negative Predictive Value = 89.2
l) The ROC curve for the validation data set is presented below.
The predicted probability that provides the best prediction for the need for a transfusion is 0.80 Read More