Eyes or subjects: are ophthalmic randomized controlled trials properly designed and analyzed?

OBJECTIVE

In ophthalmic randomized controlled trials (RCTs), each subject may have 2 potential data points (i.e., eyes) contributing to the clinical trial. Hence, various study designs may arise requiring different statistical tools. This study aimed to assess the appropriateness of study design, statistical tools used, and reporting of results in ophthalmic RCTs.

DESIGN

A systematic review of 69 ophthalmic RCTs.

METHODS

The study design, sample size calculation, statistical analysis, and reporting methodology of all RCTs published in 4 major general clinical ophthalmology journals in 2009 were assessed.

MAIN OUTCOME MEASURES

The study design of each article under review is evaluated.

RESULTS

The most common study design was a one-eye design (48%). Within this group, only half described the method of selecting the study eye, among which 5 chose the study eye by random selection. In the remaining trials, there were paired-eye design (13%), subject design (19%), and two-eye design (19%). Among the 13 two-eye design studies, 4 allocated both eyes of the subject to the same group, 4 allocated the eyes to different groups, and 4 did not restrict the allocation. None of these studies adjusted for the clustering effect in sample size calculation. Only 5 studies used statistical methods adjusting for nonindependence.

CONCLUSIONS

There is currently substantial heterogeneity in the quality among published ophthalmic RCTs in terms of proper use of study design, sample size calculation, randomization method, and statistical tools. Future ophthalmic researchers are suggested to consult a statistician and to follow some guidelines such as the CONSORT statement when performing an RCT to improve further the quality of clinical trial.