Response to comment on 'Efficient statistical tests to compare Youden index: accounting for contingency correlation'We thank Drs. Reiser and Nakas for their comments about the variance formula in Cleophas [1] that our article [2] cited and utilized. Its mathematical derivation is more involved than a direct application of the Delta method. We respectfully disagree with their assessment on the variance formula and correlation in our article [2] for reasons detailed in the succeeding text.(1) On the estimation of the variance of the sum of sensitivity and specificity.As in Cleophas et al and Goodman's approximate method [1,3], let X be sensitivity, Y be specificity, and f (X, Y) is a differentiable function of X and Y. Let E(X) = x 0 , E(Y) = y 0 . With the Delta method, the variance of f(X, Y) can be obtained by the following:Taking variance of both side yields the variance of f(X,Y).As we are now dealing with the variance of a function of correlated sensitivity and specificity, to obtain the variance of X + Y, the key is to obtain the covariance. To achieve it, several auxiliary functions are needed.Using above definition, when X and Y are dependent, and the Delta method, we can obtain the following equations:The very reason we chose f(X,Y) = X 2 ± Y 2 and f(X,Y) = XY as the auxiliary functions is that they share the same covariance with f(X,Y) = X + Y. With these auxiliary functions, we haveIn applying Taylor's expansion, f(X, Y) = X 2 ± Y 2 is expanded to order two, and f(X,Y) = XY is expanded to order one.Equation 4 ) is 16x 0 y 0 Cov(X,Y). Using the three auxiliary functions earlier, the common covariance can be obtained as follows: