Preschool vision screening.

OBJECTIVES To undertake a systematic review of the effectiveness of preschool vision screening. To provide evidence on which decisions about the future provision of this service can be made. To indicate areas for further research. STUDY SELECTION The Centre for Reviews and Dissemination guidelines for systematic reviews were used. The research questions were formulated using the Wilson and Jungner criteria for evaluating screening programmes. They concerned prevalence, natural history, disability, treatment and screening in relation to three target conditions: amblyopia, refractive errors and squints which are not cosmetically obvious. Studies were considered for inclusion according to pre-determined criteria for the age group studied, the outcomes measured and the study design. The following types of study design were considered: cross-sectional studies of prevalence, cohort studies of natural history, any type of study (e.g., cross-sectional surveys, case-series, qualitative studies) of disability attributable to a target condition, controlled trials, observational studies and audits of screening programmes, and prospective controlled trials of treatment. DATA SOURCES The following electronic databases were searched: Biological Abstracts, CINAHL, Embase, ERIC, IAC Health Periodicals, IAPV, Medline, Psychlit, Science Citation Index, System for Information on Grey Literature in Europe, DHSS-Data, Faculty of Public Health Medicine Database of Dissertations, Index of Scientific and Technical Proceedings, Dissertation Abstracts, Index of Theses, NHS Research Register, Public Health Information Sharing Database. A limited amount of handsearching was undertaken. Reference lists were scanned to identify other relevant studies, and requests for unpublished data were made to people working in the field. DATA EXTRACTION Data was extracted by the first author and then checked by the second. DATA SYNTHESIS Quantitative analysis was undertaken where possible. Qualitative analysis was performed where studies were too heterogeneous for the data to be combined, or for research questions that were not suitable for quantitative synthesis. RESEARCH FINDINGS The electronic search yielded over 5000 references, and over 500 abstracts were downloaded from the databases for further scrutiny. A total of 85 studies were included in the main analysis. PREVALENCE No studies were found with the primary aim of establishing the prevalence of visual defects in preschool children. Data from studies of screening programmes report a range of yields for all the target conditions combined of 2.4-6.1%. NATURAL HISTORY: No studies designed with the intention of documenting the natural history of the target conditions in children aged 3 or 4 years were found. Other studies that provide some natural history data suggest that mild degrees of amblyopia may resolve spontaneously. In the absence of information about natural history it is impossible to estimate the effect of treatment from studies without a control group that was not treated. DISABILITY: A total of 21 studies exploring disability in relation to the target conditions were included. The literature provides a reasonable basis for generating plausible hypotheses about the ways in which the target conditions might disable people, but is insufficient to draw any firm conclusions about their impact on quality of life. The research to date is not sufficient to determine appropriate outcomes for controlled trials of treatment. TREATMENT Five randomised controlled trials of treatment and six prospective controlleld trials without randomisation were found. No studies compared treatment with no treatment. Most of the studies were methodologically flawed.(ABSTRACT TRUNCATED)