Author: Horwood, A.M., Griffiths, H.J., Carlton, J. et al.
Geographical coverage: United States, The US, Europe and East Asia.
Equity focus: Children.
Study population: Children under 7 years of age subject to screening for amblyopia.
Review type: Other review
Quantitative synthesis method: Narrative synthesis
Qualitative synthesis method: Not applicable
Background: Amblyopia screening can target reduced visual acuity (VA), its refractive risk factors, or both. VA testing is imprecise under 4 years of age, so automated risk-factor photo-screening appears an attractive option. However, it is still unclear how children should be screened to achieve optimal visual outcomes while avoiding excessive costs, false referrals and unnecessary (or unnecessarily early) treatment.
Objectives: To consider photo-screening used in community services, focusing on costs, cost-effectiveness and scope of use, compared with EUSCREEN project Country Reports describing how photo- and automated screening used internationally.
The evidence suggesting that photo-screening decreases the prevalence of amblyopia or strabismus or enhances overall outcomes is not strong. Similarly, its cost-effectiveness compared to subsequent visual acuity (VA) screening is also not well-established. At present, the most economical approach appears to be a later, expert-conducted VA screening with the provision for a re-test prior to referral.
A total of 60 papers were included in the review. Most studies did not report local or national screening schemes as such. In the United States, some well-established research groups have published many papers developing the evidence base on the most effective photo-screening tests, their validation and refinement of referral criteria. 55% of the 370 abstracts reviewed were from the US, 17% from Europe and 14% from East Asia.
Of 370 abstracts reviewed, 55 reported large-scale community photo-screening projects. Five addressed cost-effectiveness specifically, without original data. Photo-screening was a stand-alone, single, test event in 71% of projects.
In contrast, 25 of 45 EUSCREEN Country Reports showed that if adopted, photo-screening often supplements other tests in established programmes and is rarely used as a stand-alone test. Reported costs varied widely and evidence of cost-effectiveness was sparse in the literature, or in international practice. Only eight (13%) papers compared the diagnostic accuracy or cost-effectiveness of photo-screening and VA testing, and when they did, cost-effectiveness of photo-screening compared unfavourably.
Authors conclude that currently, the most cost-effective option seems to be a later, expert VA screening with the opportunity for a re-test before referral.
The authors included studies that used or assessed the feasibility of photo-screening in unselected children under 7 years old in actual or potential community projects, specifically looking for papers where costs were explicitly mentioned. Exclusions were made for non-English publications, letters, editorials, conference or poster abstracts, duplicates, non-commercially available methods, test development, validation of tests and new equipment against each other or other refraction methods, refractive error prevalence studies, studies deriving or refining referral criteria, and studies on older and non-typically developing populations. Recently commercially available birefringence scanning for foveal fixation was not considered.
Online searches were conducted on Medline, EMBASE, PsycInfo via Ovid, CINAHL, and The Cochrane Library, dating back to 1980, on September 10, 2018. The University of Liverpool’s orthoptic literature database was used to search various orthoptic journals and associations, as well as the International Orthoptic Congress’s full paper conference proceedings not listed electronically. The search included full-text original articles written in English, excluding conference abstracts, opinion pieces, guidelines, and editorials.
One author carried out the primary title stage extraction and excluded publications that were not written in English, letters, editorials, conference or poster abstracts and duplicates. Abstracts from the remaining papers were screened by two reviewers independently of each other.
Selection was carried out iteratively using pre-determined inclusion/exclusion criteria. Any disagreements between reviewers were to be resolved by a third researcher at the full paper stage if disagreement remained.
Data extraction for the full-text studies was carried out by one author using a pre-prepared data extraction sheet using criteria agreed by other authors. A sample 20% were cross-checked by a second researcher. The checklist authors used to quality assess included studies that do not recommend a formal score, but suggest it is used as a guide to assess the rigour of the evaluation
Within the EUSCREEN project, Country Representatives have been collecting national and local data on vision and hearing screening and have completed extensive questionnaires, containing 126 questions on vision screening across nine domains. Reports prepared with detailed information on societal background, general screening programmes, vision screening programmes, protocols, outcomes and costs in each country. If photo-screening was being used, the EUSCREEN team requested additional details as necessary.
Applicability/external validity: Although the authors do not directly point out any shortcomings in their review, their findings clearly indicate that there is a lack of substantial research evidence concerning photo-screening, especially regarding its cost-effectiveness.
Geographic focus: The majority of the studies included in the review were carried out in the U.S. The authors did not discuss how their results might differ across various geographical regions, even though the EUSCREEN reports would have country-specific data available. However, they did not provide a list of the countries that submitted these reports.
Summary of quality assessment:
There are a number of limitations in the approaches used to identify, include and critically appraise relevant studies. The search was limited to items published in English and there is no evidence that the reference lists of included items were reviewed. While the use of a narrative synthesis to analyse the data was generally robust, data was not extracted by more than one author in the majority (80%) of instances. In addition, the included data was not discussed separately with relation to the quality assessment of included studies. For these reasons, a low confidence was attributed in the findings of this review.