Abstract
Purpose :
Trachoma is the leading infectious cause of blindness worldwide. Prevalence can be assessed by field grading or grading photographs. We hypothesize that poor image quality may contribute to differential grading of trachomatous inflammation—follicular (TF) between field and photo graders.
Methods :
Field grades for TF and tarsal conjunctival photographs were obtained by a field grader from one or both eyes in 3118 children from Kongwa District, Tanzania, 2 years post-MDA. 2 additional graders graded and adjudicated each photo. We identified the eyes where the field and photo TF grades did not match and assigned each eye an image quality score and a potential reason for mismatch (e.g., different interpretation of follicle number or size, field documentation error, or no obvious reason). We also assigned image quality scores to a random sample of 180 eyes with matching field and photo TF grades.
Results :
5220 eyes had matching field and photo TF grades while 177 eyes had mismatched field and photo TF grades (Fig 1). There was no difference in image quality between mismatch eyes where the field grader assigned a grade of TF and mismatch eyes where the photo graders assigned a grade of TF (p=0.6209), but there was a significant difference in image quality between eyes with matching field and photo grades and eyes with mismatched field and photo grades (p=0.0186). However, no image quality issue stood out as significantly different between eyes with matching vs. mismatched TF grades (Fig 2). Disagreement over follicle number (54.8% of mismatch eyes) or size (42.9%) were the most common potential reasons for grade mismatch.
Conclusions :
Since there was a significant difference in image quality between eyes with matching vs. mismatched field and adjudicated photo grades, image quality seemed to contribute to grading disagreements between field and photo graders. Poor image quality may alter photo graders’ view of the tarsal conjunctiva and affect their determination of follicle size or number. However, as there was no difference in image quality between mismatch eyes with a field vs. photo grade of TF, image quality did not seem to influence the direction of the mismatch (i.e., which graders assigned a grade of TF).
This is a 2021 ARVO Annual Meeting abstract.