Rapid Responses to:

PAPERS: C Williams, K Northstone, R A Harrad, J M Sparrow, I Harvey, and ALSPAC Study Team Amblyopia treatment outcomes after screening before or at age 3 years: follow up from randomised trial BMJ 2002; 324: 1549 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

The ALSPAC study did not address the utility of preschool vision screening

Michael P Clarke, Philip Griffiths, same appointment and address   (2 July 2002)

Should children be screened repeatedly between 8-37 months for amblyopia?

P.R. Sankari, Mr.J.S.Mars, Paediatric Ophthalmologist, Same address.   (16 July 2002)

Author's responses

C Williams, Richard Harrad, John Sparrow, Ian Harvey, Kate Northstone and Jean Golding (director of ALSPAC)   (1 August 2002)

The ALSPAC study did not address the utility of preschool vision screening 2 July 2002
Michael P Clarke, Senior lecturer, ophthalmology Royal Victoria Infirmary, Newcastle upon Tyne, Philip Griffiths, same appointment and address Send response to journal: Re: The ALSPAC study did not address the utility of preschool vision screening	Williams et al (BMJ 2002;324:1549, 29th June) have overinterpreted their data and reached an unjustified conclusion that preschool vision screening is effective in reducing amblyopia prevalence. Their study compares an intensive programme of screening involving 6 tests in the first 37 months against one standard preschool orthoptic screen at 37 months. The authors claim that intensive screening reduced amblyopia prevalence compared to standard screening. As such an intensive programme of screening would be clinically impractical on a national basis, this does not add significant information to inform the debate about the utility of preschool vision screening. Their conclusion that intensive screening reduced amblyopia prevalence is also open to question. It is not clear that the outcome variables were prestated (and in other respects the report does not adhere to CONSORT reporting criteria). The method of randomisation is questionable, and the 45% of children lost to follow up were more likely to have had significant amblyopia. The amblyopia B measure which was significantly affected by intensive screening (<0.3 LogMAR = 6/12 Snellen equivalent) would exclude 70% of children currently failing screening tests with acuities in the worse eye of 6/12 and 6/9. The possibility therefore exists that any reduction in amblyopia prevalence from intensive or earlier screening would be gained at the expense of large numbers of children receiving unnecessary or ineffective treatment.
Should children be screened repeatedly between 8-37 months for amblyopia? 16 July 2002
P.R. Sankari, Staff Grade in Ophthalmology Royal Albert Edward Infirmary, Wigan Lane,Wigan, WN1 2NN., Mr.J.S.Mars, Paediatric Ophthalmologist, Same address. Send response to journal: Re: Should children be screened repeatedly between 8-37 months for amblyopia?	We read the article by Cathy Williams et al with interest (1). We recognise the study contributes towards important treatment issues regarding amblyopia. But what is not clear from the article is the distribution of children with poor initial visual acuity and type of amblyopia between both the groups. Authors have discussed the distribution of various factors between both the groups like birth weight, breast- feeding maternal education use of car etc but not the above-mentioned factors. If by chance, children with poor initial visual acuity, strabismic anisometropic amblyopes were present (statistically significant?) in the control group, then that could have affected the (long term)final visual outcome (2&3) rather than the intensive screening alone. An earlier report from the study (4) suggested that cover test and visual acuity test were poorly sensitive until the children were 37 months. Yet the authors have subjected the children below 37 months to such an intensive and expensive screening programme. In spite of such intensive screening, 52.5 %(21/40) of amblyopia was detected at and after 37 months of age in the intensive group. We agree with Michael. P. Clarke?s (Rapid response 2nd July) response that such an in intensive screening programme with repeated testing on national basis will be impractical. References 1.C.Williams, K.Northstone, R.A Harrad, J.M.Sparrow, I.Harvey. Amblyopia treatment outcomes after screening before or at age 3 years: follow up from randomised trial. BMJ 2002; 324:1549(29 June) 2.Levartorksy SH, Oliver M, Gottesman N, M.Shimshoni. Longterm effect of hypermetropic anisometropia on the visual acuity of treated amblyopic eyes. . Br J Ophthalmology 1998:82:55-58(Jan) 3.Levartorksy SH, Oliver M, Gottesman N, et al. Factors affecting long-term results of successfully treated amblyopia. Br J Ophthalmol 1995; 79:225-228[Abstract]. 4. Williams C, Harrad RA, Harvey I, Sparrow JM, ALSPAC Study Team. Screening for amblyopia in preschool children: results of a population- based, randomised controlled trial. Ophthalmic Epidemiol 2001; 8: 279- 295[Medline].
Author's responses 1 August 2002
C Williams, Consultant OPhthalmologist Bristol Eye Hospital, Lower Maudlin Street, Bristol BS1 2LX, Richard Harrad, John Sparrow, Ian Harvey, Kate Northstone and Jean Golding (director of ALSPAC) Send response to journal: Re: Author's responses	Response to PR Sankari and JS Mars (Rapid Response 16 July 2002) We agree that of course such an intensive screening programme as used in our intervention group would be impractical on a national basis and at no point in our paper do we recommend that it should be considered as such. Rather, we explicitly state that research into feasible strategies is needed, given the benefits of early treatment suggested by our results. The aim of our study was to estimate the difference in outcome between earlier vs. later treatment for amblyopia, not to assess a specific population screening programme. The study for which this paper gives follow-up data, was designed to fit in with the parent ALSPAC study and had adequate power to detect changes in the outcomes of interest for all children with amblyopia, rather than for smaller sub-groups considered separately. We therefore present the data accordingly. In our earlier paper however, we had already stated that there were no differences between the groups in the cumulative incidence data for children with strabismic, anisometropic or mixed amblyopia, by the age of 37 months, although they had presented at different ages (1). Therefore the differences in the outcome results are not due to differences in the types of amblyopia in the two groups. Our earlier report stated that cover testing and visual acuity testing were poorly sensitive until 37 months of age. These findings were secondary outcomes from the study and our data suggested that photorefraction would have been more sensitive in that age group. We had carried out the early study specifically to determine the relative effectiveness of each of a range of tests conducted up to 37 months, as such information was not available in the literature. Our recent paper describes follow-up data on these same children at 7 ½. We did not repeat the intensive protocol to provide data for this follow-up study, as Sankari and Mars seem to suggest. 1. Williams C, Harrad RA, Harvey I, Sparrow JM and the ALSPA study team. "Screening for amblyopia in preschool children: results of a population-based, randomised comtrolled trial". Ophthalmic Epidemiology 2001: 8 (5) 279-95. Response to Clarke and Griffiths (Rapid Response 2 July 2002) Our study was designed to estimate the effect of earlier treatment for amblyopia vs. later treatment, not to pilot a particular programme. At no point do we recommend that the intervention programme should be considered for population screening. Rather we explicitly say that research into feasible programmes is needed. However, by providing robust data illustrating improved outcomes for children treated for amblyopia earlier rather than later, we have contributed to the evidence base for one of the several criteria needed to justify a population screening programme. Other important issues (eg coverage of population, referral routes and compliance with referral and treatment) clearly need other, dedicated studies. The report as originally submitted conformed entirely to the CONSORT guidelines. Subsequent revisions and abbreviations were made in response to suggestions from the BMJ editorial board. We will happily supply any further methodological details. Our outcome variables were pre-stated in the grant proposal that funded the study. The method of randomisation is clearly described and was devised for the ALSPAC study because of particular pragmatic constraints present when carrying out nested randomised studies within the observational birth cohort study. We explicitly state that all the staff who performed the eye testing were unaware of the rules governing allocation into groups, the mothers dates of birth or of the child's screening history. We therefore think it highly unlike any systematic bias could have arisen. The demographic differences between attenders and non-attenders are clearly stated and do not permit the assertion that the non-attenders are "more likely to have had significant amblyopia". The final outcome was assessed at a multidisciplinary clinic involving a number of researchers and promoted by providing transport costs and a gift for the child. The amblyopia B measure would naturally exclude children with very mild degrees of amblyopia. Whether treatment of such mildly affected children is unnecessary or ineffective was not an issue we sought to address in this study. The hypothesis being tested was that children detected as amblyopic earlier would achieve better results than would children detected later and our data clearly support this hypothesis. Furthermore they indirectly provide the first evidence of which we are aware that treatment for amblyopia is effective, which also adds to the evidence base needed to assessing the utility of a screening programme.