Purpose:
Strong relationships between dry eye symptoms and severity of lid parallel conjunctival folds (LIPCOF) are reported. This study investigates (i) intra-observer repeatability and (ii) inter-observer differences in novice and experienced observers classifying LIPCOF and (iii) the ability to differentiate between LIPCOF and micro-folds by fold thickness.
Methods:
45 images of LIPCOF and micro-folds were projected onto a TFT screen in randomized order and classified by two experienced observers (O1/O2) and two novice observer (ON1/ON2) in two consecutive sessions. 95% limits of agreement were calculated from the distributions of differences between observer and observations. Limits of agreement were compared using O`Briens Test of homogeneity of variances. Differences between LIPCOF and micro-folds were evaluated by digital analyses and analyzed by unpaired t-test.
Results:
The mean LIPCOF grades evaluated by the four observers, were not significantly different (ANOVA, p=0.560) (Table 1).
Intra-observer 95% limit of agreement (LoA) for O1 (LoA=0.8) and O2 (LoA=0.9) were similar (p=0.589); however LoA of ON1/ON2 was significantly larger (p=0.001; LoAON1=1.5; LoAON2=2.3). Inter-observers LoA was smaller in experienced observers (O1-O2=1.7; O1-ON1=1.7; O1-ON2=2.1; O2-ON1=2.2; O2-ON2=2.0; ON1-ON2=2.0 (Session II)), but not significantly different (p>0.294). LIPCOF were significant (unpaired t-test; p<0.0019) thicker (0.078mm) than micro-fold (0.024mm).
Conclusions:
Intra-observer limit of agreement was better than the inter-observer limit of agreement. Experience in the classification of LIPCOF increases repeatability. Any change within observations of more than 0.9 should to be considered as abnormal. LIPCOF can easily be differentiated from micro-folds by comparing thickness, since they are three times thicker.
Keywords: clinical research methodology