Abstract
Purpose :
Previous research has demonstrated that a number of factors related to patient characteristics and testing procedures may affect the reliability indices of Humphrey visual field (HVF) testing. The goal of our study is to investigate whether severity of glaucoma as measured by mean deviation affects the reliability indices of Humphrey visual field tests.
Methods :
This is a retrospective clinical study of patients who have undergone HVF testing at Kresge Eye Institute (KEI) or one of its satellite locations (Sinai Grace, Bingham Farms) over the last 20 years. Patients were included if they performed testing on the 24-2 Swedish Interactive Threshold Algorithm (SITA) program using a size III stimulus. Patients were excluded if testing involved a program other than 24-2 SITA or a stimulus size other than III. The primary outcome was to compare the rates of reliable and unreliable HVF tests between different severity levels of glaucoma. Severity of glaucoma was defined as mild if mean deviation (MD) was between 0 and 7, moderate between 7 and 14, moderate-severe between 14 and 21, and severe if above 21. An unreliable test is defined as either fixation loss greater than or equal to 20%, false positive error and false negative error greater than or equal to 33%. Other variables relating to patient medical and demographic characteristics as well as testing procedures were analyzed. Chi-squared test was used for data analysis.
Results :
Both eyes of 289 patients are included in our study for a total of 578 HVFs – 371 HVFs classified as mild glaucoma, 72 as moderate, 29 as moderate-severe, and 30 as severe. 76 HVFs had MDs greater than 0. 43.6% of all HVFs were considered unreliable. Patients with severe glaucoma had a higher percentage of unreliable HVFs (83.3%), than those with moderate-severe (58.6%), moderate (38.9), or mild (40.2%) glaucoma (P<0.001).
Conclusions :
Although other cofounding variables may affect the reliability indices of HVFs, our study shows that mean deviation may be associated with differences in HVF testing reliability.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.