Diagnostic accuracy of vision screening tests for the detection of amblyopia and its risk factors: a systematic review

Methodological quality of the review: High confidence

Author: Schmucker C, Grosselfinger R, Riemsma R, Antes G, Lange S, Lagrèze W, Kleijnen J

Region: United Kingdom (UK), United States of America (USA), Canada, Sweden and Israel.

Sector: Vision screening

Sub-sector: Amblyopia, preschool vision screening, diagnostic test accuracy.

Equity focus: Children aged up to six years old

Review type: Effectiveness review

Quantitative synthesis method: Narrative analysis

Qualitative synthesis methods: Not applicable

Background

Amblyopia and its risk factors such as strabismus and refractive errors are the most common visual disorders in children. Across different ethnic groups and a wide range of western and developing countries, the prevalence rate of amblyopia ranges between 0.3% and 3%.

Currently, there is a variety of recommendations for identifying amblyopia or its risk factors in children. Despite the large number of screening tests available, little is known about the effectiveness and diagnostic accuracy of these tests. To be effective, a screening test must identify both a high proportion of children who have the high target condition (high sensitivity) and high proportion of children without any visual disorders (high specificity).

Research objectives

To evaluate the diagnostic accuracy of pre-school vision screening tests for the detection of amblyopia and its risk factors.

Main findings

Two studies with a longitudinal design and 25 cross-sectional studies met the inclusion criteria. Studies included data from the USA, the UK, Canada, Sweden and Israel.

One of the longitudinal studies compared a screening programme in children between one and two years of age with a re-examination at the age of eight. The sensitivity for the screening programme was 86% (range: 64–97%) and the specificity 99% (range: 98–99%). The second longitudinal study compared screening examinations at eight, 12, 18, 25 and 31 months, with a re-examination at the age of 37 months. In this study, the sensitivity of the screening examination increased with age, while the specificity remained unchanged. The cross-sectional studies evaluated different screening settings, visual acuity tests, auto- or photo-refractors and stereo tests. A large variety of reference tests, differing criteria for defining amblyopia and its risk factors and methodological limitations of the studies prevented a valid data interpretation.

The authors noted that diagnostic test accuracy of pre-school vision screening tests can only be sufficiently investigated after establishing age-related values defining amblyopia, refractive errors and binocular disorders. To address these questions, authors recommended a controlled longitudinal study design.

Methodology

The authors included studies where: children up to the age of six years old from the general population participated; sufficient information to calculate test accuracy of the screening test was provided; any reference test was included; and no restrictions were applied in terms of study design.

A search was conducted in nine databases including Medline, Embase, CINAHL, PsycINFO and Cochrane Central from inception until 2007. No restriction to specific language was applied. The searches were supplemented by searching the bibliographies of included studies manually. Additionally, the authors contacted manufacturers of screening tests for clarification.

Two authors independently screened articles for inclusion, extracted data and assessed the methodological quality of included studies. The authors used 2×2 tables to calculate test sensitivity and specificity. The methodological quality of studies was assessed using the Quality of Diagnostic Accuracy Studies (QUADAS) checklist.

Authors used a narrative approach to analyse findings and did not perform a meta-analysis due to the heterogeneity of the included studies.

Applicability/external validity

The authors did not use any model or approach to improve the applicability/external validity of the findings nor did they discuss the generalizability of the results.

Geographic focus

The review did not focus on specific income settings. Authors only found studies from high-income settings, therefore included in the review. Nevertheless, findings of the study may be applicable to low- and middle-income settings, as all previous systematic reviews conducted agreed that a consensus needs to be reached for the definition of amblyopia and its risk factors.

Quality assessment

Overall, there is high confidence in the conclusions about the effects of this study. This review is based on comprehensive searches of the literature to ensure all relevant studies were included in the review, avoiding language bias. The authors used appropriate methods to select studies, and extract and quality assess included studies, avoiding risk of bias. Authors used a narrative approach to analyse the findings, which seems appropriate due to the heterogeneity of included studies. Reviewers acknowledged the lack of well-established gold standards to detect children with amblyopia or its risk factors which may therefore lead to large differences in the prevalence of visual impairments. As such, no strong policy conclusions were drawn.