Agreement Between Tear Film Tests Performed by Novice Examiners

PURPOSE: To explore the agreement among tear break-up time measurements performed by novice examiners using manual non-invasive tear break-up time (mNIBUT), automated non-invasive tear break-up time (aNIBUT) and fluorescein tear break-up time (FBUT), and to examine the association between tear film metrics and symptoms. METHODS: A total of 61 volunteers (44 females) with a mean age of 21.5 ± 2.7 years underwent tear-film assessment performed by novice examiners (undergraduate students). The tear volume (Schirmer test), tear film stability by mNIBUT, aNIBUT and FBUT, along with lipid layer pattern, were assessed. Symptoms associated with tear-film quality and ocular surface discomfort were evaluated using the Ocular Surface Disease Index (OSDI) questionnaire. RESULTS: Normal values were seen in 88.5%, 93.4%, 73.8% and 39.3% of participants using the Schirmer test, FBUT, mNIBUT and aNIBUT, respectively. The intraclass correlation coefficient (ICC) for tear break-up measurements indicated poor reliability between tests, except when comparing mNIBUT and FBUT (ICC = 0.67 and 0.60 for averaged and first measurements, respectively). Bland-Altman analysis showed that the 95% limits of agreement for mNIBUT-aNIBUT and FBUT-aNIBUT were wider than those for FBUT-mNIBUT. The lack of agreement did not change when averaging the three measurements. Results from the OSDI indicated that most participants did not report symptoms. No significant correlations were observed between the OSDI total score and any of the tear metrics (p > 0.05). The aNIBUT demonstrated the best agreement with symptoms in this young and largely asymptomatic population. CONCLUSIONS: Tear break-up time measurements obtained by novice examiners showed limited reliability and agreement when comparing traditional methods (FBUT and mNIBUT) and the aNIBUT. Examiner-related variability may influence the assessment of tear-film stability, although causal inferences regarding examiner experience cannot be established from this study design. Repeating measurements did not improve the reliability and agreement between tests when performed by novice examiners.