How evidence-based are publications in clinical ophthalmic journals?

To evaluate the methodological quality and level of evidence of publications in four leading general clinical ophthalmology journals.

All 1919 articles published in the American Journal of Ophthalmology, Archives of Ophthalmology, British Journal of Ophthalmology, and Ophthalmology in 2004 were reviewed. The methodological rigor and the level of evidence in the articles were rated according to the McMaster Hedges Project criteria and the Oxford Centre for Evidence-Based Medicine levels of evidence.

Overall, 196 (24.4%) of the 804 publications that were included for assessment met the Hedges criteria. Articles on economics evaluation and those on prognosis achieved the highest passing rate, with 80.0% and 74.4% of articles, respectively, meeting the Hedges criteria. Publications on etiology, diagnosis, and treatment fared less well, with respective passing rates of 28.3%, 20.2%, and 14.7%. Published systematic reviews and randomized controlled trials were uncommon in the ophthalmic literature, at least in these four journals during 2004. According to the Oxford criteria, 57.6% of the articles were classified as level 4 evidence compared with 18.1% classified as level 1. Articles on prognosis had the highest proportion (43.0%) rated as level 1 evidence. Generally, articles that reached the Hedges threshold were rated higher on the level-of-evidence scale (Spermans rho = 0.73; P < 0.001).

The methodological quality of publications in the clinical ophthalmic literature was comparable to that in the literature of other specialties. There was substantial heterogeneity in quality between different types of articles. Future methodological improvements should focus on the areas identified as having the largest deficiencies.