A Third of Reviews of Corneal-Disease Treatments Not Reliable

By Reuters Staff

May 24, 2019

NEW YORK (Reuters Health) - A third of the systematic reviews of corneal-disease treatments are unreliable, according to a new study.

"The present investigation reveals a cause for concern. It contributes to the evidence that only a fraction of systematic reviews in ophthalmology are deemed to be reliable (subspecialty-specific estimates range from 28% to 70%)," write Dr. Ian J. Saldanha of Brown University School of Public Health in Providence, Rhode Island, and colleagues.

"A particularly concerning aspect is that the percentage of unreliable systematic reviews has not diminished, but rather has increased somewhat . . . since 2012, the year after the landmark Institute of Medicine standards for systematic reviews were published," they note in JAMA Ophthalmology, online May 9.

Systematic reviews are reliable if they "(1) use appropriate methods to search for all available evidence, assess the risk of bias in the included evidence, and qualitatively and quantitatively synthesize it in ways that minimize bias; and (2) are reported completely and transparently," the authors explain.

They evaluated the reliability of 98 systematic reviews on treatment of 15 corneal diseases published in 1997-2017 and included in the Cochrane Eyes and Vision US Satellite database.

The American Academy of Ophthalmology (AAO) updated its Preferred Practice Patterns (PPPs) for six of the diseases in 2018: bacterial keratitis, blepharitis, conjunctivitis, corneal ectasia, corneal edema and opacification and dry-eye syndrome.

Thirty-three of the reviews (34%) were unreliable, most often because the authors had not conducted a comprehensive literature search (67%) or didn't assess individual included studies for risk of bias (39%).

Among the 17 systematic reviews that included quantitative synthesis, 12 (71%) did not use appropriate methods.

Forty-two of the 65 reliable reviews concerned diseases included in the 2018 PPPs, while the updated PPPs cited 33 of the reliable studies.

"As with other fields, it indeed is time for those involved in the ecosystem of evidence-based health care in ophthalmology to address the problem of an ever-growing number of poorly conceived and unreliable systematic reviews," the researchers write.

"Careful adherence by systematic reviewers to best practices (eg the Institute of Medicine standards for systematic reviews) and by journal editors to recommendations regarding reporting and editorial review can help improve the reliability of future systematic reviews in eyes and vision," they conclude.

SOURCE: https://bit.ly/30AHse9

JAMA Ophthalmol 2019.