Abstract
Purpose :
To comprehensively describe large publicly available health databases used in published big data research in ophthalmology.
Methods :
To identify studies using large publicly available health databases for ophthalmic research, a literature search was formulated by a health sciences librarian and conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines using the PubMed and Embase digital literature databases from inception (1809 and 1947, respectively) through August 1, 2021. Databases were identified from full-text articles. Study selection and data extraction were performed independently by two investigators, and a third investigator arbitrated any disagreements.
Results :
In total, 204 databases were identified in the included studies; most (52%; 107/204) were available online. Most were provided by government agencies (52%; 55/106) or academic institutions (30%; 32/106) and were from high-income (87%; 92/106) or English-speaking (68%; 72/106) countries; no publicly available databases were identified from Latin America and the Caribbean, the Middle East and North Africa, or Sub-Saharan Africa. Half were comprised of administrative data (50%; 53/106). Few published details about their sampling methods (23%; 24/106) or study populations (41%; 43/106). Some common requirements to access data included an application requirement (29%; 31/106), training requirement (10%; 11/106), purchase requirement (8%; 9/106), or other requirements such as citizenship or employment (8%; 8/106). One-third were freely available (33%; 35/106).
Conclusions :
Big data research in ophthalmology was concentrated on databases based in high-income countries. Access to many publicly available databases was restricted by application, training, purchase, and other requirements. Infrequent disclosure of sampling methods and study populations may impact the quality, transparency, and generalizability of big data research in ophhtalmology.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.