Editor—O'Brien seems to attach greater weight to a study of apparently declining acuity in a non-compliant group than to our observation in a randomised controlled study of no deterioration in our control group, who showed the same eventual result from deferred treatment. This simply illustrates again the pitfalls inherent in natural history studies, which are plagued by unmeasurable biases that randomised controlled trials are designed to avoid.
Srinivas comments on patching dose. A recent study has shown that two hours' patching is as effective as six for children whose vision is in this range.1
Harrad et al agree that an acuity of 6/9 may be within the normal range for age, and so it is hard to see why they do not endorse our conclusion that treatment of such children may be inappropriate. Screening does detect defective vision rather than amblyopia, but children found to have refractive errors without amblyopia are still currently treated.
We did have ample power to detect a useful effect in the mild subgroup (6/9 and 6/12 vision at recruitment). This contained 100 children and their maximum potential for acuity gain from treatment was 0.2-0.3 log units—for example, from 6/9.5 to 6/6. We had 80% power to detect a difference of the order of 0.1 log units and > 99% power to detect a difference of 0.2 log units. However, the observed difference in this group between deferred and full treatment was only 0.04 log units (approximately half a Snellen line).
We believe that our study adds weight to the case for early school entry vision screening. However, in the current state of knowledge, treatment should be targeted only at children with poorer acuity to avoid inappropriate treatment of physiological refractive errors.
Competing interests: None declared.
References
- 1.Pediatric Eye Disease Investigator Group. A randomised trial of patching regimens in the treatment of moderate amblyopia in children. Arch Ophthalmol 2003:121: 603-11. [DOI] [PubMed] [Google Scholar]