Kay Picture Test referral criteria
Best practices for pre-school vision screening
When setting up pre-school vision screening programmes, establishing clear vision screening pathways and referral thresholds is essential. Doing so ensures children who need follow-up care are correctly identified while keeping false positives and false negatives balanced.
Clinicians often ask orthoptist Hazel Kay how to convert Kay Picture Test normative data into an effective vision screening protocol.
Optimising distance visual acuity screening for children aged 3 years 10 months to 5 years involves complex factors. Below is a summary of key factors to consider in community child health settings.
1. Distance visual acuity screening protocols – prioritise linear presentation
To obtain accurate and repeatable visual acuity results during vision testing in pre-school children, the protocol should mirror clinical assessments that orthoptists or paediatric optometrists perform as closely as possible.
Linear presentation: We recommend a linear format of pictures for this age group. Using our Kay Picture Test Linear Crowded Book, screeners should present five pictures at and around the smallest size the child can see.
Single crowded test: The linear test is faster for routine screening. However, our Single Crowded Book is a valuable alternative for children who find full lines overwhelming. It isolates a single optotype which crowding bars surround. This provides an accurate, crowded logMAR measurement. It also keeps the child’s focus on one target at a time.
Keeping engagement high: Using tools such as our Screening Book, Screening Chart, or iSight Test Pro app allows screeners to move quickly through larger pictures to build the child’s confidence. This leaves sufficient time for careful checking at smaller optotype sizes without causing the child to become frustrated or bored.
2. Deciding referral criteria and pass levels
A common question among child health nurses and screening teams is how to set visual acuity referral thresholds.
Normative data and the lower limit: Our published normative research data provides a baseline for expected vision levels in children. Screening teams often set pass levels at the lower limit, which represents the vision level that 97.5% of children achieve when researchers confirm normal vision during studies.
Balancing false positives and negatives: Any screening programme will encounter a trade-off between false negatives (missing a condition) and false positives (referring a child unnecessarily). Choosing the exact cutoff depends on budget/cost constraints and local programme goals.
3. Establishing local referral pathways and criteria
When establishing local referral pathways, the receiving eye professionals’ opinion on referral criteria is paramount. Screening administrators should prioritise early consultation with these professionals (including hospital, community, and optometry providers) and evaluate:
Target population: Consider the specific demographic and needs of the population that screeners assess.
Screening environment: Determine if the testing environment is quiet and free of distractions to allow full concentration.
Frequency of testing: Distinguish whether the test is a single, one-time screening opportunity or part of a multi-stage vision screening series. A one-time vision screening service may set a higher bar for passing and accept more false positives so that so that screeners miss fewer children.
Local standards: Align with the level of vision that the local community health framework considers acceptable.
4. Don’t forget near vision assessment
While many vision screening pathways focus primarily on distance screening, evaluating near vision is highly recommended.
Young children possess a strong ability to over-focus (accommodate) to overcome significant hypermetropia (long-sightedness) over brief periods. As a result, a child might successfully pass a distance visual acuity test but fail to maintain clear vision during near tasks. Incorporating a near vision check helps capture hypermetropic errors that distance testing alone might miss.
