We appreciate Drs Na, Hwang, and Park’s comments regarding the biostatistical analysis of the data in our manuscript. We agree with their comments regarding the treatment of multiple outcome measures and the importance of applying the Bonferroni correction to solve multiple testing problems attributable to the dependency among a number of hypotheses. We would like to clarify a number of points they raised.
There were 3 designated primary outcome measures defined in our protocol as listed in clinicaltrials.gov (Identifier: NCT01135719 ). The 3 designated primary outcome measures were changes in best spectacle-corrected visual acuity, improvement in uncorrected visual acuity, and changes in low-contrast visual acuity. All 3 primary outcome measures had a designated primary endpoint of 12 months. We also included a number of secondary outcome measures (changes in higher-order aberrations, topographic changes, and quality-of-vision changes) with a 12-month primary endpoint. We should have detailed our study design more clearly in our manuscript, and we regret the omission of this important information.
Based on the 3 primary outcome measures with a single 12-month endpoint, the appropriate adjustment for a type 1 error using Bonferroni correction specifies that a P value be less than .017 in order to be considered statistically significant.
The 2 primary endpoints that we reported as statistically significant (improvement in uncorrected visual acuity and changes in low-contrast visual acuity) were still statistically significant using Bonferroni correction at the 12-month primary outcome, with P values of .016 and .004, respectively. Therefore, our conclusions remain unchanged using the adjusted P values.
We would also like to point out that we reported our data in accordance with requirements detailed by the editorial boards of the Journal of Cataract and Refractive Surgery , the Journal of Refractive Surgery , and Cornea with regard to publishing refractive surgical manuscripts. The American Journal of Ophthalmology policy, as written in the author information, requests that authors use these same requirements when reporting refractive surgical data.