Abstract
Purpose: :
It is estimated that 50 to 75% of persons with glaucoma are undiagnosed. Over half of undiagnosed persons are seen regularly in the primary care setting but remain unidentified and under–referred. The American Academy of Ophthalmology (AAO) developed the Glaucoma Risk Factor Score (GRFS) for the identification and referral of patients at high risk for glaucoma. The AAO recommends the GRFS as a feasible and effective primary care screening procedure. There is currently no literature on the validity of GRFS. The purpose of this study is to determine the validity of the GRFS in identifying glaucoma based on standard automated perimetry (SAP) examination.
Methods: :
The study population is a 33,500 patient database screened for glaucoma (including SAP) and collected by the Congressional Glaucoma Caucus Foundation (CGCF). The four GRFS independent variables of age, race, family history, date of the last eye examination were extracted from the CGCF’s database. The dependent variable was the diagnosis of glaucoma as determined by SAP. The null hypothesis of no relationship between the GRFS and a diagnosis of glaucoma was evaluated by multiple logistic regression, ROC curve analysis, and estimation of binomial probability.
Results: :
Only subjects that had SAP results recorded for both eyes were included (n = 27,646). There was a significant relationship between weighted total score of GRFS and SAP. Using a cut off point of all four GRFS variables there was a significant relationship between GRFS and SAP. The sensitivity is 44.3% (95% CI: 0.432 to 0.454) and specificity is 70.7% (95% CI: 0.701 to 0.713).
Conclusions: :
Glaucoma risk factor score is not highly associated with glaucoma based on SAP in this large screening population study. Modification of the GRFS cut off point does improve the sensitivity. Changing the cut off point from all 4 variables to 3 increases the sensitivity to 76%, but decreases the specificity to 38.8%.
Keywords: clinical (human) or epidemiologic studies: risk factor assessment • clinical (human) or epidemiologic studies: prevalence/incidence • clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology