Children's visual acuity tests without professional supervision: a prospective repeated measures study

Abstract

Background: Home visual acuity tests could ease pressure on ophthalmic services by facilitating remote review of patients. Home tests may have further utility in giving service users frequent updates of vision outcomes during therapy, identifying vision problems in an asymptomatic population, and engaging stakeholders in therapy.

Methods: Children attending outpatient clinics had visual acuity measured 3 times at the same appointment: Once by a registered orthoptist per clinical protocols, once by an orthoptist using a tablet-based visual acuity test (iSight Test Pro, Kay Pictures), and once by an unsupervised parent/carer using the tablet-based test.

Results: In total, 42 children were recruited to the study. The mean age was 5.6 years (range 3.3 to 9.3 years). Median and interquartile ranges (IQR) for clinical standard, orthoptic-led and parent/carer-led iSight Test Pro visual acuity measurements were 0.155 (0.18 IQR), 0.180 (0.26 IQR), and 0.300 (0.33 IQR) logMAR respectively. The iSight Test Pro in the hands of parents/carers was significantly different from the standard of care measurements (P = 0.008). In the hands of orthoptists. There was no significant difference between orthoptists using the iSight Test Pro and standard of care (P = 0.289), nor between orthoptist iSight Test Pro and parents/carer iSight Test Pro measurements (P = 0.108).

Conclusion: This technique of unsupervised visual acuity measures for children is not comparable to clinical measures and is unlikely to be valuable to clinical decision making. Future work should focus on improving the accuracy of the test through better training, equipment/software or supervision/support.

Fig. 1. Systematic bias of parent/carer-measured and orthoptist-measured iSight Test Pro visual acuity.

Modified Bland-Altman plots for parent/carer measured iSight Test Pro (A) and orthoptist measured iSight (B) visual acuity against standard of care measurements. Description: X-axis = mean of iSight Test Pro and standard of care measurements. Y-axis = iSight Test Pro minus standard of care measurements. Systematic bias = median difference between iSight Test Pro and standard of care tests. Limit of Agreement (LOA) = 2.5% and 97.5% quantiles denoted by solid black upper and lower lines. Confidence intervals (CI) = 90% CI (bootstrapping) for LOA and 95% CI for median systematic bias denoted by dotted lines and shading.

Fig. 2. Proportional bias of parent/carer-measured and orthoptist-measured iSight Test Pro visual acuity.

Linear regression of (A) parent/carer iSight Test Pro and (B) orthoptist iSight Test Pro against standard of care visual acuity. Description: The 95% confidence intervals (CI) of the slope gradient span 0; the null hypothesis that level of standard of care visual acuity does not affect iSight Test Pro measure is accepted.