The negative impact of amblyopia from a population perspective: untreated amblyopia almost doubles the lifetime risk of bilateral visual impairment

Short abstract

Van Leewen et al's work provides valuable data, moving us closer to determining whether preschool vision screening can be justified

Ten years ago, a systematic review on the effectiveness of preschool vision screening for amblyopia and related target conditions was published in the UK, the Snowdon and Stewart‐Brown report.¹ After summarising the scientific evidence available at the time, the authors came to the following conclusion: purchasers are advised against implementing preschool vision screening programmes and providers should consider discontinuing them. The response from the paediatric ophthalmology community was, of course, massive, and a lively international discussion ensued. Even though the initial currents in the debate were quite harsh towards the conclusion of the report and even towards the authors, in retrospect we should be very grateful that it was published. In the last few years, we have had the pleasure of reading an impressively large number of well‐designed scientific studies, significantly moving forward our knowledge about diagnosis, screening and treatment of amblyopia, as well as challenging some old “truths” (for a recent summary, see Holmes and Clarke²).

However, data on the negative impact of amblyopia from a population perspective have not been as abundant. One of the main problems identified by The Snowdon & Stewart‐Brown Report was the lack of evidence showing that the target conditions (amblyopia, significant refractive error and non‐cosmetically obvious squint) are disabling if left untreated. In discussing the rationale for preschool vision screening programmes, more results on possible associations between amblyopia and increased lifetime risk of visual impairment, as well as quality of life/utility measures for unilateral amblyopia, are required. The paper by van Leewen et al, published in this issue of the BJO (see page 1450), is therefore an important contribution.³ In a world with limited economic resources and ever‐growing expenses for medical services, we will most likely see an increasing demand for evaluation, evidence of benefit and proof of cost‐effectiveness for government‐financed screening programmes.

In a paper on the cost‐effectiveness of screening for amblyopia published in 2004, König and Barry⁴ conclude that the risk of losing the better eye would not alone justify vision screening. The numbers used for their calculations come from a study by Rahi et al⁵ where the projected lifetime risk of vision loss (worse than 6/12) for an individual with amblyopia is reported to be 1.2%. Van Leewen et al report the lifetime risk of bilateral visual impairment (BVI) to be as high as 18% for those with amblyopia (compared with 10% for non‐amblyopic subjects), and that individuals with amblyopia on average spend an extra seven months with BVI. The numbers reported by van Leewen et al are considerably higher than those reported by Rahi et al. The two studies differ in methodology, which may explain these differences. The study by Rahi et al was a surveillance study, while The Rotterdam Study was population‐based. Moreover, the Rotterdam Study includes older age groups in whom the risk is higher. One cannot help but wonder if the conclusion from König and Barry⁴ would have been different using the numbers presented by van Leewen et al.

In discussing the cost‐effectiveness of screening and treatment for amblyopia, the possible functional disability related to unilateral amblyopia must also be taken into consideration. In economics, utility is a measure of relative happiness or satisfaction. Utility measures are preference‐based estimates of health‐related quality of life. These measures are increasingly being used in economic evaluations of healthcare interventions. In the last 5 years, a number of studies have been published on cost–utility analyses of treatment and screening for amblyopia,⁴,⁶,⁷,⁸ all with the limitation that utility/quality of life measures were not included,⁸ or derived from studies on individuals with impaired vision due to causes other than amblyopia.⁴,⁶ There is reason to believe that amblyopia may be very different from other ophthalmic diseases. Individuals with acquired visual disorders, many of which are progressive, previously have had healthy eyes and normal vision. Moreover, in many ophthalmic diseases, there are additional visual problems such as reduction of visual field and reduced colour vision, which those with amblyopia do not have. There is a clear need for objective studies on the possible relationship between unilateral amblyopia and functional disabilities. In such studies, comparisons should be made between three groups: (1) normal controls, (2) non‐treated amblyopes, and (3) amblyopes after successful treatment. Cost–utility studies assume that those with successfully treated amblyopia have the same utility value as normal controls. If it is shown that amblyopia is related to some kind of disability or loss of utility, then it is important to establish whether successful treatment reduces this disability or utility loss. At the recent 2007 ARVO meeting, Webber et al presented a study on the effect of amblyopia on motor and psychosocial skills in children.⁹ The study compared normal controls with children who had been diagnosed and treated for amblyopia, the latter group consisting of both successfully and non‐successfully treated subjects. The results showed those with amblyopia to have poorer fine motor skills compared with control subjects. This finding was not related to inter‐ocular difference in visual acuity, suggesting that successful treatment does not restore normal function.

Van Leewen et al should be praised also for presenting results on the numbers needed to treat (NNT). NNT is defined as the number of patients who need to be treated in order to prevent one unfavourable outcome. According to their data, 12.5 individuals with amblyopia need to be successfully treated to prevent one case of BVI. This number is well in line with other healthcare interventions,¹⁰ such as calcium and vitamin D supplementation to prevent fractures in older people (NNT 15) or statin treatment post‐myocardial infarction to prevent major coronary events (NNT 11).

In order to evaluate if preschool vision screening is worth while from a population point of view, several issues need to be clarified.¹¹ First, the lifetime risk associated with untreated amblyopia must be established, both in terms of increased risk of bilateral visual impairment and in terms of quality of life. Second, the total cost of screening and treatment, and the actual benefit of a screening programme to the population,¹²,¹³ must also be assessed. As of today, we do not have all of the information required, but van Leewen et al's work has provided us with very valuable data, moving us closer to determining whether preschool vision screening can be justified.