Skip to main content
editorial
. 2022 Dec 20;3(1):100262. doi: 10.1016/j.xops.2022.100262

Table 1.

Types of Clinical Data Used in Ophthalmology Research

Data Type Utility Limitations Examples
Research data
 Population-based studies Analyzing associations between variables and outcomes, epidemiology, public health
  • •

    Limited sample diversity with predominantly White participants

  • •

    Often only allows for cross-sectional analyses

  • •

    Access may be restricted for researchers external to host institution(s)

 Publicly available: UK Biobank, National Health and Nutrition Examination Survey (NHANES), NIH All of Us Research Program
Not publicly available: Rotterdam Eye Study, Baltimore Eye Study, Los Angeles Latino Eye Study
 Aggregated data Analyses specific to clinical uses or conditions
  • •

    Require data entry and maintenance

  • •

    Nonstandardized data formats among institutions

  • •

    Eye-specific registries may lack systemic diagnosis and medication information

 Intelligent Research in Sight (IRIS), the European Registry of Quality Outcomes for Cataract and Refractive Surgery (EUREQUO), the Japan-Retinal Detachment Registry, the Swedish National Cataract Register
 RCTs Can be used to determine causal associations
  • •

    A substantial proportion of major ophthalmology RCTs were conducted 1–2 decades ago

  • •

    Limited sample sizes

  • •

    Limited sample ethnic diversity

 Early Manifest Glaucoma Trial, Ocular Hypertension Treatment Study, Age-Related Eye Disease Studies, Pediatric Eye Disease Investigator Group
Administrative data
 Institutional data Longitudinal analyses within an institution
  • •

    Limited data quality with missing, duplicate, or misclassified data

  • •

    Nonstandard nomenclature

  • •

    Cannot be generalized

 Data from all clinical practices
 Insurance claims-level data Longitudinal analyses with large and diverse samples
  • •

    Clinical care cannot reliably be inferred from claims data

  • •

    Data can be missing or misclassified

  • •

    Recorded diagnosis and treatment codes may be influenced by institutional coding patterns

 Optum Medical Claims, National Health Insurance Research Database of Taiwan, The Centers for Medicare and Medicaid Services (CMS) Medicare data

RCT = randomized controlled trial.