The prevalence of visual impairment in the UK
A review of the literature
Rosemary Tate, Liam Smeeth, Jennifer Evans, Astrid Fletcher
Dept of Epidemiology & Population Health
London School of Hygiene & Tropical Medicine
Chris Owen St George’s Hospital Medical School
Alicja Rudnicka Wolfson Institute of Preventive Medicine
Report commissioned by the Royal National Institute for the Blind
Acknowledgements
Funding for the review was provided by the Royal National Institute for the Blind
Advisory Committee
Liz Bates (Director Corporate Policy, Ashton, Leigh & Wigan Primary Care Trust)
Prof Ian Bruce (VOLPROF, City University)
Prof Alistair Fielder (Professor of Ophthalmology, Imperial College London and Western Eye Hospital)
Dr Carol Lupton (Policy Research Programme, Department of Health
Dr Angela McCullagh (Research Director, The Pocklington Trust)
RNIB
Lesley-Anne Alexander (Chief Executive)
Nigel Charles (Research Development Manager)
Fazilet Hadi (Director of Policy)
Dr Adam Ockelford (Assistant Director, Education and Employment)
Many thanks also to Sue Keil at RNIB who provided extra information for Chapter 5. Special thanks to Nigel Charles for commissioning this project and for invaluable support and advice.
Authors’ details
Rosemary Tate PhD, research fellow, London School of Hygiene & Tropical Medicine (now at Institute of Child Health)
Liam Smeeth MRCP, Senior clinical lecturer in epidemiology, London School of Hygiene & Tropical Medicine
Jennifer Evans PhD, lecturer in epidemiology, London School of Hygiene & Tropical Medicine
Astrid Fletcher PhD, Professor of epidemiology, London School of Hygiene & Tropical Medicine
Chris Owen PhD, Senior Research Fellow in Epidemiology, St George's Hospital Medical School
Alicja Rudnicka PhD, Lecturer in Epidemiology & Medical Statistics, Wolfson Institute of Preventive Medicine
Contribution of authors
Rosemary Tate undertook the literature review and preparation of main tables. Liam Smeeth reviewed and updated Chapter 5 on Children. Jenny Evans contributed the section on blindness registrations and provided critical comments on the review. Chris Owen and Alicja Rudnicka carried out the analyses and wrote the material for Chapter 6.
Astrid Fletcher took overall responsibility for the project and wrote the Summary, Chapter 1 and Chapter 7.
Authors’ potential conflict of interest
Data on self reported visual difficulties and visual acuity measures reviewed in this report comes from the MRC funded Trial of assessment and management of older people in the community (Principal Investigator Astrid Fletcher), the associated causes of vision impairment study funded by the The Pocklington Trust (with Jenny Evans and Richard Wormald) and the nested MRC funded trial of screening for vision impairment within the main MRC Trial (with Liam Smeeth).
Summary
We identified three approaches to describing visual problems in the population.
- Visual difficulties - based on self-report of difficulties with vision related functions ranging from single item questions to disability scales. Most studies have used questions on difficulty in “reading newsprint” as minimum criteria for difficulties with seeing.
- Clinical measures, predominantly visual acuity. Distance visual acuity is the basis for categorising vision impairment in the WHO International Classification of Diseases (ICD). Most studies have measured distance acuity using an illuminated Snellen chart. Visual acuity can be measured with usual aids if worn i.e. contact lenses or glasses (“presenting vision”) or after full refraction for refractive error (“best corrected visual acuity”). In surveys full refraction may not be available and studies may use pinhole correction to try to remove some of the refractive error (pinhole corrected visual acuity).Results may be presented as binocular vision or as vision in the better eye. The WHO ICD classifications use best corrected visual acuity ie vision in the better eye after full correction. Vision impairment is defined as a Snellen acuity <6/18 and blindness is defined as visual acuity <3/60. Visual acuity <6/12 is not used internationally to define impairment but is reported by some studies as it represents a reduction in visual acuity sufficient to affect lifestyle e.g. it corresponds to the requirements for sight for the UK driving licence. A recent report from the WHO Prevention of Blindness group recommends that binocular presenting vision should also be reported in population based surveys as this represents the actual impairment experienced by the individual in their everyday life (whether due to underlying disease such as cataract or due to uncorrected refractive error). Information on the causes of poor acuity provides important information for policy and service provision. Among children, a commonly used criteria is visual loss sufficiently bad to mean a child is identified as being in need of special educational or social services.
- Vision related “quality of life” describing the impact of vision problems on everyday functioning and well-being using specially developed scales. This is a relatively new area and to date there has been little experience in the use of such scales in population based studies in the UK.
A review of the literature was undertaken for all UK population based studies and reports which had included information on at least one of the above. Vision- related quality of life data were available for one study only (from the authors of this report) and are presented in Chapter 6.
Findings
Visual difficulties
The estimates of visual difficulties (described as visual disability in some studies) derive predominantly from large government surveys and are reviewed in Chapter 2. The results show, when using similar survey methods and criteria for visual disability, (as in the OPCS surveys of 1985 and 1996/7) some consistency in the results in the older population: 22% of the over 75s in private households reported difficulties with reading newsprint or worse and 26% when people in communal establishments were included. There was less consistency in the results for the younger age groups where the two surveys differed over 2 fold in their estimates for the population aged 16-64 from 0.8% to 2.0 % mainly due to a difference in the selection criteria into the survey. If the proportion of the younger population with visual difficulties is over estimated this will substantially inflate the numbers with sight difficulties in the population since the numbers of people in the population aged 16-64 is much greater than in the older age groups. Thus the estimated numbers with sight difficulties in private households in England and Wales based on the 1996/7 survey was estimated at nearly 2 million compared to 1.4 million from the 1985 survey.
There are doubts concerning the reliability of all estimates based on the criteria of “difficulties with reading newsprint” or worse. The lack of reliability is based on findings from the OPCS/RNIB 1998/1999 survey. The sample for this study included a re- survey of respondents who had reported sight difficulties in the 1996/7 survey. A substantial proportion of respondents who initially self reported with a sight problem denied on re-survey they had had a sight problem. The main change in response was to the questions: “Difficulty in seeing a friend across the road” and “Difficulty reading ordinary newsprint”. With more severe task difficulties the level of misclassification was minimal. The estimates of visual disability in the OPCS/RNIB 1998/1999 study took account of this misreporting and were considerably lower than either of the two previous surveys with a prevalence for difficulties with reading newsprint or worse of 14% for the over 75s and 1.1% for the 16-59 age group. However, estimates from the OPCS/RNIB 1998/1999 study may be too low because they exclude some people who initially had a problem and had subsequently received treatment.
Confidence in the results on is further limited by the lack of detailed information on response rates in the government surveys. These surveys used a “sift” approach to screen out people with no apparent disability with successive reductions in response with successive sifting. Although the final response rates in the sample from which the estimates are derived are not clearly presented they are of the order of 50% or less in most government surveys. The representativeness of the surveyed population is therefore difficult to assess.
Visual acuity
The studies which have measured visual acuity in the population are reviewed in Chapter 3. The overwhelming majority of studies have been in the older population. Because of differences between studies in reporting results and criteria for defining visual impairment it was not possible to pool the results to obtain a single estimate.
The prevalence of vision impairment varied according to the definitions for defining visual impairment and whether refractive error was included in the estimates.
Prevalence of visual impairment
Studies using visual acuity measurements have used various criteria for definitions and cut points of visual impairment. The results from two studies: National Diet and Nutrition Study (NDNS), and MRC Assessment Trial that were nationally representative of the older population and use uncorrected presenting VA and similar cut points are given in the table below (unpublished data from NDNS provided by Dr van der Pols and unpublished data on VA <6/12 for MRC Assessment Trial using same LogMar criteria as NDNS).
65-74 / 95% CI / 75-84 / 95%CI / 85+ / 95% CINDNS community sample1 / 475 / 429 / 222
VA <6/18 / 5.6 / 3.5, 7.6 / 13.3 / 10.1, 16.5 / 31.7 / 25.5, 37.8
VA<6/12 / 15.8 / 12.6, 19.1 / 28.3 / 24.1, 32.6 / 54.0 / 47.4, 60.6
MRC Assessment Trial2 / Age group not included / 11500 / 3100
VA <6/18 / 8.5 / 7.1, 9.8 / 26.8 / 23.9, 29.7
VA<6/12 / 18.7 / 16.5, 20.9 / 45.8 / 42.2, 49.5
1 Best score (Glasgow Acuity Cards) of two eyes without pinhole correction
2 Binocular acuity without pinhole correction
These results show that the point prevalence estimates of vision impairment (VA <6/18) are : 6% NDNS only) at ages 65-74, 8% to 13% at ages 75-84 and 27% to 32% at ages 85+. Estimates of the prevalence of minor visual loss (visual acuity <6/12) are 16% (NDNS only) at ages 65-74, between 19% to 28% at ages 75-84 and 46% to 54% at ages 85+.
Considering the upper and low bounds of the 95% confidence intervals the results show that the estimates of vision impairment (VA <6/18) range from 4% to 8 % (NDNS only) at ages 65-74, 7% to 16% at ages 75-84 and 24% to 38% at ages 85+. Similarly the range of estimates of the prevalence of minor visual loss (visual acuity <6/12) are from 13% to 19% (NDNS only) at ages 65-74, between 16% to 33% at ages 75-84 and 42% to 61% at ages 85+.
Estimates for the nursing home population from NDNS show high levels of pinhole corrected vision impairment (VA <6/18): 12.1% at ages 65-74 (95% CI 9.7%, 23.2%), 30.0% of those aged 75-84 (95% CI 20.5%, 39.5%) and 46.9% of those aged 85+ (95% CI 37.7%, 56.1%). However as these are the only recent estimates available for the nursing home population we recommend that they are viewed with some caution.
Two studies provided estimates of blindness using international criteria of VA < 3/60. In the MRC assessment trial the estimates for blindness for the 75+ age group were 2.1 (95% CI 1.8, 2.4) and in a small study in London the estimates for blindness were very similar (1.9%) although the confidence intervals were very wide (0.2, 6.6)
In Chapter 4 we reviewed the few UK studies which have provided some information on the causes of vision impairment. Untreated refractive error and cataracts are the major remediable causes of vision impairment in older people ranging from 50% to 70% as a proportion of visual impairment. With increasing age the proportion of treatable vision impairment declines as age related macular degeneration becomes more prevalent.
In Chapter 3 we reviewed an earlier report from the RNIB that raised concerns that a large proportion of the visually impaired population were not on the Blind/partial Sight register. As the results for visual impairment from that report were based on sight tests alone without excluding possible remediable conditions (such as refractive error and cataracts) the registerable component is likely to have been considerably overestimated (around two fold). We showed that, based on other studies which have collected data on the causes of vision impairment, the numbers likely to be registered because of permanent vision loss and the actual numbers registered from Department of Health statistics were of a similar magnitude (and differed by about 10%, possible range 0% to 20%). We conclude that, although registration rates may show some under ascertainment in the older age groups, the evidence does not support substantial under-reporting.
Chapter 5 summarises the literature relating to the prevalence of visual impairment and visual disability in children in the UK. Most of the available data about visual loss in children comes from registers or from surveys of providers of health care, social care or educational services to children with visual loss. There are marked variations in the criteria used to define visual disability and visual impairment in different areas of the UK and in the various surveys performed. Caution is required in interpretation of any single estimate of the prevalence of visual impairment or blindness among children in the UK. nited Kingdom. However, using a broad and pragmatic definition of visual loss sufficiently poor as to mean a child is identified as being in need of special educational or social services, the existing data suggest a prevalence of visual impairment in the region of 10-20/10,000 children. For severe visual loss, the best estimate available suggests a cumulative incidence of blindness or severe visual impairment by the age of 16 years of 5.9/10,000 children. Notably, around half the children receiving support from visual impairment services may have additional disabilities, and this proportion may be even higher for children with severe visual loss.
Only one UK population based study reported vision related quality of life. Using the NEI-VFQ 25, 69% reported their vision was excellent or good and 31% reported their vision as fair or poor. Around a half reported problems with near or distance activities. Visual acuity levels were strongly associated with all sub-scales except ocular pain. The most important scales related to visual acuity were general vision, near activities, and social functioning and, to a lesser extent, dependency and colour vision. However visual acuity only explained about a fifth of the variation between people in their reporting on these scales.
Most studies have been done in the older population and there is a scarcity of data in younger adult age groups in the UK. A review of epidemiological studies performed in Western Europe, North America and Australia covering the age group 20 to 59 years found the prevalence of blindness was 0.08 and of visual acuity 6/24 to 6/48 was 0.07%. These figures agree well with the prevalence of registrations in a similar age range and we conclude that registration data provide reasonably accurate estimates of the prevalence of serious vision impairment in the younger adult age groups. Information is lacking on less severe levels of visual acuity.
Conclusions and Recommendations
Policy
Estimates of the numbers of people in the population with sight problems are critically dependent on the validity of the methods, definitions used and the quality of the studies (such as response rates and representativeness of the population).
Estimates based on self report from government surveys are less reliable than estimates based on visual acuity because of concerns about the validity of the questions and the poorer response rates. In addition, these estimates are likely to be a substantial overestimate of the numbers since they include people with minimal problems.
Estimates based on visual acuity measurements from the representative studies of the older population are sufficiently consistent for general purposes of planning and estimation of the proportion and numbers of people in the UK with vision impairment. However we caution that these estimates are seen as approximations and not as “tablets of stone” as prevalence estimates will vary from place to place and over time and because all estimates carry a range of uncertainty as shown by the 95% confidence intervals.
Estimating the prevalence of vision impairment for future older populations requires assumptions about temporal trends in the underlying conditions, and future levels of service provision and uptake. No data are available on temporal trends in the incidence of refractive error, cataracts or age related macular degeneration and such data would be prohibitively expensive to obtain. The estimated increase in the proportion of people with age related macular degeneration eligible for registration has been calculated to increase by 11% over a 11 year period to from 2000 to 2020 assuming no change in the prevalence of this condition or improved treatments
Studies in the older population have consistently shown that about 50% of visual impairment is due to treatable conditions i.e. cataract and refractive error. Appropriate action e.g. media campaigns and advocacy is required to ensure that eye care providers are aware of the current unacceptably high levels of untreated conditions. Campaigns should also target older people and their carers to raise awareness of poor vision in later life and the potential for improving visual impairment through spectacle correction and cataract surgery.
For children, agreed criteria to define visual disability, visual impairment and the need for specialist support are needed. The needs of children with visual disability who also have other disabilities should also be a priority.
Recommendations for adoption of standards for measurement and reporting
We have shown the considerable variation in definitions and cut points used in the UK based studies. We strongly recommend the adoption of standards for measurement and reporting which have been recommended by international organisations (WHO and the International Council of Ophthalmology) and are described in Chapter 7.
Vision related quality of life scales are a useful adjunct to clinical measurements, but should not be used as a substitute for visual acuity. We recommend the use of validated questionnaires or scales for measuring self reported vision problems or vision related quality of life. We emphasize the need to thoroughly test all questions before use in surveys.
We encountered difficulties in understanding the methods and basis for calculations in some of the reports we reviewed due to a lack of clarity and transparency in the reporting of data. We recommend that reports which present novel research findings (but are not submitted for publication in academic journals) should conform to the same standards of reporting as the formal peer review process of academic journals (for example as published by the International Council of Medical Editors). Further details are given in Chapter 7.