We thank Drs Dave and Narayanan for their interest in our recently published paper. With regard to the power calculations please note the following. The population covered by the present study consists of all subjects potentially affected with myopic CNV. It is a population of modalities, and therefore, is unlimited. The sample was selected according to inclusion and exclusion criteria, mentioned in the “Methods.” From a statistical point of view, it is a simple, random, no orderly sample without replacement: it is a nonprobabilistic sample with reasoned choice, which does not allow the application of the Horvitz-Thompson estimator to estimate variance, and then errors and best sample size. We elected to enroll every eligible patient who sought treatment from February 2008 through December 2008; therefore, the sample size was left to chance.

The 1-tailed paired t test we used to compare the best-corrected visual acuity (BCVA) results applies with the compound hypothesis; therefore, the test power 1−β cannot be determined by the Neyman-Pearson lemma. The Neyman-Pearson test, indeed, applies to the case of a simple null hypothesis against a simple alternative hypothesis; on the contrary, the 1−β varies with an alternate hypothesis.

We used an approach based on statistical results obtained in previous surveys in which both BCVA and foveal center thickness showed the same coefficient of variation; therefore, BCVA or foveal center thickness can be examined in the same way to determine test power.

With a sample size of 16 patients in each arm, we found a standard deviation equal to 8.413 for BCVA in the bevacizumab group. We can assume a prudential alternate hypothesis of BCVA improvement equal to 7 letters; given an α error of 5%, we obtain a power of 0.942, and therefore a β error of 5.8%.

As for the proportional confidence interval of a complication rate, it is absolutely arbitrary to evaluate as 25% because we could not examine the problem, not being detected over all of the 32 patients. This means that the probability of complications is less than 3%, corresponding to a maximum proportional confidence interval equal to 11.7%, with an α error of 5%.

However, the indication for a careful preoperative fundus examination and prophylactic laser treatment as needed in these highly myopic eyes was just general advice, as underlined in the “Discussion,” given that myopia a known risk factor for retinal detachment.

Only gold members can continue reading. Log In or Register to continue

Jan 17, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on Reply

Full access? Get Clinical Tree

Get Clinical Tree app for offline access