Critical appraisal of Australian and New Zealand paediatric vision screening clinical practice guidelines using the AGREE II tool

ABSTRACT

Clinical Relevance

Vision disorders in children impact health-related quality of life, with early detection and intervention improving outcomes and educational performance. Eye health professionals should be aware of paediatric vision screening guidelines and their development to understand the components of local programmes and the differences in sensitivity and specificity between protocols.

Background

High-quality clinical practice guidelines (CPGs) for vision screening enable the early detection of common vision disorders; however, they require rigorous development to ensure optimal accuracy in detecting vision disorders, enabling timely interventions. This study evaluated the quality of available vision screening CPGs on vision screening of children in Australia and New Zealand.

Methods

A systematic search of academic databases, guideline databases, professional associations and Google search engines was conducted to identify relevant paediatric vision screening CPGs. Four independent reviewers used the Appraisal of Guidelines, Research and Evaluation (AGREE II) instrument to assess the quality of individual guidelines and scores were aggregated and reported as the percentage of the total possible score across the six AGREE II domains: scope and purpose, stakeholder involvement, rigour of development, clarity of presentation, applicability, and editorial independence.

Results

Initial 2,999 items were evaluated, with seven guidelines included. AGREE-II quality score agreement ranged from 43.3% to 95.8%. All guidelines scored >60.0% in the scope and purpose, however, most had poor scores of <26.5% in the rigour of development and <3.3% in editorial independence domains. All guidelines recommended screening using measures of habitual distance vision.

Conclusion

Of the guidelines developed for use in Australia and New Zealand, most guidelines scored poorly when assessed against the AGREE II tool, because of lack of editorial independence and rigour of development. Paediatric vision screening guidelines should prioritise systematic review of literature to inform practice and include statements regarding competing interests.

KEYWORDS:

• AGREE II
• children
• clinical practice guideline
• paediatric
• vision screening

Introduction

Vision screening is an important public health initiative for detecting childhood vision conditions, particularly in vulnerable or economically disadvantaged communities, as it delivers economical vision assessments and improves outcomes for common childhood vision disorders.¹ Untreated vision problems in children can lead to amblyopia,² poor vision from abnormal visual experience early in life, reduce quality of life through to adulthood,³ and negatively impact learning⁴ and future earning potential.³

Vision screening programs identify children with reduced vision and refer them for comprehensive assessment. The predominant causes of poor vision in childhood are refractive error, strabismus or amblyopia, however poor vision can be subsequent to ocular or systemic pathology.^2,3 Whilst identifying and treating vision disorders is a cost-effective health intervention for improving productivity and quality of life,³ there are challenges to effectively capturing a paediatric population and developing a screening programme that balances efficiency and cost-effectiveness with a high sensitivity for detecting paediatric eye conditions.

Most developed countries have some form of children’s vision screening or testing,⁵ but there is limited consensus regarding evidence-based screening protocols that are effective in detecting common vision disorders in children.⁶ Children’s vision screening programmes vary significantly in terms of their objectives, testing approach and personnel employed.⁷

Given the importance of identifying children at risk of amblyopia and other common vision disorders, a paediatric vision screening programme must have high sensitivity, but also high specificity to ensure appropriate delivery without overwhelming health care resources. Therefore, protocols and procedures should be rigorously developed, using the appropriate evidence and stakeholder engagement. Clinical practice guidelines (CPGs), which are statements developed to help people make clinical, system- and policy-related decisions, can vary widely in content and quality.⁸

The Appraisal of Guidelines, Research and Evaluation (AGREE) Collaboration established a generic instrument, the AGREE II tool, to assess the process of guideline development and quality.^5,6,9 With 23 items, AGREE II critically assesses published guidelines across six domains, and two global scores, looking at the overall quality of the guideline and its suitability for use with or without modification. The domains evaluate scope and purpose covering target population and overall aims; stakeholder involvement assessing the input from relevant groups; rigour of development analysing the links to the evidence and how recommendations were developed; clarity of presentation of the overall guideline; applicability including barriers to implementation; and editorial independence.⁹

Individual items within each domain are outlined in Supplementary Table S1. Using a 7-point response scale, with a maximum score of 7 (strongly agree) and a minimum score of 1 (strongly disagree), the instrument differentiates between high quality and low-quality guideline content. There is a correlation between favourable outcomes of guideline implementation and positive ratings in multiple AGREE domains with every 10% change in the rigour of development domain predicting a statistical improvement in clinician endorsement.⁹

Despite the prevalence of clinically significant refractive errors in children ranging from 13.0% to 30.0%,^10,11 and amblyopia prevalence reported as between 1.9% and 6.1%^2,10 there has been no systematic investigation of the quality of paediatric vision screening guidelines developed for in Australia and New Zealand. Therefore, the aim of this study was to review the quality, current evidence, and recommendations from published paediatric vision screening CPGs in Australia and New Zealand using the AGREE II tool.

Methods

Search strategy

A single, systematic literature search for published vision screening guidelines from Australia and New Zealand in academic databases (Medline, PsycINFO, Web of Science Core Collection, CCENTRAL, LILACs and ProQuest Scopus) was conducted in January 2023, using pre-defined selection criteria and search strategies outlined in Table 1 and Supplementary Table S2. Additionally, guideline databases and optometry and ophthalmology professional society websites (Supplementary Table S3) were searched using the terms and criteria in Supplementary Table S3.¹²

Table 1. Criteria used for systematic search and selection of the clinical practice guidelines (CPGs) and exclusion criteria.

Inclusion criteria
Clinical Practice Guideline (CPG)	Clinical practice guidelines (‘guidelines’) are systematically developed statements to assist practitioner and patient decisions about appropriate health care for specific clinical circumstances
Type of participants	Children from 0–18 years of age, or guidelines that included vision screening for both adults and children (but only the paediatric section was addressed)
Type of intervention	Only vision screening guidelines with a screening or testing protocol and referral criteria were included. Guidelines that were designed to provide guidance on comprehensive eye and vision examinations in children were excluded.
Date of Publication	From January 2008 – January 2023
Exclusion criteria
Title and Abstract Screening	Identified literature did not include the words ‘guideline’ or ‘protocol’ The guideline was not published in the last 15 years The guideline was not in English
Full Text Retrieved	The guideline was not developed for vision screening The target population did not include patients between the ages of 0–18 years The CPGs were not developed for use in Australia or New Zealand
Full Text Screening	The guideline described comprehensive vision examinations and not a vision screening protocol Did not specifically screen for refractive error, strabismus, amblyopia risk factors or reduced visual acuity in the specified age group Guidelines that screened for other visual conditions such as retinopathy of prematurity (ROP), retinoblastoma, diabetes, juvenile arthritis etc.

Finally, a detailed Google search was undertaken to identify any relevant CPGs in public domains using specific search terms, as outlined in Supplementary Table S3.¹³ The Google search was terminated for each dataset when 10 consecutive irrelevant results were found. All searches were limited to the last 15 years and only to English language CPGs, to ensure evidence was current.

Clinical practice guideline selection

Two authors (CH and RC) independently screened the titles and abstracts, where available, of items identified from the systematic literature, Google, professional society, and guideline database searches. CPGs that met the exclusion criteria outlined in Table 1 were removed. CPGs that met title screening underwent independent full text screening by CH and RC.

Guidelines that did not meet the inclusion criteria outlined in Table 1, or were not based in Australia and New Zealand, were excluded. Any discrepancies were resolved by discussion between the two authors conducting the full text screening or, in the event a consensus could not be reached, by discussion with a third author (NA).

Quality appraisal and analysis

The quality of the included guidelines was assessed by four independent reviewers (CH, KK, JH and NM) using all 23 key items across the six domains and the two global ratings of the AGREE II tool. Differences of more than two points for any item were resolved through discussion within the appraisal group or, in the event a consensus could not be reached, by discussion with a fifth reviewer (NA).

A quality score was calculated for each of the six AGREE II domains by summing individual item scores and by scaling the total as a percentage of the maximum possible score for that domain,¹⁴ with the minimum possible score for each domain of one and the maximum score for each domain of seven. Each domain score was calculated by obtaining the sum of scores from all reviewers, then producing an agreement percentage using the equation⁹:

$\frac{\mathit{obtained}\mathit{score}-\mathit{minimal}\mathit{possible}\mathit{score}}{\mathit{maximal}\mathit{possible}\mathit{score}-\mathit{minimal}\mathit{possibe}\mathit{score}}.\mathit{\hspace{0.17em}}\mathit{\hspace{0.17em}}$

A good-quality guideline was defined as the one with ≥60.0% score for the rigour of development, the scope and purpose, and at least one other domain.¹⁵

Data on recommendations were extracted from included CPGs using a standardised online form. Key vision screening recommendations including the tests used for screening, age of screening, referral criteria, individuals involved in the screening programme, how participants were notified about the outcomes, and the preferred referral pathways were extracted by one author (CH), and independently checked by a second author (NA).

Results

Search results

Figure 1 shows the number of items screened, assessed for eligibility, and included in the final review. Initially, 2,999 results were identified, with only 115 (3.8%) remaining after title and abstract screening. Systematic literature search of published vision screening guidelines yielded no results. All CPGs were found on guideline databases, professional association websites, or general websites. Full text screening further excluded 59 items due to relevance and 47 because they were not developed for vision screening in Australia or New Zealand, leaving nine potential CPGs.

Figure 1. Preferred reporting items for systematic literature search, search through guideline databases and professional association websites, and Google search.

After a detailed appraisal, an additional two CPGs were excluded; one because it was not a vision screening guideline, and the other because it had been replaced by a newer version. This left seven guidelines selected for quality appraisal using the AGREE II tool: Western Australia (WA) Clinical Nursing Manual,¹⁶ the Victorian Maternal and Child Health Service Practical Guidelines and Supplementary material Melbourne Initial Screening test,^17,18 Vision 2020 National Framework for Vision Screening (Vision 2020), Lions Eye Health Program Children’s Vision Screening Screener Manual,¹⁹ New Zealand Well Child Programme and supplementary material,^20,21 the B4 School Check (B4 School Check)²² and the National Vision and Hearing Screening Protocols (New Zealand National Protocols).²³

Guideline characteristics

The seven guidelines, four from Australia and three from New Zealand, were published between 2008 and 2023. Five guidelines were commissioned by government bodies, and two were published by charities with an interest in vision screening (Table 2). Two CPGs also included supplementary materials.

Table 2. Characteristics of the seven clinical practice guidelines (CPGs) selected for quality appraisal using the AGREE II tool.

Guideline Name	CPG Name	Institution/Guideline Development Group	Year	Country	Screeners
WA Clinical Nursing Manual, Vision and Eye Health^Citation16	WA Clinical Nursing Manual	Government of Western Australia, Child and Adolescent Health Service^Citation16	2021	Australia	Nurse Practitioners
Victorian Maternal and Child Health Service, Practical Guidelines Supplementary material: Melbourne Initial Screening test (MIST) booklet^Citation17,Citation18	Melbourne Initial Screening Test	Victorian Maternal and Child Health Service^Citation17,Citation18	2009	Australia	Nurse Practitioners and screeners
Vision 2020 National Framework for Vision Screening^Citation24	Vision 2020	Vision 2020^Citation24	2020	Australia	Anyone conducting vision screening
Lions Eye Health Program, Children’s Vision Screening Screener Manual^Citation19	Lions Eye Health	Lions Club^Citation19	2022	Australia	Trained Layperson
New Zealand Well Child Tamariki Ora Programme Supplementary material: Brief Evidence Reviews for the Well Child Tamariki Ora Programme^Citation20,Citation21	New Zealand Well Child	Ministry of Health, Manatu Hauora^Citation20,Citation21	2014/2019	New Zealand	Nurse Practitioners and Trained Layperson Vision Hearing Technician
The B4 School Check^Citation22	B4 School Check	Ministry of Health, Manatu Hauora^Citation22	2008	New Zealand	Registered nurses, vision technicians
National Vision and Hearing Screening Protocols^Citation23	New Zealand National Protocols	Ministry of Health, Manatu Hauora^Citation23	2021	New Zealand	Vision Hearing Technicians

CPG quality assessment using the AGREE II tool

Domain-specific results for all guidelines are summarised in Table 3 and Figure 2. The mean (± standard deviation) overall percentage score for all the guidelines was 69.0 ± 15.3%. The New Zealand Well Child and Supplementary material: Brief Evidence Reviews for the Well Child Tamariki Ora Programme CPG had the highest overall score of 95.8%, and the Lions Eye Health Program had the lowest overall score of 43.3%. The only guideline that met the accepted standard for a high-quality guideline with ≥60.0% in the scope and purpose, rigour of development and one other domain¹⁵ was the New Zealand Well Child CPG.

Figure 2. Radar Chart showing individual AGREE II domain scores (%) for the seven selected clinical practice guidelines (CPGs).

Table 3. Individual AGREE II domain scores (%) for the seven selected clinical practice guidelines (CPGs). Percentage scores for each domain were calculated by summing individual item scores and by reporting the total as a percentage of the maximum possible score for that domain.

Guideline name	Scope and Purpose Domain (%)	Stakeholder Involvement Domain (%)	Rigour of Development Domain (%)	Clarity of Presentation Domain (%)	Applicability Domain (%)	Editorial Independence Domain (%)	Overall Score (%)
WA Clinical Nursing Manual, Vision and Eye Health^Citation16	86.1	33.3	19.1	64.8	22.9	0.0	69.4
Victorian Maternal and Child Health Service, Practical Guidelines Supplementary material: Melbourne Initial Screening test (MIST) booklet^Citation17,Citation18	93.3	50.0	20.0	65.6	32.5	3.3	66.7
Vision 2020 National Framework for Vision Screening^Citation24	91.1	60.0	21.7	83.3	72.5	18.3	66.3
Lions Eye Health Program, Children’s Vision Screening Screener Manual^Citation19	68.9	25.6	2.9	42.2	22.5	0.0	43.3
New Zealand Well Child, Tamariki Ora Programme Supplementary material: Brief Evidence Reviews for the Well Child Tamariki Ora Programme^Citation20,Citation21	95.8	63.9	83.3	87.5	47.9	60.4	95.8
The B4 School Check^Citation22	97.2	44.4	22.4	84.7	63.5	0.0	70.8
National Vision and Hearing Screening Protocols^Citation23	95.8	33.3	8.9	86.1	29.2	0.0	70.8
Average	89.8	44.4	25.5	73.5	41.6	11.7	69.0
Standard deviation	9.9	14.5	26.5	16.8	20.1	22.5	15.3

All other reviewed CPGs scored poorly in the rigour of development domain with scores ≤22.4%. The editorial independence domain also scored poorly, with five CPGs scoring <3.3%. Conversely, the scope and purpose domain scored well, with all CPGs scoring >60.0%, and the clarity of presentation domain scored >60.0% for all but one CPG.²⁵

The final two questions of the AGREE II tool ask the assessors to rate the overall quality of the guideline and whether the assessor would recommend the guideline for use, with or without modification. Although no CPGs were considered unsuitable for use by the examiners, modifications were recommended in most of them. Only the New Zealand Well Child CPG had 100.0% agreement from examiners that it should be used, while 60.0% of examiners agreed the Melbourne Initial Screening Test CPG should be used without modification. There was agreement that the B4 School Check, the WA Clinical Nursing Manual CPG, and the Lions Eye Health CPGs could be used, but with modifications.

Results of the individual AGREE II domains

Domain 1 scope and purpose was the best scoring domain, with a mean ± SD score of 89.8% ± 9.9%. All CPGs clearly outlined the screening population, the conditions targeted and why vision screening is an important health intervention for these populations. Clarity of presentation (Domain 4) was the next best scoring domain, with a mean score of 73.5% ± 16.8% with six out of seven CPG scoring above 60.0% agreement. Overall, the CPGs scored poorly (44.4% ± 14.5%) in the stakeholder engagement domain due to lack of engagement with professional groups, and minimal consultation with schools, teachers, parents, and children on the screening programme.

Likewise, the mean score for the applicability of the guideline (Domain 5), which examined funding, barriers to guideline application, and monitoring or auditing criteria, was low (41.6% ± 20.1%) with five of the seven CPGs (WA Clinical Nursing Manual, Melbourne Initial Screening Test, Lions Eye Health, the B4 School Check, and the New Zealand National Protocols) achieving the lowest possible score for monitoring and auditing CPG protocols, recommendations or outcomes.

The two poorest scoring domains were rigour of development (Domain 3, mean score 25.5% ± 26.5%) and editorial independence (Domain 6, mean score 11.7% ± 22.5%). Domain 3 represents the largest section of the AGREE II tool, with questions addressing the evidence basis behind the development of the CPG, how evidence was selected and how guidelines will be reviewed and updated in the future. Most CPGs had limited protocols for updating the guideline and lacked linkages between protocol selection and the evidence base used. The editorial independence domain was the worst scoring of the six, as most guidelines, except for the New Zealand Well Child and Vision 2020, scored 0–3.3% for this section.

Clinical recommendations

Table 4 summarises the key recommendations from each of the CPGs, all of which recommend conducting paediatric vision screening by measuring habitual distance vision. Most guidelines recommended that children should be referred for a comprehensive eye examination when acuity was 0.3 logMAR (Snellen equivalent 6/12) or worse, with the guidelines from Aotearoa New Zealand also identifying the need to rescreen children who had visual acuity of 6/9 (0.175 logMAR) in the worse eye and 6/6 (0.0 logMAR) in the better eye.

Table 4. Clinical recommendations from the selected clinical practice guidelines (CPGs).

Clinical Guideline	Age of Screening	Recommended VA test	VA Referral Criteria (for one or both eyes)	Who performed screening	Supplementary tests	Referral Pathway
WA Clinical Nursing Manual, Vision and Eye Health^Citation16	4 years	Lea Symbols Chart 3 m^39	6/12 refer 6/19 Urgent refer	Nurses	Red reflex, cover test, corneal light reflex	Non-urgent referrals are re-checked within 3 months. Urgent referrals to ‘medical practitioner’
Victorian Maternal and Child Health Service, Practical Guidelines/Melbourne Initial Screening test (MIST) booklet^Citation17,Citation18	3–5 years	MIST Letter matching chart 3 m^Citation17,Citation18	3/5 (Snellen equivalent approx. 6/9.5)	Nurses	n/a	Ophthalmologist, eye clinic at a public hospital, optometrist, orthoptist
Vision 2020 National Framework for Vision Screening^Citation24	3.5–5 years	Crowded or linear optotypes e.g., HOTV, Lea^Citation26	6/15 refer 6/18 urgent refer	Lay screeners, nurses and teachers	n/a	Ophthalmologists, optometrists, orthoptists
Lions Eye Health Program, Children’s Vision Screening Screener Manual^Citation19	Pre-school and school-aged children	Lea symptoms chart 3 m^39	6/12 refer	Lay-person who has completed a screening course	‘Quick 6’- Basic colour vision screening, Stereo Fly, Spot Vision Screener	Optometrist or Ophthalmologist
New Zealand Well Child Tamariki Ora Programme/Brief Evidence Reviews for the Well Child Tamariki Ora Programme^Citation20,Citation21	4–5 year-old 11–12 year-old	Parr 4 m letter-matching with confusion bars (or equivalent Sheridan Gardner)^Citation27	6/9 + 6/6 Re-screen 6/12 Refer	Vision and Hearing Technicians, or other competent health practitioner	None	Ophthalmologist or Optometrist
The B4 School Check^Citation22	4 year-old	Snellen chart, or single-letter charts with crowding bars at 4 m^39	6/9 + 6/6 re-screen 6/12 refer	Registered nurses, vision and hearing technicians	None	Orthoptist
National Vision and Hearing Screening Protocols^Citation23	4 year-old	Parr 4 m letter matching, Snellen chart^Citation27	6/9 + 6/6 rescreen 6/12 refer	Vision and Hearing Technician	Ishihara	Optometrist

Most CPGs recommended screening at 4–5 years of age, with some offering repeat screening at different ages. There was no consistency in recommendations between the visual acuity chart used, with some CPGs recommending letter-based charts, while others recommended picture optotypes. Only the Vision 2020, New Zealand Well Child, B4 School Check and New Zealand National Protocols specifically mentioned using vision charts with crowding.

Three of the seven guidelines also included the use of supplementary tests to screen for colour vision defects,^19,23 strabismus,^16,19 refractive error¹⁹ and ocular pathology.¹⁶ All CPGs recommended that the vision screening be conducted by nurses or lay screeners with children who failed the vision screening referred to orthoptists, optometrists, or ophthalmologists.

Discussion

This study identified seven paediatric vision screening guidelines that have been developed for use in Australia and New Zealand, although none of these guidelines had been published in peer-reviewed journals. Most guidelines scored poorly when assessed against the AGREE II tool, primarily due to lack of editorial independence and poor rigour of development. Only the New Zealand Well Child CPG was considered as a good guideline as it had AGREE II scores >60.0% for the rigour of development, the scope and purpose, and at least one other domain.²⁵

Most of the guidelines reviewed in this study were rated as poor to moderate quality based on quantitative scores, which was also reflected in the appraisers' overall qualitative rating based on the final two holistic AGREE II questions. All assessors agreed that the only guideline suitable to use in its current form was the New Zealand Well Child check with all other guidelines requiring some modifications. In general, most guidelines could be improved by including a well-designed systematic literature review and justifying how the evidence base had been used to develop the CPG recommendations.

Furthermore, explicit information regarding the editorial board and guideline developers, as well as a conflict of interest disclosure were missing from most guidelines. The AGREE II tool recommends explicit discussion on any conflict of interest from members of the development group, and clear links between funding bodies and the guideline content to ensure clear editorial independence in the interests of best practice for patients.⁹ Finally, further stakeholder engagement is warranted during the development process to ensure patient-centred care and applicability in a clinical setting.

Clinical practice guidelines are designed to optimise patient care and assist health care practitioners with clinical decision-making and should be informed by a systematic review of the evidence, therefore Domain 3 rigour of development and Domain 6 editorial independence are particularly critical during the CPG development process.²⁸ The rigour of development domain is regarded as having the most direct effect on CPG quality as it captures the need for a systematic review of the literature and the use of a clear evidence base for guideline development.²⁹

Alonso-Coella et al. noted that the leading deficiencies in this area of guideline development were lack of methodological expertise and a lack of information provided about the methods used to influence guideline development.²⁹ The authors therefore recommended involving librarians and guideline methodologists with expertise in methodological design, clinical epidemiology and statistics in guideline development, and to specifically include information about search strategies, literature selection processes and systems used to evaluate the quality of evidence as an addendum. The New Zealand Well Child CPG scored well in Domain 3 through the associated brief evidence review that was published by paediatric vision researchers in 2019,²¹ 6 years after the CPG was first developed, which provides the evidence base to support the screening protocol.

However, the lack of evidence base utilised in developing these paediatric vision screening CPGs may be reflected in the range of screening tests suggested. While all guidelines recommend measuring distance vision as the primary screening technique, the recommended acuity chart varies between different screening guidelines with both letter matching (Melbourne Initial Screening Test chart,^17,18 Parr 4 m^20,21,23 and HOTV²⁴ chart) and picture optotype (Lea symbols^16,24) at either three or 4 m suggested.

Some guidelines (New Zealand Well Child, B4 School Check and New Zealand National Protocols) recommend a re-screen when vision was 6/9 in one eye and 6/6 in the fellow eye, with other guidelines recommending referral with VA worse than 6/12 (0.3 logMAR) or 3/5 (0.2 logMAR for Melbourne Initial Screening Test), other than the Vision 2020 with a referral criterion of 6/15.

Recent evidence by Findlay et al. found that removing restrictions around the two-line inter-eye difference improved the sensitivity of the screening, with no reduction in specificity for amblyopia risk factors.²⁷ Acute or urgent referral is recommended by the WA Clinical Nursing Manual at 6/19 and the Vision 2020 at 6/18. Almost all programmes other than the B4 School check, which suggests orthoptist referral, and the WA Clinical Nursing Manual, which recommends ‘medical practitioner’ referral, recommend referral to optometrists and ophthalmologists (Table 4). Only the Lions Eye Health programme recommended the use of autorefraction (Spot Vision Screener) as an adjunct test, despite the literature suggesting that handheld autorefractors demonstrate higher specificity and equivalent sensitivity to Lea symbols, when screening for significant refractive error and amblyopia risk factors.^30,31

The WA Clinical Nursing Manual, Lions Eye Health and New Zealand National Protocols CPGs all recommend supplementary testing including cover test, corneal light reflex test, stereoacuity, and colour vision screening, however there is little published evidence to support the use of these procedures. While the inclusion of cover test (in screenings conducted by eye health professionals) or stereopsis (when the screening is conducted by lay screeners) improves the sensitivity of detecting strabismus,³⁰ measuring stereoacuity alone does not increase the sensitivity for detecting amblyopia, except when combined with autorefraction, which may increase the efficacy of screening protocols.³²

There was broad general agreement across all guidelines recommending screening children between 3 and 5 years of age, with the aim of identifying amblyopia and amblyopia risk factors such as strabismus, anisometropia and significant refractive error.³³ Screening children at younger ages was discussed only in the WA Clinical Nursing Manual, and repeat vision screening of school-aged children only referenced in the New Zealand Well Child and the Lions Eye Health CPGs. These CPG referral recommendations are broadly consistent with international best practice,³⁴ which shows there is limited evidence for screening children younger than 3 years,³⁴ and most children with vision worse than 0.2 logMAR require refractive correction.³⁵

All the seven CPGs recommend that vision screening should be conducted by nurses or lay screeners. Nurses could achieve high sensitivity and specificity with appropriate training.³⁶ The Vision in Preschoolers Study compared the performance of different vision screening tests in 3- to 5-year-old children by trained nurses and lay screeners, using the results of cycloplegic eye examinations administered by an ophthalmologist or optometrist as the gold standard.^30,37

Although the sensitivities for all tests, including Lea symbols, autorefractor and stereoacuity testing, were marginally higher when administered by nurse screeners compared to lay screeners, these differences were not statistically different. The study concluded that nurse and lay screeners can achieve similar sensitivity to optometrists and ophthalmologists for detecting vision disorders in preschool children, when specificity is set at 0.90.³⁷ It is recommended that a collaborative effort of eye-care professionals, nurses and lay screeners may be an ideal and cost-effective approach for vision screening in paediatric populations.⁵

A major strength of the study is that it included a comprehensive literature review encompassing guideline databases, professional association websites, and Google searches of all available paediatric vision screening CPGs in the last 15 years. Another strength of the study was the inclusion of four independent reviewers as recommended by the AGREE II protocol, which improves the reliability of scores in each individual domain.³⁸

A limitation of this study is that the AGREE II tool does not assess the quality of the guideline recommendations, and some studies have found that high AGREE II scores do not guarantee optimal CPG recommendations.³⁹ Furthermore, given the nature of guidelines, there may be other CPGs that met our inclusion criteria but were not published in literature or available publicly on the internet. Comparison of guidelines from other geographical areas may have strengthened the study, but due to the diverse nature of eye health care practices internationally including the different status of eye-care professionals, this study focused only on paediatric vision screening guidelines in Australia and New Zealand, which share similar health care systems through trans-Tasman mutual recognition.

Conclusion

This review identified seven paediatric vision screening CPGs developed in Australia and New Zealand, all of which were found on guideline databases, professional association websites or through a detailed Google search. The majority of these paediatric vision screening CPGs scored poorly on the AGREE II tool, primarily due to poor scores in the domains of editorial independence and rigour of development.

Conversely, most CPGs scored well on the scope and purpose and the clarity of presentation domains on the AGREE II tool. The highest scoring CPG was the New Zealand Well Child check, which scored >60.0% in six AGREE-II domains, including rigour of development. There were significant variations in the recommendations from these CPGs regarding referral criteria for comprehensive eye examination, the age of screening and the tests for screening refractive errors and other vision disorders.

These findings suggest the need for the development of a robust evidence-based paediatric vision screening guideline for use in Australian and Aotearoa New Zealand through improved editorial independence, increased rigour of development by incorporating current evidence and published literature and better stakeholder engagement.

Acknowledgements

We would like to thank our librarian, Ms Shannon Brown, for helping us with the literature search.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08164622.2024.2339276.