Systematic Review Of Sample Size Calculations And Reporting In Randomized Clinical Trials In Ophthalmology Over A 20-Year Period

Randomized Controlled Trials (RCTs) are considered the gold standard for the practice of evidence-based medicine. The purpose of this study is to systematically assess the reporting of sample size calculations in ophthalmology RCTs in 5 leading journals over a 20-year period. Reviewing sample size calculationsin ophthalmology RCTs will shed light on the methodological quality of RCTs and, by extension, on the validity of published results.

Sample size calculation is important in RCT methodology reporting, since underestimated or overestimated sample sizes are difficult to interpret.  Studies in various medical fields have demonstrated,that a significant number of published RCTs still fail to mention all necessary parameters, thus rendering sample size re-estimation impossible.

The MEDLINE database was searched to identify full reports of RCTs in the journals Ophthalmology, JAMA Ophthalmology, American Journal of Ophthalmology, Investigative Ophthalmology and Visual Science, and British Journal of Ophthalmology between January and December of the years 2000, 2010 and 2020. Screening identified 559 articles out of which 289 met theinclusion criteria for this systematic review. Data regarding sample size calculation reporting and trial characteristics was extracted for each trial by independent investigators.

In 2020, 77.9% of the RCTs reported sample size calculations as compared with 37% in 2000 (p < 0.001) and 60.7% in 2010 (p = 0.012). Studies reporting all necessary parameters for sample size recalculation increased significantly from 17.2% in 2000 to 43.0% in 2020 (p<0.001). Reporting of funding was greater in 2020 (98.8%) compared with 2010 (89.3%) and 2000 (53.1%). Registration in a clinical trials database occurred more frequently in 2020 (94.2%) compared to 2000 (1.2%; p < 0.001) and 2010 (68%; p < 0.001). In 2020, 38.4% of studies reported different sample sizes in the online registry from the published article. The number of patients enrolled in a study and the number of eyes studied was significantly greater in 2020 (p < 0.001).

Sample size calculation reporting in ophthalmology RCTs has improved significantly between the years 2000 and 2020 and is comparable to other fields in medicine. However, reporting of certain parameters remains inconsistent with current publication guidelines.