Article Text

Download PDFPDF

Prevalence and risk factors for common vision problems in children: data from the ALSPAC study
  1. C Williams1,
  2. K Northstone2,
  3. M Howard3,
  4. I Harvey4,
  5. R A Harrad5,
  6. J M Sparrow5
  1. 1
    Centre for Child and Academic Health, University of Bristol, Bristol, UK
  2. 2
    Department of Social Medicine, University of Bristol, Bristol, UK
  3. 3
    Department of Community Based Medicine, University of Bristol, Bristol, UK
  4. 4
    School of Medicine, Health Policy and Practice, University of East Anglia, Norwich, UK
  5. 5
    Bristol Eye Hospital, Bristol, Bristol, UK
  1. Ms C Williams, 24 Tyndall Avenue, Bristol BS8 1TQ, UK; cathy.williams{at}


Objective: To estimate the distribution and predictors of some common visual problems (strabismus, amblyopia, hypermetropia) within a population-based cohort of children at the age of 7 years.

Methods: Children participating in a birth cohort study were examined by orthoptists who carried out cover/uncover, alternate cover, visual acuity and non-cycloplegic refraction tests. Prospectively collected data on potential risk factors were available from the study.

Results: Data were available for 7825 seven-year-old children. 2.3% (95% CI 2.0% to 2.7%) had manifest strabismus, 3.6% (95% CI 3.3% to 4.1%) had past/present amblyopia, and 4.8% (95% CI 4.4% to 5.3%) were hypermetropic. Children from the lowest occupational social class background were 1.82 (95% CI 1.03% to 3.23%) times more likely to be hypermetropic than children from the highest social class. Amblyopia (p = 0.089) and convergent strabismus (p = 0.066) also tended to increase as social class decreased.

Conclusions: Although strabismus has decreased in the UK, it and amblyopia remain common problems. Children from less advantaged backgrounds were more at risk of hypermetropia and to a lesser extent of amblyopia and convergent strabismus. Children’s eye-care services may need to take account of this socio-economic gradient in prevalence to avoid inequity in access to care.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Common visual problems of childhood in the UK include refractive errors, amblyopia and strabismus.1 Contemporary population-based data are not available on the prevalence or distribution of these problems in children in the UK. In the 1960s, 5.3% of 7 year-old schoolchildren in Cardiff had a manifest strabismus.2 Among 11-year-old children in the 1970 UK birth cohort, 5.9% had presumed myopia, 1.5% presumed hypermetropia, and 1.5% presumed amblyopia. While vision defects overall were more common in deprived children, there was no clear association between factors associated with amblyopia and social class.3

Up-to-date, population-specific prevalence data are necessary for planning the quantity and delivery of ophthalmic and optometric services, particularly as expansions in community-based eye care have been recommended.4 We therefore present results from an ongoing birth cohort study to estimate some of these data in children. The relevant conditions that could be reliably identified were hypermetropia, strabismus and amblyopia. We also looked at potentially important risk factors identified from the literature.3 5 6


The subjects were all participants in the Avon Longitudinal Study of Parents and Children.7 All children born to mothers resident in Avon (in the southwest of England) with an estimated date of delivery between 1 April 1991 and 31 December 1992 were eligible for the study. A total of 14 541 children (estimated 85% of those eligible) were recruited, of whom 13 988 were still alive and participating at the age of 1 year. The original cohort was largely representative of the UK 1991 Census for Avon and for the UK as a whole.7 However, deprived families, families where the mother was teenaged at the birth of her child, and families of non-white ethnic origin are under-represented.

Data have been collected by self-completion questionnaires sent to the mother and her partner, from weights and measurements at birth and from yearly assessments since the children were 7 years old, where a variety of physical, cognitive and psychosocial characteristics have been assessed. This paper concerns the children’s visual status when they attended the assessment at the age of 7 years. The data were collected between October 1998 and October 2000.


Vision problems at 7 years

Refraction was assessed without cycloplegic drops, by a Canon R50 (Canon, Tokyo) autorefractor. A nested validation study8 indicated that the autorefractor data had a sensitivity of 71% and specificity of 99% when used to screen for true hypermetropia of ≥+2.00 D spherical equivalent in a sample with 25% having amblyopia or strabismus and 61% sensitivity and 99% specificity when only children without these conditions were included. Visual acuity was measured monocularly, using the “2000” series ETDRS charts (a different one for each eye). The acuity for each eye was determined as the best of presenting (with glasses if worn) and presenting-plus-pinhole. Any misalignment was quantified by an orthoptist using a simultaneous prism cover test and alternate prism cover test, both at 33 cm and at 6 m, and with and without glasses if worn. Strabismus was classified according to the direction (convergent, divergent, vertical or mixed) and whether it was present in normal viewing with both eyes open (manifest) or only when the eyes were dissociated and/or one covered (latent).

Data on potential risk factors

The Standard Occupational Classification9 was used as an indicator of the family social class, based on the mother’s report of her own and her partner’s occupation at 32 weeks of pregnancy. This scale goes from I (professional) through II, III (subdivided into Manual and Non-Manual), IV and V. We examined housing tenure (owned/mortgaged; subsidised rented; other), recorded when the child was 8 months old as another indicator of socio-economic status. We also considered as potential risk factors parental self-reported hypermetropia (yes/no), history of strabismus/amblyopia in a first-degree relative (yes/no); smoking during pregnancy (either first trimester or last 2 weeks, yes/no), child’s ethnicity (white/non-white), gestation (weeks), birth weight (g) and sex.


Hypermetropia was defined as an autorefraction of ⩾+2.00 D spherical equivalent in either eye. Presenting visual acuity in the best-seeing eye was categorised as “normal” (⩽0.3 logMAR, 6/12 or better), or “reduced” (>0.3 logMAR, worse than 6/12). Manifest strabismus was defined as any deviation >1 prism dioptre (pd) on prism cover testing. “Clinically significant” strabismus included all manifest cases and children with arbitrarily defined “large” latent deviations: ⩾10 pd if convergent and ⩾15 pd if divergent. Children with amblyopia were defined as those with a history of patching treatment and/or with an interocular difference in best acuity for each eye of >0.2 logMAR units where the worst-seeing eye had a best acuity of worse than 0.3 logMAR, and the eye looked normal on dilated funduscopy.

We used the higher (more advantaged) of the mother’s or her partner’s occupational social class (SC) where both were available, or just the mother’s where necessary. Social class categories IIIM, IV and V were combined for analysis as the numbers were small. We defined prematurity as gestation less than 37 weeks and intrauterine growth retardation (IUGR) was defined as birth weight-adjusted-for-gestation more than 2 SD lower than the mean for the whole sample.

Sex-specific and ethnicity-specific prevalences for the conditions under investigation were calculated. Prevalences were compared between social class groups. Logistic regression was used to calculate unadjusted and adjusted odds Ratios (OR) and 95% CI (adjusting these for the factors listed above). All analyses were performed using SPSS for Windows v12 (SPSS, Chicago).


Of the original ALSPAC cohort, 7843 attended the 7-year vision assessment. Eighteen children had ocular pathology and were excluded, leaving 7825, of whom 7538 (96.3%) provided autorefraction data for both eyes: these are the children presented in this paper.

Compared with the ALSPAC children who did not attend or did not provide data, the children with data were more likely to be white, female, born at term, in SC I or II, to live in owner/occupier housing and not to have been exposed to maternal smoking in utero: all p<0.001 except sex (p = 0.017). The characteristics of the participants included in this analysis are shown in table 1.

Table 1 Distribution of variables used in the analysis (n = 7388)

Prevalence of strabismus, amblyopia, hypermetropia and reduced presenting visual acuity (table 2)

Table 2 Overall and sex- and ethnicity-specific prevalences (95% CI) of strabismus, amblyopia, reduced presenting acuity in the better eye and hypermetropia in 7538 seven-year old children

Over a quarter of children had a latent strabismus, while only 2.3% had a manifest strabismus. Most of the latent deviations were small, over three-quarters being 10 pd or less, and 73.0% were divergent, 26% convergent and 1% vertical. By contrast, the manifest deviations were predominantly convergent (73.4%), with 21.4% divergent and 5.2% having a vertical component. Overall, 3.4% of children had a “clinically significant” strabismus. Less than 1% (68 children, 0.9%) had previously had squint surgery. These include 39 children (57% of those who had surgery) who still had some deviation and are included in the prevalence data and 29 children who were orthophoric when examined.

About one child in every 30 had previously been treated for amblyopia or was amblyopic on examination. However, very few children (0.6%) had impaired acuity in their better-seeing eye.

Hypermetropia affected nearly one child in every 20 but was less frequent in non-white children. Apart from this, there was no evidence of major differences in risk between the sexes or the two ethnic groups.

Associations between conditions

As would be expected, strabismus, hypermetropia and amblyopia often clustered in the same children. Of the 365 children with hypermetropia at least +2.00 D in either eye, 124 (34.0%) had a clinically significant convergent strabismus, and 158 (43.3%) had past or present amblyopia; overall, 199 (54.5%) children with hypermetropia had one or other condition, while 83 (22.7%) had both.

Prevalence rates according to parental occupation (table 3)

Table 3 Numbers (percentages) of children with common visual defects according to their family Standard Occupational Classification (1991)

These unadjusted figures show a tendency towards increasing prevalence of both amblyopia and clinically significant convergent strabismus with decreasing SC, with the main feature being a lower prevalence for children in SC I. For hypermetropia, however, there is stronger evidence that the prevalence increases as SC decreases (becomes less advantaged). The combined prevalence of children with hypermetropia and/or convergent strabismus and/or amblyopia was 30% lower in children from SC I backgrounds than in children from SC IIIM/IV/V. Very few children had impaired vision in their better eye, and there was no relationship with SC. Similarly, divergent strabismus did not vary with SC.

Sixty-eight per cent of children with hypermetropia had glasses in the clinic, and another 7.2% had left their glasses at home or had abandoned wearing them. There was no difference in the proportion of children who had glasses with them, in different categories of SC (data not shown): p = 0.68.

Results of risk-factor analyses for strabismus, amblyopia and hypermetropia (table 4)

Table 4 Risk-factor analysis for strabismus, amblyopia and hypermetropia: odds ratios (OR) before and after mutual adjustment in 7538 children

Convergent strabismus was predicted by a family history of hypermetropia or of strabismus/amblyopia. Both of these were reduced in importance once the child’s own hypermetropia was taken into account. Prematurity remained a strong predictor of convergent strabismus whatever adjustments were made. Divergent strabismus was reliably predicted by only two risk factors (maternal smoking in late pregnancy and intrauterine growth retardation), and there was a suggestion that it was more common in girls. Amblyopia was predicted by a family history of strabismus/amblyopia or of parental hypermetropia and by family housing and by maternal smoking. After adjustment for concurrent vision problems, however, smoking and family history were the only predictive factors.

These adjusted analyses also show a relationship between hypermetropia and SC: there was an increase in risk of 20% (1.20; 95% CI 0.02 to 1.41) between each SC group, with the most disadvantaged children being over 80% more likely to be hypermetropic than the most advantaged children. The child’s housing in early life also reflected a similar and apparently independent trend. Hypermetropia was less frequent in non-white children. Children with at least one hypermetropic parent, or a first-degree relative with strabismus/amblyopia, were more likely to be hypermetropic themselves.

Although smoking as a dichotomous variable did not predict hypermetropia (data not shown; p>0.1 in adjusted analysis), children whose mothers smoked in pregnancy had slightly more hypermetropic spherical equivalent refractions for their right eyes than children whose mothers did not: +0.26 D vs +0.19 D, ANOVA: F = 6.26, df = 1, p = 0.012 for smoking in the first trimester and +0.25 D vs +0.19 D, ANOVA: F = 3.37; df = 1, p = 0.066 for smoking in late pregnancy.


Overall, the prevalence of strabismus in this cohort (2.3% manifest, 3.4% including clinically significant latent deviations) is lower than the 5.3% described in the 1960s for a group of schoolchildren and than the 3.98% reported recently for Ireland.10 Our data are more similar to those for an Australian sample of 12-year old children (2.7%)11 and support the suggestion, based on an audit of surgical practice, that the incidence of constant esotropia in the UK might be reducing.12

A family history of strabismus or amblyopia was a robust predictor of the same conditions in the children, as would be expected. The tendency for strabismus to be inherited has been observed for many years,13 but as yet no genetic loci have been identified for strabismus or amblyopia. Associations between maternal smoking and divergent strabismus6 or convergent strabismus have already been described.6 14 Prematurity has also previously been associated with an increased risk of strabismus,5 15 as has low birth weight.6 14 The cumulative incidence of amblyopia (3.6%) reported here is higher than figures recently reported for children in Australia (1.8%)16 and similar to the 3.1% residual or present amblyopia in Ireland.10

A previous study of vision screening using a subset (1914, 25%) of this cohort17 may have affected the age at presentation of amblyopia, but not the cumulative incidence as estimated here. It is possible that the prevalence of esotropia at 7 and of children who had required surgical intervention for esotropia, may have been reduced if the early screening resulted in a shorter period of misalignment.18 Thus, these data may slightly underestimate the prevalence found in other areas of the UK.

Social disadvantage, either absolute or relative to others, is an important cause of ill-health in all societies.19 Although myopia in children is known to be associated with social advantage and/or education20 this complementary association between disadvantage and hypermetropia in childhood has not been emphasised in the literature.

Mild hypermetropia (<+2.00 D) is the norm for older children and young adults, but more severe hypermetropia has already been described as a major risk factor for convergent strabismus and/or amblyopia. We selected +2.00 as a definition of hypermetropia to aid comparison with other prevalence studies (eg, the Refractive Error in Children Studies2123) and with reports investigating patterns of spectacle usage in children in which hypermetropia >+2.00 D was used at the least level which might require correction.24

There is variability in the prevalence of hypermetropia worldwide, from 0.7% in rural India,21 to 21.6% for 5–7-year-olds in Chile.23 The corresponding figure for this cohort is 4.8%, which will be an underestimate as the data were collected without cycloplegia. Some groups report an excess of hypermetropia in girls,22 which was not seen in this study. While there is clear evidence that refractive error is largely heritable,25 and less frequent in non-white children,26 our finding of “social patterning,” or variation according to SC suggests that social/lifestyle factors may also be involved, for example smoking, as seen here and as reported earlier.27

Cross-sectional data22 suggest that the prevalence of hypermetropia decreases during childhood, and it is likely that some or many children included as hypermetropic here will not be so by their mid-teens. Hypermetropia has already been suggested as a marker of delayed development,28 and social disadvantage may be predictive of developmental delay.29

The strengths of the study are the large sample size and broadly representative nature of participating children, the use of orthoptists to collect accurate ophthalmic data and the variety of prospectively collected demographic and risk factor data. This study is limited by the facts that the refraction data were collected without the use of cycloplegic drops, that screening interventions were previously offered to a minority of the children and that the sample of children under-represents non-white children and those from the lower end of the socio-economic spectrum. This latter limitation is further compounded by the fact that social class is missing for 8% of the children in the study. This resulted in excluding children from the analyses who may well have been from the lower social classes. It is therefore likely that our results are biased in favour of children from more affluent families, who generally have higher response rates, such that we underestimate the true rates of the conditions in the UK as a whole. We may have underestimated the influence of heredity by using very simple questionnaire data. Cover testing was carried out near and at 6 m, but not for far distance viewing, so we will have missed some cases of divergent strabismus. We used data reflecting socio-economic status early in life and may have missed relevant contributions from later changes in status.

Strabismus and amblyopia remain common problems within the UK, each affecting at least one in 30 children. We present new data indicating that disadvantaged children are more at risk of hypermetropia. Further research into social and environmental determinants of vision problems in children is needed. Policy makers and clinicians need to be aware of socio-economic gradients in prevalence to reduce the risk of inequity in access to eye care for children.


We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The UK Medical Research Council, the Wellcome Trust and the University of Bristol provide core support for ALSPAC. This publication is the work of the authors, and CW will serve as guarantor for the contents of this paper. This collection of the vision data presented here was specifically funded by the Research and Development Directorate of the Southwest Regional Health Authority.


View Abstract


  • Competing interests: None.

  • Ethics approval: Ethical approval for the study was obtained from the ALSPAC Law and Ethics Committee and the Local Research Ethics Committees.

  • Patient consent: Patient consent was obtained.

Linked Articles

  • At a glance
    Harminder S Dua Arun D Singh