Abstract
Purpose: :
To investigate the agreement between letter scores obtained on the Early Treatment Diabetic Retinopathy Study (ETDRS) chart at 4 meters and 1 meter.
Methods: :
This was a retrospective chart review conducted in a hospital-based ophthalmic outpatient department. 205 patients were assessed with the ETDRS chart. Patients unable to read 40 letters on the ETDRS chart at 4 meters had to read the chart again at 1m. This was performed by the same ophthalmic practitioner in a standardized setting with standard recording protocols. To reflect a "real-world" scenario, no forced-choice testing or termination rules were utilized. All patients were instructed to read down to the smallest letter they could see. After excluding patients who could not read any letters at 4m or who were able to read all 70 letters at 1m (with pinhole), 65 patients remained for further analysis. The raw ETDRS letter scores at 4m were obtained and grouped for analysis i.e. Group 1: 1-10 letters; Group 2: 11-20 letters; Group 3: 21-30 letters and Group 4: 31-40 letters. A Bland-Altman analysis was performed to assess the level of agreement between the two measurements. The expected difference between the raw 4m ETDRS score and the1m score is 30. The mean difference (score at 1m minus raw score at 4m) and 95% limit of agreement (LOA) between the 4m and the 1m scores was calculated.
Results: :
There were 38 males and 27 females in the study. The mean age (SD) was 67 (18) years (range 26-91, median 65). The mean difference (95% LOA) for each group was; (32, 95% LOA 21-44, n=16, Group 1 ), (30, 95% LOA 19-42, n=16, Group 2 ), (32, 95% LOA 17-47, n=17, Group 3 ), (28, 95% LOA 21-36, n=16, Group 4 ). The mean difference (95% LOA) of the entire cohort was 31 (20 -43) letters.
Conclusions: :
The difference between the raw score at 4m and the total score at 1m is close to the expected value of 30, although this can vary between 20 to 40 letters. This implies that in some patients, the 4m and 1 m scores are not directly interchangeable. The disagreement between 1 and 4 m scores should be borne in mind when interpreting visual acuity outcomes and designing vision testing protocols for both research and clinical purposes.
Keywords: visual acuity • clinical research methodology • clinical (human) or epidemiologic studies: systems/equipment/techniques