Screening for amblyopia in childhood
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The primary objective of this review will be to evaluate the effects of vision screening in reducing the prevalence of amblyopia in comparable screened versus unscreened populations. Subgroup analyses are planned, if appropriate, to determine the effect of the type of personnel conducting the testing, the age at screening, and the threshold applied for failure. Secondary objectives will be to report available evidence regarding the disability associated with living with uncorrected amblyopia and to document reports of the potential harms and costs associated with screening.
Background
Introduction
Amblyopia (lazy eye) can be defined as a reduction in vision with no demonstrable abnormality of the visual pathway, and which is not immediately resolved by refractive correction. It develops during maturation of the visual pathway and is reversible during the first seven to eight years of life. This is known as the 'critical period'. In some situations the critical period may be extended (Simmers 1999).
The developing visual system relies on good quality visual images. Amblyopia can develop when the image coming into one or both eyes is either blurred or obscured. Amblyopia develops through an abnormal binocular cortical interaction and results in a loss of acuity, contrast sensitivity, and/or positional disorder (Levi 1999).
Amblyopia is usually classified by cause:
(1) strabismic when it is due to the presence of a squint;
(2) anisometropic where the refractive error is significantly stronger for one eye than the other (a difference of more than or equal to 0.75 dioptre is generally thought to be significant);
(3) meridional where there is a significant (more than or equal to 1 00 dioptre) degree of astigmatism;
(4) stimulus deprivation where, for example, a cataract or ptosis (droopy lid) obscures the visual axis;
(5) ametropic where the refractive error is such that neither eye receives a good quality image.
It is not uncommon for the types to co‐exist.
Untreated amblyopia can have a negative impact in adult life. Some career choices have specific visual acuity requirements. The number of careers barred to adults with reduced vision increases with the severity of the deficit (Adams 1999). The lifetime risk of serious visual impairment through loss of vision in the better seeing eye is estimated to be between 1.2% and 3% (Rahi 2002).
Screening programmes to detect amblyopia have been set up on the assumption that early detection and treatment results in more favourable outcomes. Many of the screening programmes in existence do not have a single specific target condition. Data regarding the percentage of children identified with amblyopia, and those with other correctable causes of reduced vision needs to be collected. This review will consider the screening process and the impact it has on the prevalence of amblyopia.
Epidemiology
Amblyopia is a common cause of reduced visual acuity. Estimates of the prevalence of amblyopia vary between 2% and 5%. Factors that contribute to the range of estimates include the population studied and the definition of amblyopia applied. It is, however, widely accepted that in the general population the incidence is between 2% and 2.5% (von Noorden 1996).
Presentation and diagnosis
Amblyopia most commonly affects only one eye. Some types of amblyopia present because they are associated with signs such as strabismus or ptosis whereas other types are only detected by specifically testing for the condition. Affected children therefore often present as a result of vision screening programmes.
There are four main steps in the diagnosis of amblyopia:
(1) monocular visual acuity assessment using an age appropriate vision test;
(2) refraction. Although autorefraction is relatively quick and easy to perform, cycloplegic retinoscopy is recommended. In young children cycloplegia is required to paralyse accommodation as most young children are unable to maintain distance fixation during testing;
(3) fundus and media examination to exclude any pathology;
(4) rechecking the visual acuity with the glasses correction in place.
Some improvement in visual acuity is expected following a period of glasses wear. For the purposes of this review amblyopia will be defined as visual acuity of worse than 6/9 Snellen or 0.2 LogMAR in the affected eye or eyes. A minimum period of four weeks wearing glasses will be required before retesting and diagnosis.
Screening
Screening programmes for amblyopia rely on reduced visual acuity as a marker for the disease. Any screening programme aiming to detect amblyopia this way will, inevitably, also pick up children with other causes of reduced vision, for example uncorrected refractive error.
Currently there are a variety of recommendations for vision screening programmes and a number of different approaches to providing the service. The battery of tests carried out usually includes monocular visual acuity testing with an age‐appropriate test, plus or minus assessment of extra‐ocular muscle function, binocular function, and colour vision assessment. Protocols vary with regard to the vision and/or binocular function test used, threshold for failure, and age at which children are screened. The type of personnel carrying out the testing varies, as does the setting. Some screening programmes are conducted in a community setting. This type of programme allows early screening to take place but makes achieving a high coverage rate more difficult. They may have unacceptably high false positive and recall rates because the participants are too young to co‐operate with testing. To overcome this some programmes are set up to screen children during their first year attending school. Whilst this should improve the percentage reached by screening and decrease the number of false positives it is possible that later treatment of any detected amblyopia may not have such good outcomes.
Screening programmes for amblyopia are concentrated in the developed world and urban areas of the developing world. They can be part of the government health care system or private, commerce‐driven schemes. Screening has a role not only in detection of the target condition but also in improving equity of access to care.
This review is concerned with screening programmes that have amblyopia as the target condition. We will not therefore consider studies that have screened for refractive error, for example photorefraction studies or fundus screening programmes.
Treatment options
Treatment options and the management plan will depend on the type of amblyopia diagnosed. The role of the various treatment options is being examined in a series of Cochrane reviews. Treatment broadly consist of a combination of spectacle correction, patching, or penalisation using drugs or lenses. Factors affecting the outcome of treatment include the age of the patient and the density of the amblyopia. Compliance with treatment prescribed is also important.
The impact of treatment can be quantified by measuring the change in acuity in the amblyopic eye either as:
(1) the number of lines change in visual acuity at the end of treatment;
(2) the proportion of the amblyopia deficit that has been corrected (Stewart 2003);
(3) final visual acuity on an age specific test. A final visual acuity of 6/9 Snellen or 0.2 logMAR or better is generally agreed to represent the elimination of amblyopia.
Rationale for a systematic review
Uncertainty regarding the value of screening for amblyopia has been raised and the need for good quality research into the natural history of the condition and the efficacy of orthoptic treatment for it has been identified (Snowden 1997). This stimulated new research to be undertaken. Hall commented that there has been some interest in preventing amblyopia by screening for causative factors and intervening before it can develop (Hall 2003). He concluded that since the relationship between the presence of these potentially amblyogenic factors and development of the condition is unclear the measure of success of a screening programme should be its ability to detect amblyopia as early as possible rather than to prevent it.
Since amblyopia is commonly monocular it may not be detected within the critical period. The aim of screening for amblyopia is to detect the condition in time to allow effective treatment to take place. The optimum protocol to achieve that remains unclear. For this reason it is necessary to look again at the screening process and its impact on the prevalence of amblyopia and to evaluate the effectiveness of different screening strategies. The costs of providing screening and the potential for a screening programme to cause harm will also be considered.
Objectives
The primary objective of this review will be to evaluate the effects of vision screening in reducing the prevalence of amblyopia in comparable screened versus unscreened populations. Subgroup analyses are planned, if appropriate, to determine the effect of the type of personnel conducting the testing, the age at screening, and the threshold applied for failure. Secondary objectives will be to report available evidence regarding the disability associated with living with uncorrected amblyopia and to document reports of the potential harms and costs associated with screening.
Methods
Criteria for considering studies for this review
Types of studies
This review will include randomised controlled trials, including cluster trials. If no randomised controlled trials are found we will describe other relevant studies in order to explain current practice. No language or date restrictions will be imposed.
Types of participants
We will include studies that screened participants before they started school or as they entered school. Where the age of participants at screen and outcome is not stated we will attempt to clarify this with the authors. Participants with ocular pathology identified by post screening fundus and media examination will be excluded.
Types of interventions
We will include trials of screening by formal, monocular, visual acuity testing, using any age‐appropriate test, and any screening protocol. The following comparisons are of interest:
-
screening to no screening;
-
a failure threshold of worse than 6/9 to a threshold of 6/9 or better;
-
testing personnel with different professional qualifications, for example nurses, teachers, and eye trained personnel;
-
participants screened under the age of five years and those screened aged five years and over.
Types of outcome measures
Primary outcomes
The primary outcome for this review will be the prevalence of amblyopia in comparable screened versus unscreened populations 12 months from screening.
Secondary outcomes
Secondary outcomes for this review will be:
1) the prevalence of amblyopia at more than 12 months from screening;
2) the prevalence of amblyopia at six months from screening;
3) coverage rates achieved by screening programmes in different settings (for example pre‐school and school‐entry screening programmes) defined by the percentage of the target population that was actually screened;
4) any reports attempting to quantify the disability associated with living with uncorrected amblyopia will be described.
Prevalence data will be described in the context of age at outcome assessment and the likelihood of amblyopia remaining amenable to treatment.
Adverse effects
Since screening is performed on asymptomatic people there is the potential for some participants in a screening program to experience harm. Some of the ways in which a screening program might cause harm are (Doust 2004):
-
adverse physical and psychological effects of the screening test itself;
-
the psychological effects of an abnormal screening test result which is found to be false on further testing (a false positive result);
-
any delay in diagnosis due to false reassurance from a 'normal' screening result (a false negative result) and consequent disappointment, anger and increasing judiciary claims;
-
adverse effects of further testing or treatment after an abnormal screening result;
-
the physical and psychological effects of a diagnosis which is made earlier than otherwise with no effective change in health outcome.
Any reports found of the above adverse effects associated with vision screening for amblyopia will be documented.
Quality of life measures
We will describe any report of quality of life measures (for example improved academic performance, job opportunities or income in later life) associated with having had amblyopia corrected by treatment following screening.
Economic data
We will attempt to compare coverage rates achieved by programmes carried out in different settings; cost data for providing screening programmes, including personnel, and treatment, cost per case detected, and cost per case successfully treated.
Follow up
The minimum time between screening and outcome assessment will be six months .
Search methods for identification of studies
Electronic searches
Trials will be identified from the Cochrane Central Register of Controlled Trials (CENTRAL) (which contains the Cochrane Eyes and Vision trials register) on The Cochrane Library, MEDLINE, EMBASE and LILACS (Latin American and Caribbean Health Sciences Literature Database). There will be no language or date restrictions in the electronic searches. The searches below will also be used in a separate Cochrane review of screening for correctable visual acuity deficits (Powell 2004) and therefore contain some additional terms not strictly relevant to this review.
The following strategy will be used to search CENTRAL on The Cochrane Library:
#1 SCHOOLS
#2 CHILD DAY CARE CENTERS
#3 CHILD
#4 INFANT
#5 (#1 or #2 or #3 or #4)
#6 (child* or adolesc* or juvenile* or minor* or school* or kindergarten* or pre‐school* or (pre next school*) or nurser*:ti)
#7 (child* or adolesc* or juvenile* or minor* or school* or kindergarten* or pre‐school* or (pre next school*) or nurser*:ab)
#8 (#5 or #6 or #7)
#9 VISION SCREENING
#10 VISION DISORDERS di:pc
#11 (vision or visual:ti)
#12 (vision or visual:ab)
#13 (test* or screen*:ti)
#14 (test* or screen*:ab)
#15 ((#11 or #12) and (#13 or #14))
#16 VISION TESTS
#17 MASS SCREENING
#18 (#16 or #17) (1966 ‐ 1988)
#19 (#9 or #10 or #15 or #18)
#20 STRABISMUS
#21 AMBLYOPIA
#22 REFRACTIVE ERRORS
#23 (#20 or #21 or #22)
#24 (amblyopi* or squint* or strabism* or anisometropi* or myopi* or hypermetropi* or astigmati* or ammetropi* or hyperopi*)
#25 (lazy near eye*)
#26 (eye* or sight* or vision* or visual*)
#27 (problem* or defect* or impair* or deficit or reduc*)
#28 (#26 and #27)
#29 (#23 or #25 or #28)
#30 (#9 and #23 and #29)
The following strategy will be used to search MEDLINE on SilverPlatter:
#1"SCHOOLS"/ all subheadings
#2 explode "CHILD‐HEALTH‐SERVICES"/ all subheadings
#3 "CHILD‐DAY‐CARE‐CENTERS"/ all subheadings
#4 explode "CHILD"/ all subheadings
#5 (CHILD* or ADOLESC* or JUVENILE* or MINOR* or SCHOOL* or KINDERGARTEN* or PRE?SCHOOL* or NURSER*) in TI,AB
#6 (#1 or #2 or #3 or #4 or #5)
#7 "VISION‐SCREENING"/ all subheadings
#8 explode "VISION‐DISORDERS"/ diagnosis , prevention‐and‐control
#9 (VISUAL or VISION) near4 (TEST* or SCREEN*)
#10 explode "VISION‐TESTS"/ all subheadings
#11 #10 and (PY=1966‐1988)
#12 explode "MASS‐SCREENING"/ all subheadings
#13 #12 and (PY=1966‐1988)
#14 (#7 or #8 or #9 or #11 or #13)
#15 explode "STRABISMUS"/ all subheadings
#16 "AMBLYOPIA"/ all subheadings
#17 explode "REFRACTIVE‐ERRORS"/ all subheadings
#18 (EYE* or SIGHT* or VISION* or VISUAL*) near4 (PROBLEM* or DEFECT* or IMPAIR* or DEFICI* or REDUC*)
#19 LAZY near EYE*
#20 AMBLYOPI* or SQUINT* or STRABISM* or ANISOMETROPI* or MYOPI* or HYPERMETROPI* or ASTIGMATI* or AMMETROPI* or HYPEROPI*
#21 (#18 or #19 or #20) in TI,AB
#22 #15 or #16 or #17 or #21
#23 #14 and #22
#24 #6 and #23
To identify randomised controlled trials, this search will be combined with the Cochrane Highly Sensitive Search Strategy phase one and two as contained in the Cochrane Reviewers' Handbook (Alderson 2004).
The following strategy will be used to search EMBASE on SilverPlatter:
1. exp School/
2. exp day care/
3. exp child health care/
4. Child/
5. (child$ or adolesc$ or juvenile$ or minor$ or school$ or kindergarten$ or preschool$ or pre‐school$ or nurser$).ab,ti.
6. 1 or 2 or 3 or 4 or 5
7. exp Vision Test/
8. exp Visual Disorder/pc, di [Prevention, Diagnosis]
9. ((vision or visual) adj3 (test$ or screen$)).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
10. 7 or 8 or 9
11. exp STRABISMUS/
12. exp AMBLYOPIA/
13. exp Refraction Error/
14. ((eye$ or sight$ or vision or visual) adj5 (problem$ or defect$ or impair$ or deficit$ or reduc$)).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
15. (lazy adj3 eye$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
16. (amblyop$ or squint$ or strabism$ or anisometropi$ or myopi$ or hypermetropi$ or astigmati$ or ammetropi$ or hyperopi$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
17. 11 or 12 or 13 or 14 or 15 or 16
18. 6 and 10 and 17
To identify randomised controlled trials this search will be combined with the following:
1. Randomized Controlled Trial/
2. Controlled Study/
3. randomization/
4. Double Blind Procedure/
5. Single Blind Procedure/
6. Clinical Trial/
7. Crossover Procedure/
8. follow up/
9. exp prospective study/
10. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9
11. exp ANIMAL/
12. Nonhuman/
13. Human/
14. 11 or 12
15. 14 not 13
16. 10 not 15
17. (clinica$ adj3 trial$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
18. ((singl$ or doubl$ or trebl$ or tripl$) and (mask$ or blind$ or method$)).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
19. exp PLACEBO/
20. placebo$.mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
21. random$.mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
22. exp Methodology/
23. (latin adj3 square$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
24. ((control$ or prospectiv$ or volunteer$) adj3 (trial$ or method$ or stud$)).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
25. (cross adj3 over$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
26. (crossover$ or cross‐over$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
27. 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26
28. 27 not 15
29. 16 or 28
LILACs will be searched using terms from the MEDLINE search strategy.
Manual searches
The following conference proceedings will be manually searched, if possible.
-
European Strabismus Association.
-
International Strabismus Association.
-
American Association of Paediatric Ophthalmology and Strabismus.
-
Royal College of Ophthalmologists.
A manual search will be carried out of The British Orthoptic Journal from 2003 to the present, previous years having already been searched. We will attempt to contact specialists in the field for details of relevant trials.
Data collection and analysis
Assessment of search results
One reviewer will check the initial search results to identify and exclude any studies that are clearly not relevant. A minimum of two reviewers independently will then scrutinise each of the abstracts of any studies which are potentially relevant and, using a spreadsheet, divide them into three categories: include, exclude, or unclear. The gradings will then be compared. Studies that are excluded by both reviewers will be excluded at this stage. We will retrieve the full articles for those abstracts included by both reviewers. Any articles not appearing in the same category on the lists will be discussed and if necessary the third reviewer will be asked to read and evaluate the abstract. We will request full copies of articles where it is unclear from the abstract whether or not they should be included and in any cases where there is still a disagreement. The above process will then be repeated. If necessary we will seek further information from authors. Studies to be included will then be assessed for quality.
Assessment of methodological quality
We will assess trial quality using the guidelines in section 6 of the Cochrane Reviewers' Handbook and the Cochrane Eyes and Vision Group Review Development Guidelines.
We will assess three main sources of bias:
(1) Selection bias ‐ controlled by randomisation and allocation concealment.
(2) Detection bias ‐ whether or not examiners responsible for measuring outcomes were masked to the group allocation of participants.
(3) Attrition bias ‐ how participants lost to follow up were accounted for. We will consider whether follow up rates for groups were similar and whether all participants were analysed as randomised i.e. was an intention to treat analysis performed.
Performance bias (masking of participants and personnel) is not relevant in this context.
Each parameter will be graded as (A) yes ‐ requirements met; (C) no ‐ requirements not met; or (B) unable to determine. Clarification will be sought from authors of studies graded B. We will exclude in sensitivity analyses studies graded B or C to examine whether they have an impact on the size and direction of effect.
Data collection
Two reviewers independently will extract data using the Cochrane Eyes and Vision Group data collection form. Both reviewers will enter data into RevMan using the double data‐entry system to check for discrepancies.
Data synthesis
Studies included in the review will be checked for homogeneity by:
1) examining the characteristics of the included studies;
2) looking for poor overlap of the confidence intervals on the forest plot;
3) the result of the chi squared test.
If appropriate a meta‐analysis will be carried out using the RevMan software. The fixed effect model will be used if there are fewer than three trials to analyse. If more trials have been included the random effects model will be used.
Subgroup analysis are planned of trials with:
-
failure thresholds of 6/9 or 0.2 or better and those with worse than 6/9 or 0.2;
-
carried out by different personnel i.e. teachers, school nurses and eye trained professionals;
-
participants screened under the age of five and those screened at five years or over.
Cluster trials will be dealt with according to the guidelines in the Cochrane Reviewers' Handbook.
It is anticipated that there will be two sets of data: screening versus no screening (intervention versus no treatment), and one screening protocol compared to another. These will be analysed separately.
Within each group the proportion of participants with and without amblyopia should be reported as an outcome measure i.e. dichotomous data. We plan, therefore, to use the risk ratio as the measure of effect. For continuous data we will present the weighted mean difference. If different instruments have been used to measure outcomes, but are similar enough to be combined, the standardised mean difference will be calculated.
Sensitivity analysis
We plan to conduct the following sensitivity analyses:
1) excluding trials graded C on any aspect of methodological quality;
2) excluding trials graded B or C on any aspect of methodological quality;
3) excluding industry funded studies;
4) excluding unpublished studies.
