We read with interest the article by Schojai et al. in which the authors conclude that “Seven years after implantation [of a light-adjustable intraocular lens] and refractive adjustment, eyes with an LAL had stable refraction, good visual acuity, and no IOL-associated pathologies.”1
We commend the authors on such a long-term follow-up with a novel IOL. However, the statistical results of this observational study are our primary interest. In the article, the 4 main outcomes are presented in separate box plot figures for uncorrected distance visual acuity, corrected distance visual acuity, spherical equivalent, and average central corneal thickness with both 1-year and 7-year measures after cataract surgery. The statistical section of the manuscript provides only the software that was used for the analysis; therefore, we were hoping to learn more about what type of summary measures and statistical testing were used. Median values are referred to in the text (which indicate that the variables have a nonparametric distribution); however, the values provided in the text do not always align with the median values observed in the box plots. Standard deviations are also provided in the text, but there are no mean values in the text or box plots. It is important to note that when the distribution of the values is not symmetrical or contains outliers, as indicated by the use of a median to report the center of the distribution, then a standard deviation may not accurately describe the spread of the data values. We are curious whether the authors accounted for the fact that these data are repeated (not independent) measures and that there is also correlation within patients who have 2 eyes included, both of which can influence standard errors and statistical comparisons.2,3
Of particular interest, the authors conclude that refraction as well as uncorrected and corrected visual acuities were stable over the 2 time periods, but if we understand the P values correctly, it appears that these measures all worsened from year 1 to year 7. Further clarification of the hypotheses tested and adding information regarding which statistical methods were performed would help us and other readers to interpret these findings.
Finally, we also noticed the low response rate of 14% of the original set of patients. Although we understand that attrition is a common issue with prospective studies, if there are data available to compare year 1 outcomes among the study cohort and full cohort, this could offer some understanding as to whether the study cohort is representative of the entire cohort. We want to reiterate how much we valued reading this study and look forward to learning more from the authors.
1. Schojai M, Schultz T, Schulze K, Hengerer FH, Dick HB. Long-term follow-up and clinical evaluation of the light-adjustable intraocular lens implanted after cataract removal: 7-year results. J Cataract Refract Surg 2020;46:8–13
2. Ying G, Maguire MG, Glynn R, Rosner B. Tutorial on biostatistics: linear regression analysis of continuous correlated eye data. Ophthalmic Epidemiol 2017;24:130–140
3. Finch WH, Bolin JE, Kelley K. Multilevel Modeling Using R. 2014