![]() ![]() ![]() Our model was statistically significant ( P < 0.001), which means that our model significantly fits the data better than using the average rate of visual field change of the sample (2) R 2 values are used to measure the goodness of fit of a linear model for the “reference” sample. Nevertheless, two aspects of our analyses need to be heightened: (1) The coefficient of determination and statistical significance of the model ( P value) should be examined in combination. ![]() This limitation is carefully described, elucidated, and extensively discussed in our manuscript, in which we have recommended caution when interpreting the results of our risk calculator. We appreciate their thoughtful comments and recognize that our study is limited by the low R 2 statistic (coefficient of determination) of our model, in which the dependent variable “rates of VF change” was predicted by a set of baseline clinical independent variables using a linear model. 2 In brief, they commented on the limitations and low accuracy of our model, which might make it unsuitable for use in clinical practice. We read with great interest the letter by Drs Crabb, Saunders, and Russell 1 regarding our recent publication.
0 Comments
Leave a Reply. |