For Peer Review Only Journal: BMJ Open

BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

A comparison of the causes of blindness certifications in England and Wales in working age adults (16-64 years), 1999-2000 to 2009-2010

For peer review only Journal: BMJ Open

Manuscript ID: bmjopen-2013-004015

Article Type: Research

Date Submitted by the Author: 12-Sep-2013

Complete List of Authors: Liew, Gerald; Moorfields Eye Hospital, Medical Retina Michaelides, Michael; University College London, Bunce, Catey; University College London,

Primary Subject Ophthalmology Heading:

Secondary Subject Heading: Ophthalmology, Public health

Medical retina < OPHTHALMOLOGY, PUBLIC HEALTH, Diabetic retinopathy Keywords: < DIABETES & ENDOCRINOLOGY

http://bmjopen.bmj.com/

on September 27, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 A comparison of the causes of blindness certifications in England and Wales in 3 4 5 working age adults (16-64 years), 1999-2000 to 2009-2010. 6 7 Short title: Blindness in England and Wales 2009-2010 8 9 Gerald Liew, Michel Michaelides, Catey Bunce 10 11 12 13 14 Moorfields Eye Hospital NHS Foundation Trust, London. EC1V 2PD, UK. 15 Gerald Liew, MedicalFor Retina Researchpeer Fellow; review only 16 Michel Michaelides, Consultant Ophthalmologist. 17 18 Centre for Vision Research, Westmead Millenium Institute, University of Sydney, Sydney, Australia. 19 Gerald Liew, Hononary Senior Lecturer 20 21 Department of Genetics, University College London, Institute of Ophthalmology, Bath Street, London EC1V 22 9EL, UK; 23 Michel Michaelides, Clinical Senior Lecturer 24

25 26 National Institute for Health Research Biomedical Research Centre at Moorfields Eye Hospital NHS 27 Foundation Trust and UCL Institute of Ophthalmology. 28 Catey Bunce, Senior Statistician 29 30 31 Financial Disclosure(s): 32 The authors have no proprietary or commercial interest in any materials discussed in this article. 33 34 Conflict of interest: http://bmjopen.bmj.com/ 35 The authors have read and understood the BMJ Group policy on declaration of interests and declare no 36 conflict of interest. All authors have completed the Unified Competing Interest form at 37 www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare that 38 (1) No authors had support from any commercial companies for the submitted work; (2) GL received a 39 40 fellowship from the Royal Australia New Zealand College of Ophthalmologists (RANZCO) Eye 41 Foundation/Novartis scholarship for unrelated work during the period this report was written. No other authors have relationships with any companies that might have an interest in the submitted work in the 42 on September 27, 2021 by guest. Protected copyright. 43 previous 3 years; (3) No authors, their spouses, partners, or children have financial relationships that may 44 be relevant to the submitted work; and (4) No authors have no non-financial interests that may be relevant 45 to the submitted work. 46 47 Sources of Funding: 48 This study was supported by grants from the RNIB and the National Institute for Health Research 49 Biomedical Research Centre at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of 50 Ophthalmology. The data captured by CVI are DH copyright and this work was made possible by 51 collaboration with the Royal College of Ophthalmologists. The views expressed in this report are those of 52 53 the authors and do not necessarily reflect those of the funding bodies. MM is a recipient of a Foundation 54 Fighting Blindness Career Development Award. 55 56 Licensing 57 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all 58 authors, a worldwide licence to the Publishers and its licensees in perpetuity, in all forms, formats and 59 60 1

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 media (whether known now or created in the future), to i) publish, reproduce, distribute, display and store 3 the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include 4 within collections and create summaries, extracts and/or, abstracts of the Contribution, iii) create any other 5 derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the 6 inclusion of electronic links from the Contribution to third party material where-ever it may be located; 7 and, vi) licence any third party to do any or all of the above. 8 9 10 11 Corresponding author: 12 C. Bunce, Research and Development, Moorfields Eye Hospital, City Road, London EC1V2PD, UK 13 14 Tel: +442075662820 15 Fax: +442086086925For peer review only 16 Email: [email protected] 17 18 19 Word count: 2367 20 21 22 Abstract: 233 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Abstract 3 4 Objectives. To report on the causes of blindness certifications in England and Wales in working age adults 5 6 (16-64 years) in 2009-2010; and to compare these with figures from 1999-2000. 7 8 Design. Analysis of the national database of blindness certificates of vision impairment (CVIs) received by 9 10 the Certifications Office. 11 12 Setting and Participants. Working age (16-64 years) population of England and Wales. 13 14 15 Main outcome measuresFor. Number peer and cause ofreview blindness certifications. only 16 17 Results. The Certifications Office received 1756 CVIs for blindness from persons aged between 16 and 64 18 19 inclusive between 2009-2010. The main causes of blindness certifications were hereditary retinal disorders 20 21 (354 certifications comprising 20.2% of the total), diabetic retinopathy/maculopathy (253 persons, 14.4%), 22 23 and optic atrophy (248 persons, 14.1%). Together these three leading causes accounted for almost 50% of 24 25 all blindness certifications. In 1999-2000, the leading causes of blindness certification were diabetic 26 27 28 retinopathy/maculopathy (17.7%), hereditary retinal disorders (15.8%) and optic atrophy (10.1%). 29 30 Conclusions. For the first time in at least five decades, diabetic retinopathy/maculopathy is no longer the 31 32 leading cause of certifiable blindness in England and Wales, having been overtaken by inherited retinal 33 34 disorders. This change may be related to factors including the introduction of nationwide diabetic http://bmjopen.bmj.com/ 35 36 retinopathy screening programmes in England and Wales, and improved glycaemic control. Inherited 37 38 retinal disease now representing the commonest cause of certification in the working age population has 39 40 41 both clinical and research implications, including with respect to the provision of care/resources in the NHS

42 on September 27, 2021 by guest. Protected copyright. 43 and the allocation of research funding. 44 45 46 47 240 words 48 49 50 51

52 53 54 55 56 57 58 59 60 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 “Article Summary” 3 4 Article Focus. 5 6 - Blindness certifications are an important public health indicator in England and Wales. 7 8 - Every year, data on the number of persons certified blind is collected and the main causes listed. 9 10 - In 1999-2000, the leading causes of blindness certifications were diabetic retinopathy/maculopathy, 11 12 hereditary retinal disorders and optic atrophy. We report results for 2009-2010. 13 14 15 For peer review only 16 17 Key messages. 18 19 - We report that for the first time in over five decades, diabetic retinopathy/maculopathy is no longer the 20 21 leading cause of blindness in working age adults, having been overtaken by hereditary retinal disorders. 22 23 - This decline in blindness certifications from diabetic retinopathy/maculopathy may be related to the 24 25 introduction of nationwide public health measures in England and Wales. 26 27 28 - The results have implications for resource allocations for clinical care delivery and research. 29 30 31 32 Strengths and Limitations 33 34 - Strengths of the data include nationwide coverage, collection of uniform data fields with pre-specified tick http://bmjopen.bmj.com/ 35 36 boxes, and clear and consistent definitions of sight impairment. 37 38 - Limitations include comparisons across two slightly different data collection forms. 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 4

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Introduction 3 4 England and Wales have collected data on the number of blind people in the country since 1851.1;2 From 5 6 1930 until 2003, the causes of blindness have also been collected through use of a designated certificate, 7 8 the BD8, which was completed by the examining ophthalmologist.2 The BD8 was replaced in England in 9 10 September 2005 by the Certificate of Vision Impairment (CVI), and in Wales in April 2007 by the equivalent 11 12 form, the CVI-W. One copy of the CVI and CVI-W is then sent to the Certifications Office, London, for 13 14 15 anonymised epidemiologicalFor analysis.peer The Certifications review Office is currently only funded by the RNIB (Royal 16 17 National Institute for the Blind) and operates under the auspices of the Royal College of Ophthalmologists 18 19 with CVI data under crown copyright. 20 21 22 23 Although not compulsory, certification allows patients to be registered (i.e. placed on an official local 24 25 council register) as either severely sight impaired (blind) or sight impaired (partially sighted), which then 26 27 28 permits access to certain state benefits and social service provisions. There is hence an incentive for 29 3;4 30 patients to be certified and counted, providing an estimate, albeit imperfect, of the causes and burden of 31 32 blindness in England and Wales. The importance of this data is highlighted by the fact that certification 33 34 figures have recently been adopted in 2012 as an indicator for preventable sight loss and included in the http://bmjopen.bmj.com/ 35 36 Public Health Outcomes Framework.5 In this report we present the findings from an analysis of working age 37 38 blindness certifications from 2009-2010 and compare these with figures from 1999-2000.6 Major changes in 39 40 41 leading causes of blindness certification in this age group have occurred over this period; subsequent

42 on September 27, 2021 by guest. Protected copyright. 43 publications will report on findings in persons of other age groups. 44 45 46 47 Methods 48 49 Details of data entry, collections and transmission have been reported previously.6-8 For 2009-2010 data 50 51 were obtained from CVI forms, whereas for 1999-2000, data were obtained from BD8 forms. With regards 52 53 54 to CVI forms, data were transcribed from these paper-based forms into a database at the Certification 55 56 Office. Part C of the CVI form collects information on the causes of visual loss and requires the completing 57 58 ophthalmologist to select from a list of common diagnoses. More than one cause of visual loss can be 59 60 5

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 selected in which case the main cause should be highlighted using either an asterisk or circle. Guidelines are 3 4 provided to assist with this process when the main cause is not evident. ‘Multiple’ causes are used where 5 6 the ophthalmologist completing the form has not indicated a single main cause. Possible reasons for this 7 8 selection include differing causes in the two eyes, or more than one cause within one eye and the 9 10 ophthalmologist is unable to determine which contributes the most to the visual loss. Causes for visual 11 12 impairment were coded using International Classification of Diseases (ICD)-9 codes and grouped into 13

14 6-8 15 disease categoriesFor as in previous peer reports. review only 16 17 18 19 For 1999-2000, data were extracted from paper based BD8 forms in a similar fashion.6 Part 5 of the BD8 20 21 form has 16 fields for medical information including one for the main cause of visual loss and five for each 22 23 eye for contributory causes. For records where the main cause was not filled, this field was imputed 24 25 wherever possible using an algorithm published previously.8 In short this algorithm brought forward as the 26 27 28 main cause the contributing cause if any were so listed; if more than one contributory cause was listed, the 29 30 patient was listed as having ‘multiple causes’. The main difference between data collected from the CVI and 31 32 BD8 forms is the number of certifications with ‘multiple causes’ of visual impairment which was higher in 33 34 the CVI forms.7 We have previously minimized bias from this source by re-examining the certificates with http://bmjopen.bmj.com/ 35 36 ‘multiple causes’, extracting the contributory causes listed and tabulating them with the main causes. 7 We 37 38 follow the same procedure in this report. 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 Blindness was defined according to criteria previously described on the BD8 and CVI forms as either best 44 45 corrected visual acuity in the better eye of either (1) worse than 3/60, or (2) worse than 6/60 with severely 46 47 constricted visual fields, or (3) better than 6/60 with severely constricted visual fields particularly the 48 49 inferior field. 50 51

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 on main cause’ of visual impairment had contributory causes extracted and listed, and later combined with 3 4 the main causes. Pie and bar charts were used to graphically illustrate the distribution of the main causes of 5 6 visual impairment. Proportions of blindness registrations due to each cause are presented rather than 7 8 adjusted incidence rates in order to indicate the relative contribution of each condition to the pool of vision 9 10 impairment. Currently almost every grant application for diabetic retinopathy projects commences with a 11 12 statement that diabetic retinopathy is the most common cause of visual loss in the working age population 13 14 15 – we present proportionsFor of blindnesspeer registrations review to determine if this onlystatement remains valid. 16 17 18 19 Results 20 21 For the period April 2009 to March 2010, the Certifications Office received 1756 CVIs for blindness from 22 23 persons aged between 16 and 64 inclusive. This compares with 1637 BD8 forms for blindness received 24 25 between April 1999 and March 2000, details of which have been reported previously.8 Table 1 shows the 26 27 28 number of persons certified blind for each of the disease categories. Hereditary retinal disorders, including 29 30 Stargardt Disease and Retinitis Pigmentosa, formed the largest category with 354 certifications comprising 31 32 20.2% of the total. Diabetic retinopathy/maculopathy was the second largest cause of certifiable blindness 33 34 with 253 persons (14.4%), followed by optic atrophy with 248 persons (14.1%). Together these 3 leading http://bmjopen.bmj.com/ 35 36 causes accounted for almost 50% of all blindness certifications in the working age group. Glaucoma was 37 38 responsible for 104 (5.9%) blindness certifications, followed by congenital abnormalities of the eye which 39 40 41 included congenital cataracts and retinopathy of prematurity (89 certifications, 5.1%). Multiple pathologies

42 on September 27, 2021 by guest. Protected copyright. 43 were listed for 242 persons (13.8%) and no information on the main cause was listed for 42 persons (2.4%). 44 45 When these categories were examined for contributory causes, the most common contributory causes 46 47 recorded were glaucoma (60 persons), diabetic retinopathy/maculopathy (56 persons) and optic atrophy 48 49 (46 persons). Combining the main and contributory causes resulted in small changes in the overall 50 51 proportion of certifications due to specific causes, but did not change the relative rankings of the top six 52 53 54 causes of blindness (Table 1). ‘Other conditions’ comprised 150 certifications (8.5%) of which the most 55 56 common were malignant neoplasms of the brain and nervous system (27 persons, 1.5%) and retinal 57 58 59 60 7

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 detachments (24 persons, 1.4%). Figure 1 shows the distribution of the causes of blindness certifications 3 4 graphically in a pie chart. 5 6 7 8 We next compared the main causes of blindness certifications from 1999-2000 (n=1637) with the figures 9 10 above from 2009-2010 (n=1756). The results are shown graphically in Figure 2. From 1999-2000, the 11 12 leading cause of blindness certification was diabetic retinopathy/maculopathy which accounted for 290 13 14 15 certifications (17.7%).For By 2009-2010, peer this figure review had decreased to 253 (14.4%),only and diabetic 16 17 retinopathy/maculopathy was now the second leading cause of blindness certification. In contrast, 18 19 hereditary retinal disorders which were the second leading cause of blindness certification in 1999-2000 20 21 accounting for 258 cases (15.8% of total), had increased to 354 cases (20.2%) by 2009-2010 and have now 22 23 become the leading cause of certifiable blindness in the working age group in England and Wales. Optic 24 25 atrophy remained the third leading cause in 1999-2000 and 2009-2010 with an increase from 165 cases 26 27 28 (10.1%) to 248 cases (14.1%) respectively. A notable finding was that degeneration of the macula and 29 30 posterior pole, which accounted for 7.7% of blindness registration in 1999-2000, had dropped in 31 32 percentage terms and now accounted for only 3.0% by 2009-2010. Other causes of blindness registration 33 34 remained roughly similar for the two time periods. http://bmjopen.bmj.com/ 35 36 37 38 Discussion 39 40 41 This report provides updated estimates on the causes of certifiable blindness in England and Wales in

42 on September 27, 2021 by guest. Protected copyright. 43 working age adults. Three main diseases were responsible for half of all certifications – hereditary retinal 44 45 disorders (20.2%), diabetic retinopathy/maculopathy (14.4%) and optic atrophy (14.1%). A marked change 46 47 has occurred in the relative importance of these main causes of blindness certifications since the last major 48 49 analysis in 1999-2000,6 with diabetic retinopathy/maculopathy now no longer the leading cause of 50 51 blindness in working age adults. Since at least 1963,9 diabetic retinopathy/maculopathy has been the 52 53 54 leading cause of blindness among working age adults in England and Wales; a similar situation exists in 55 10 56 other developed countries such as the United States. Over the past decade the proportion of 57 58 59 60 8

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

14 11 15 from diabetic eyeFor disease. These peer are known as reviewthe NHS Diabetic Eye Screening only Program (England) and 16 12 17 Diabetic Retinopathy Screening Service Wales (DRSSW) , and these programmes annually screen almost 2 18 19 million and 150,000 patients with diabetes, respectively. Concurrent with these screening programmes, in 20 21 2004 the Quality and Outcomes Framework13 was introduced to incentivise general practitioners in the 22 23 United Kingdom to improve primary care management of several conditions including diabetes. This effort 24 25 may have contributed to the improvement in glycaemic control documented since the late 1990s.13;14 The 26 27 28 decline in both the absolute number and relative proportion of blindness certifications due to diabetic 29 30 retinopathy/maculopathy since introduction of these public health measures may be an indicator of their 31 32 effectiveness. 33 34 http://bmjopen.bmj.com/ 35 36 Whether increased rates of certification for inherited eye diseases reflects improved certification of existing 37 38 sight impairment, or a true increase in incidence of these disorders is unclear. The progress made over the 39 40 41 last decade in molecular genetics/diagnostics and the increasing avenues of research/clinical trials for

42 on September 27, 2021 by guest. Protected copyright. 43 inherited retinal disease with widespread media coverage may plausibly have resulted in increasing clinic 44 45 visits and thereby registration,4 without a true increase in incidence rates. An observation in favour of this 46 47 scenario is that the rates of blindness certification for optic atrophy have also increased over the last 48 49 decade, in tandem with those for inherited retinal disease, while those for other non-inherited conditions 50 51 such as glaucoma have remained fairly constant. Hereditary retinal diseases occur more frequently in 52 53 54 communities with a higher rate of consanguinity, and it is conceivable that increased rates of immigration 55 56 from countries where consanguinity is more prevalent may have contributed to these findings, though at 57 58 this stage this remains speculative. 59 60 9

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 4 These findings have implications for clinical care and research budget allocation. A prolonged focus on 5 6 prevention and treatment of diabetic eye disease has likely contributed to the decline in blindness 7 8 certifications from this disorder, and the rate is expected to decline further with the recent National 9 10 Institute for Clinical Excellence (NICE) approval of ranibizumab for treatment of diabetic maculopathy. Now 11 12 that hereditary retinal diseases comprise the leading cause of blindness certifications, an increased focus 13 14 15 on clinical managementFor of these peer conditions (e.g. review with low vision aids, visualonly rehabilitation) and greater 16 17 allocation of research funding to study these disorders may be appropriate. Funding bodies may need to re- 18 19 asses their funding priorities. 20 21 22 23 Strengths and Limitations 24 25 Data from BD8 and CVI registrations represent some of the best available epidemiological data on sight 26

27 5 28 impairment in England and Wales and is regarded as a major public health indicator. Strengths of the data 29 30 include nationwide coverage, collection of uniform data fields with pre-specified tick boxes, and clear and 31 32 consistent definitions of sight impairment. 33 34 http://bmjopen.bmj.com/ 35 36 Limitations that should be highlighted include that fact that blindness certifications are not equivalent to 37 38 blindness rates. This has been discussed previously3;4 and it is estimated that up to 53% of eligible patients 39

40 7 41 may not be registered blind despite consultation with an ophthalmologist. However, arguments have been

42 on September 27, 2021 by guest. Protected copyright. 7 43 advanced that in time, most patients eligible to be registered will in fact do so, and studies have shown a 44 45 major increase in registration rates with increasing clinic visits.4 46 47 48 49 Another caveat to these results is that some patients who are certified blind may not always satisfy all of 50 51 the official criteria, with one study suggesting an inappropriate blindness certification rate of 23%.3 Such 52 53 54 inappropriate certifications may inflate the numbers somewhat but it should be borne in mind that the aim 55 56 of certification is not to identify persons meeting rigid clinical criteria but to identify and count those with 57 58 significant visual impairment who may benefit from state assistance. Indeed, current guidelines for 59 60 10

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 completion of CVI forms state that the criteria should be interpreted in the context of the patient’s 3 4 functional status rather than as strict cut-offs. Another caveat when interpreting these results is that the 5 6 figures for the two time periods were collated from different forms. The differences in these forms are 7 8 discussed elsewhere7 and one of the main complications in comparing temporal trends is the increase in 9 10 the number of forms where a main cause has not been identified. In the 1999-2000 dataset, which was 11 12 derived from BD8, approximately 4% of forms had ‘multiple pathology’; in the 2009-2010 dataset derived 13 14 15 from CVI, this hadFor increased topeer 14%. This raised review the possibility that diabetic only retinopathy/maculopathy may 16 17 have been under-reported for the 2009-2010 period. We attempted to address this by examining the 18 19 contributory causes in those without a main cause recorded and using these in place of the missing main 20 21 cause; this analysis resulted in only small changes to the percentage of blindness due to each cause and did 22 23 not change the overall ranking of the top six causes. This suggests that the rate of under-reporting of main 24 25 causes was similar for most categories and not responsible for the shift in the leading causes of blindness 26 27 28 certifications. 29 30 31 32 In summary, this report found three main causes were responsible for half of all blindness certifications 33 34 among working age adults in England and Wales from 2009-2010 - hereditary retinal disorders (20.2%), http://bmjopen.bmj.com/ 35 36 followed by diabetic retinopathy/maculopathy (14.4%) and optic atrophy (14.1%). This marks the first time 37 38 in almost five decades that diabetic retinopathy/maculopathy is no longer the leading cause of blindness in 39 40 41 working age adults, and may be related to the introduction of nationwide public health measures in

42 on September 27, 2021 by guest. Protected copyright. 43 England and Wales. The results have implications for resource allocations for clinical care delivery and 44 45 research. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 11

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Table 1. Numbers of working age adults (age 16-64) with severe sight impairment (blindness) in England 3 4 and Wales: certifications 2009-2010. The ‘Main cause’ column lists the number of certifications with the 5 6 corresponding diagnosis; the ‘Contributory cause’ column lists the contributory causes in certifications from 7 8 the ‘Multiple pathology’ and ‘No Information on main cause’ categories. 9 10 ICD-9 Codes Diagnosis Main cause Contributory Combined 11 12 (% Total) cause (% Total) 13 14 15 For peer review only(% Total) 16 17 18 19 362.7 Hereditary retinal disorders 354 (20.2) 29 (6.6) 383 (20.0) 20 21 362/34000 Diabetic retinopathy/maculopathy 253 (14.4) 56 (12.8) 309 (16.2) 22 23 24 377.1 Optic atrophy 248 (14.1) 46 (10.5) 294 (15.4) 25 26 365 Glaucoma 104 (5.9) 60 (13.7) 164 (8.6) 27 28 29 743-760 Congenital abnormalities of the eye 89 (5.1) 32 (7.3) 121 (6.3) 30 31 377.7 Disorders of visual cortex 72 (4.1) 24 (5.5) 96 (5.0) 32 33 430-438 Cerebrovascular disease 56 (3.2) 21 (4.8) 77 (4.0) 34 http://bmjopen.bmj.com/ 35 36 362.5 Degeneration of macula and posterior pole 52 (3.0) 14 (3.2) 66 (3.5) 37 38 360.2 Myopia 49 (2.8) 23 (5.2) 72 (3.8) 39 40 370-371 Corneal disorders 45 (2.6) 34 (7.7) 79 (4.1) 41

42 on September 27, 2021 by guest. Protected copyright. 43 - Multiple pathology 242 (13.8) - - 44 45 - Other conditions 150 (8.5) 100 (22.8) 250 (13.1) 46 47 - No information on main cause 42 (2.4) - - 48 49 50 Total 1756 439* 1911* 51 52 ICD refers to International Classification of Disease. 53 54 *Including contributory causes. The total number of contributory causes is greater than the sum of 55 ‘Multiple pathology’ and ‘No information on main cause’ categories because persons can have between 56 one and four contributory causes documented. 57 58 59 60 12

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Figure 1. Main causes of severe sight impairment (blindness) in England and Wales in working age adults 3 4 (age 16-64): certifications 2009-2010. 5 6 7 Main Causes of Severe Sight Impairment 8 9 (Blindness) Certifications 2009-10 10 11 Age 16-64 years 12 Corneal disorders 13 3% 14 Degeneration of the No information on 15 maculaFor and posterior peer mainreview cause only 16 pole 2% 17 3% Myopia 3% 18 19 Cerebrovascular 20 disease 21 3% 22 Hereditary retinal 23 Disorders of disorders 24 visual cortex 20% 25 4% 26 27 28 29 Congenital 30 anomalies of eye 31 5% 32 33

34 Glaucoma http://bmjopen.bmj.com/ 35 6% Diabetic retinopathy 36 14% 37 38 39 Other conditions 40 9% 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 Optic atrophy 46 Multiple pathology 14% 47 14% 48 49 50 51 52 53 54 55 56 57 58 59 60 13

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Figure 2. 10-year change in causes of severe sight impairment (blindness) in England and Wales in working 3 4 age adults (age 16-64): certifications 1999-2000 and 2009-2010. 5 6 7 Change in Main Causes of Blindness Certification 8 9 from 2000 to 2010 10 11 25 12 13 14 20 15 For20.2 peer review only 16 17 17.7 18 15 15.8 19 14.4 14.1 20

21 % of Total 10 22 10.1 23 24 7.7 25 5 1999-00 5.45.9 26 5.1 5.1 4.9 2009-10 27 2.9 3.2 3 2.8 28 0 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

34 http://bmjopen.bmj.com/ 35 (11) NHS Diabetic Eye Screening Programme. http://diabeticeye screening nhs uk/about [ 2013 36 37 (12) Diabetic Retinopathy Screening Service Wales. http://www cardiffandvaleuhb wales nhs uk/drssw [ 38 2013 39 40 (13) Calvert M, Shankar A, McManus RJ, Lester H, Freemantle N. Effect of the quality and outcomes 41 framework on diabetes care in the United Kingdom: retrospective cohort study. BMJ 2009;

42 338:b1870. on September 27, 2021 by guest. Protected copyright. 43 44 (14) Oluwatowoju I, Abu E, Wild SH, Byrne CD. Improvements in glycaemic control and cholesterol 45 concentrations associated with the Quality and Outcomes Framework: a regional 2-year audit of 46 diabetes care in the UK. Diabet Med 2010; 27(3):354-359. 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 STROBE Statement—checklist of items that should be included in reports of observational studies 3 4 Item 5 No Recommendation 6 Title and abstract 7 1 (a) Indicate the study’s design with a commonly used term in the title or the abstract 8 Y (b) Provide in the abstract an informative and balanced summary of what was done 9 and what was found 10 Introduction 11 12 Background/rationale 2 Explain the scientific background and rationale for the investigation being reported 13 Y 14 Objectives 3 State specific objectives, including any prespecified hypotheses 15 Y For peer review only 16 17 Methods 18 Study design 4 Present key elements of study design early in the paper 19 Y 20 21 Setting 5 Describe the setting, locations, and relevant dates, including periods of recruitment, 22 Y exposure, follow-up, and data collection 23 Participants 6 (a) Cohort study—Give the eligibility criteria, and the sources and methods of 24 Y selection of participants. Describe methods of follow-up 25 26 Case-control study—Give the eligibility criteria, and the sources and methods of 27 case ascertainment and control selection. Give the rationale for the choice of cases 28 and controls 29 Cross-sectional study—Give the eligibility criteria, and the sources and methods of 30 selection of participants 31

32 (b) Cohort study—For matched studies, give matching criteria and number of 33 exposed and unexposed

34 Case-control study—For matched studies, give matching criteria and the number of http://bmjopen.bmj.com/ 35 controls per case 36 37 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, and effect 38 Y modifiers. Give diagnostic criteria, if applicable 39 Data sources/ 8* For each variable of interest, give sources of data and details of methods of 40 measurement assessment (measurement). Describe comparability of assessment methods if there 41 Y is more than one group

42 on September 27, 2021 by guest. Protected copyright. 43 Bias 9 Describe any efforts to address potential sources of bias 44 Y 45 Study size 10 Explain how the study size was arrived at 46 NA 47 48 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, 49 NA describe which groupings were chosen and why 50 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding 51 Y (b) Describe any methods used to examine subgroups and interactions 52 53 (c) Explain how missing data were addressed 54 (d) Cohort study—If applicable, explain how loss to follow-up was addressed 55 Case-control study—If applicable, explain how matching of cases and controls was 56 addressed 57 Cross-sectional study—If applicable, describe analytical methods taking account of 58 59 sampling strategy 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml1 Page 17 of 18 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 (e) Describe any sensitivity analyses 3 Continued on next page 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml2 BMJ Open Page 18 of 18 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 Results 4 5 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, 6 Y examined for eligibility, confirmed eligible, included in the study, completing follow-up, and 7 analysed 8 (b) Give reasons for non-participation at each stage 9 (c) Consider use of a flow diagram 10 11 Descriptive 14* (a) Give characteristics of study participants (eg demographic, clinical, social) and information 12 data on exposures and potential confounders 13 Y (b) Indicate number of participants with missing data for each variable of interest 14 (c) Cohort study—Summarise follow-up time (eg, average and total amount) 15 For peer review only 16 Outcome data 15* Cohort study—Report numbers of outcome events or summary measures over time 17 Y Case-control study—Report numbers in each exposure category, or summary measures of 18 exposure 19 Cross-sectional study—Report numbers of outcome events or summary measures 20 21 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their 22 Y precision (eg, 95% confidence interval). Make clear which confounders were adjusted for and 23 why they were included 24 (b) Report category boundaries when continuous variables were categorized 25 26 (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful 27 time period 28 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and sensitivity 29 Y analyses 30 31 Discussion 32 Key results 18 Summarise key results with reference to study objectives 33 Y

34 http://bmjopen.bmj.com/ Limitations 19 Discuss limitations of the study, taking into account sources of potential bias or imprecision. 35 36 Y Discuss both direction and magnitude of any potential bias 37 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity 38 Y of analyses, results from similar studies, and other relevant evidence 39 Generalisability 21 Discuss the generalisability (external validity) of the study results 40 41 Y

42 Other information on September 27, 2021 by guest. Protected copyright. 43 Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, 44 Y for the original study on which the present article is based 45 46 47 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and 48 unexposed groups in cohort and cross-sectional studies. 49 50 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and 51 52 published examples of transparent reporting. The STROBE checklist is best used in conjunction with this article (freely 53 available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 54 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is 55 available at www.strobe-statement.org. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml3 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

A comparison of the causes of blindness certifications in England and Wales in working age adults (16-64 years), 1999-2000 with 2009-2010

For peer review only Journal: BMJ Open

Manuscript ID: bmjopen-2013-004015.R1

Article Type: Research

Date Submitted by the Author: 05-Dec-2013

Complete List of Authors: Liew, Gerald; Moorfields Eye Hospital, Medical Retina Michaelides, Michael; University College London, Bunce, Catey; University College London,

Primary Subject Ophthalmology Heading:

Secondary Subject Heading: Ophthalmology, Public health

Medical retina < OPHTHALMOLOGY, PUBLIC HEALTH, Diabetic retinopathy Keywords: < DIABETES & ENDOCRINOLOGY

http://bmjopen.bmj.com/

on September 27, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 A comparison of the causes of blindness certifications in England and Wales in working 3 4 age adults (16-64 years), 1999-2000 with 2009-2010 5 6 Short title: Blindness in England and Wales 2009/2010 7 8 Gerald Liew, Michel Michaelides*, Catey Bunce* 9 10 11 Moorfields Eye Hospital NHS Foundation Trust, London. EC1V 2PD, UK. 12 Gerald Liew, Medical Retina Research Fellow; 13 Michel Michaelides, Consultant Ophthalmologist. 14

15 Centre for VisionFor Research, Westmead peer Millenium review Institute, University ofonly Sydney, Sydney, Australia. 16 17 Gerald Liew, Hononary Senior Lecturer 18 19 Department of Genetics, University College London, Institute of Ophthalmology, Bath Street, London EC1V 20 9EL, UK; 21 Michel Michaelides, Clinical Senior Lecturer 22 23 National Institute for Health Research Biomedical Research Centre at Moorfields Eye Hospital NHS 24 Foundation Trust and UCL Institute of Ophthalmology. 25 Catey Bunce, Senior Statistician 26 27 28 Corresponding authors: 29 30 C. Bunce, Research and Development, Moorfields Eye Hospital, City Road, London EC1V2PD, UK 31 Tel: +442075662820 32 Fax: +442086086925 33 Email: [email protected]

34 http://bmjopen.bmj.com/ 35 M. Michaelides, Department of Genetics, University College London, Institute of Ophthalmology, Bath 36 Street, London EC1V 9EL, UK 37 Tel: +442076086864 38 Email: [email protected] 39 40 Word count: 2367 41

42 on September 27, 2021 by guest. Protected copyright. 43 Abstract: 233 44

45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Abstract 3 4 Objectives. To report on the causes of blindness certifications in England and Wales in working age adults 5 6 (16-64 years) in 2009-2010; and to compare these with figures from 1999-2000. 7 8 Design. Analysis of the national database of blindness certificates of vision impairment (CVIs) received by 9 10 the Certifications Office. 11 12 Setting and Participants. Working age (16-64 years) population of England and Wales. 13 14 15 Main outcome measuresFor. Number peer and cause ofreview blindness certifications. only 16 17 Results. The Certifications Office received 1756 CVIs for blindness from persons aged between 16 and 64 18 19 inclusive between 1st April 2009 to 31st March 2010. The main causes of blindness certifications were 20 21 hereditary retinal disorders (354 certifications comprising 20.2% of the total), diabetic 22 23 retinopathy/maculopathy (253 persons, 14.4%), and optic atrophy (248 persons, 14.1%). Together these 24 25 three leading causes accounted for almost 50% of all blindness certifications. Between 1st April 1999 to 31st 26 27 28 March 2000, the leading causes of blindness certification were diabetic retinopathy/maculopathy (17.7%), 29 30 hereditary retinal disorders (15.8%) and optic atrophy (10.1%). 31 32 Conclusions. For the first time in at least five decades, diabetic retinopathy/maculopathy is no longer the 33 34 leading cause of certifiable blindness amongst working age adults in England and Wales, having been http://bmjopen.bmj.com/ 35 36 overtaken by inherited retinal disorders. This change may be related to factors including the introduction of 37 38 nationwide diabetic retinopathy screening programmes in England and Wales, and improved glycaemic 39 40 41 control. Inherited retinal disease now representing the commonest cause of certification in the working age

42 on September 27, 2021 by guest. Protected copyright. 43 population has both clinical and research implications, including with respect to the provision of 44 45 care/resources in the NHS and the allocation of research funding. 46 47 48 49 240 words 50 51

52 53 54 55 56 57 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Introduction 3 4 England and Wales have collected data on the number of blind people in the country since 1851.1;2 From 5 6 1930 until 2003, the causes of blindness have also been collected through use of a designated certificate, 7 8 the BD8, which was completed by the examining ophthalmologist.2 The BD8 was replaced in England in 9 10 September 2005 by the Certificate of Vision Impairment (CVI), and in Wales in April 2007 by the equivalent 11 12 form, the CVI-W. One copy of the CVI and CVI-W is then sent to the Certifications Office, London, for 13 14 15 anonymised epidemiologicalFor analysis.peer The Certifications review Office is currently only funded by the RNIB (Royal 16 17 National Institute for the Blind) and operates under the auspices of the Royal College of Ophthalmologists 18 19 with CVI data under Crown copyright, meaning the copyright is owned by the British Government. 20 21 22 23 Although not compulsory, certification allows patients to be registered (i.e. placed on an official local 24 25 council register) as either severely sight impaired (blind) or sight impaired (partially sighted), which then 26 27 28 permits access to certain state benefits and social service provisions. There is hence an incentive for 29 3;4 30 patients to be certified and counted, providing an estimate, albeit imperfect, of the causes and burden of 31 32 blindness in England and Wales. The importance of these data is highlighted by the fact that certification 33 34 figures have recently been adopted in 2012 as an indicator for preventable sight loss and included in the http://bmjopen.bmj.com/ 35 36 Public Health Outcomes Framework.5 In this report we present the findings from an analysis of working age 37 38 blindness certifications from 2009-2010 and compare these with figures from 1999-2000.6 Major changes in 39 40 41 leading causes of blindness certification in this age group have occurred over this period; subsequent

42 on September 27, 2021 by guest. Protected copyright. 43 publications will report on findings in persons of other age groups. 44 45 46 47 Methods 48 49 Details of data entry, collections and transmission have been reported previously.6-8 Between 1st April 2009 50 51 to 31st March 2010 data were obtained from CVI forms, whereas between 1st April 1999 to 31st March 2000, 52 53 54 data were obtained from BD8 forms. With regards to CVI forms, data were transcribed from these paper- 55 56 based forms into a database at the Certification Office. Part C of the CVI form collects information on the 57 58 causes of visual loss and requires the completing ophthalmologist to select from a list of common 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 diagnoses. More than one cause of visual loss can be selected in which case the main cause should be 3 4 highlighted using either an asterisk or circle. Guidelines are provided to assist with this process when the 5 6 main cause is not evident. ‘Multiple’ causes are used where the ophthalmologist completing the form has 7 8 not indicated a single main cause. Possible reasons for this selection include differing causes in the two 9 10 eyes, or more than one cause within one eye and the ophthalmologist is unable to determine which 11 12 contributes the most to the visual loss. Causes for visual impairment were coded using International 13

14 6-8 15 Classification of DiseasesFor (ICD)-9 peer codes and grouped review into disease categories only as in previous reports. 16 17 18 19 For 1999-2000, data were extracted from paper based BD8 forms in a similar fashion.6 Part 5 of the BD8 20 21 form has 16 fields for medical information including one for the main cause of visual loss and five for each 22 23 eye for contributory causes. For records where the main cause was not filled, this field was imputed 24 25 wherever possible using an algorithm published previously.8 In short this algorithm brought forward as the 26 27 28 main cause the contributing cause if any were so listed; if more than one contributory cause was listed, the 29 30 patient was listed as having ‘multiple causes’. The main difference between data collected from the CVI and 31 32 BD8 forms is the number of certifications with ‘multiple causes’ of visual impairment which was higher in 33 34 the CVI forms.7 We have previously minimized bias from this source by re-examining the certificates with http://bmjopen.bmj.com/ 35 36 ‘multiple causes’, extracting the contributory causes listed and tabulating them with the main causes. 7 We 37 38 follow the same procedure in this report. 39 40 41

52 53 54 Analyses 55 56 Data regarding the main cause of visual impairment were extracted from the forms and grouped into 57 58 disease categories as previously described.6-8 For CVI data, those with ‘multiple causes’ or ‘no information 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

14 2 15 – we present proportionsFor of blindnesspeer registrations review to determine if this onlystatement remains valid. Χ tests 16 17 were performed to test differences in proportions. 18 19 20 21 Results 22 23 For the period 1st April 2009 to 31st March 2010, the Certifications Office received 1756 CVIs for blindness 24 25 from persons aged between 16 and 64 inclusive. This compares with 1637 BD8 forms for blindness received 26

27 st st 8 28 between 1 April 1999 and 31 March 2000, details of which have been reported previously. Table 1 29 30 shows the number of persons certified blind for each of the disease categories. Hereditary retinal 31 32 disorders, including Stargardt Disease and Retinitis Pigmentosa, formed the largest category with 354 33 34 certifications comprising 20.2% of the total. Diabetic retinopathy/maculopathy was the second largest http://bmjopen.bmj.com/ 35 36 cause of certifiable blindness with 253 persons (14.4%), followed by optic atrophy with 248 persons 37 38 (14.1%). Together these 3 leading causes accounted for almost 50% of all blindness certifications in the 39 40 41 working age group. Glaucoma was responsible for 104 (5.9%) blindness certifications, followed by

42 on September 27, 2021 by guest. Protected copyright. 43 congenital abnormalities of the eye which included congenital cataracts and retinopathy of prematurity (89 44 45 certifications, 5.1%). Multiple pathologies were listed for 242 persons (13.8%) and no information on the 46 47 main cause was listed for 42 persons (2.4%). When these categories were examined for contributory 48 49 causes, the most common contributory causes recorded were glaucoma (60 persons), diabetic 50 51 retinopathy/maculopathy (56 persons) and optic atrophy (46 persons). Combining the main and 52 53 54 contributory causes resulted in small changes in the overall proportion of certifications due to specific 55 56 causes, but did not change the relative rankings of the top six causes of blindness (Table 1). ‘Other 57 58 conditions’ comprised 150 certifications (8.5%) of which the most common were malignant neoplasms of 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 the brain and nervous system (27 persons, 1.5%) and retinal detachments (24 persons, 1.4%). Figure 1 3 4 shows the distribution of the causes of blindness certifications graphically in a pie chart. 5 6 7 8 We next compared the main causes of blindness certifications from 1999-2000 (n=1637) with the figures 9 10 above from 2009-2010 (n=1756). The results are shown graphically in Figure 2. From 1999-2000, the 11 12 leading cause of blindness certification was diabetic retinopathy/maculopathy which accounted for 290 13 14 15 certifications (17.7%).For By 2009-2010, peer this figure review had decreased to 253 (14.4%),only and diabetic 16 17 retinopathy/maculopathy was now the second leading cause of blindness certification. This difference was 18 19 statistically significant (P=0.009). In contrast, hereditary retinal disorders which were the second leading 20 21 cause of blindness certification in 1999-2000 accounting for 258 cases (15.8% of total), had increased to 22 23 354 cases (20.2%) by 2009-2010 and have now become the leading cause of certifiable blindness in the 24 25 working age group in England and Wales. Optic atrophy remained the third leading cause in 1999-2000 and 26 27 28 2009-2010 with an increase from 165 cases (10.1%) to 248 cases (14.1%) respectively. A notable finding 29 30 was that degeneration of the macula and posterior pole, which accounted for 7.7% of blindness registration 31 32 in 1999-2000, had dropped in percentage terms and now accounted for only 3.0% by 2009-2010. Other 33 34 causes of blindness registration remained roughly similar for the two time periods. http://bmjopen.bmj.com/ 35 36 37 38 Discussion 39 40 41 This report provides updated estimates on the causes of certifiable blindness in England and Wales in

1 8 2 certifications for hereditary retinal disorders has slowly risen, while that for diabetic 3 4 retinopathy/maculopathy has reduced, resulting in the two conditions swapping rankings. 5 6 7 8 This report is not designed to identify the reasons behind these changes or estimate the incidence of 9 10 blindness from diabetes. The prevalence of diabetes in the UK is not known with certainty, with several 11 12 limited general practice surveys suggesting a range of between 1-2% of the general population11-14; we are 13 14 15 thus not able to estimateFor the incidencepeer of blindness review from diabetes. Nonetheless, only available data suggest the 16 11;13 17 prevalence of diabetes in England and Wales has increased over the period in question , which would be 18 19 expected to lead to increased rates of blindness if other factors remained constant. In this context we 20 21 speculate that several intervening public health developments may have contributed to the reduction in 22 23 both absolute and proportional rates of registrable blindness from diabetes amongst working age adults. 24 25 Between 2003-2008 both England and Wales introduced nationwide diabetic retinopathy screening 26 27 28 services with the aim of reducing the incidence of blindness from diabetic eye disease. These are known as 29 15 30 the NHS Diabetic Eye Screening Program (England) and Diabetic Retinopathy Screening Service Wales 31 32 (DRSSW)16, and these programmes annually screen almost 2 million and 150,000 patients with diabetes, 33 34 respectively. Concurrent with these screening programmes, in 2004 the Quality and Outcomes http://bmjopen.bmj.com/ 35 36 Framework17 was introduced to incentivise general practitioners in the United Kingdom to improve primary 37 38 care management of several conditions including diabetes. Several studies have documented an 39

40 18;19 41 improvement in the quality of care for diabetes since this was introduced, and the effort may have

42 on September 27, 2021 by guest. Protected copyright. 17;20 43 contributed to the improvement in glycaemic control documented since the late 1990s. The decline in 44 45 both the absolute number and relative proportion of blindness certifications due to diabetic 46 47 retinopathy/maculopathy amongst working age adults since introduction of these public health measures 48 49 may be an indicator of their effectiveness. Nonetheless this remains speculative at present, and such 50 51 explanations should be read with caution. 52 53 54 55 56 Whether increased numbers of certification for inherited eye diseases reflects improved certification of 57 58 existing sight impairment, or a true increase in incidence of these disorders is unclear. The progress made 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 over the last decade in molecular genetics/diagnostics and the increasing avenues of research/clinical trials 3 4 for inherited retinal disease with widespread media coverage may plausibly have resulted in higher 5 6 community awareness and increased clinic visits and thereby registration,4 without a true increase in 7 8 incidence rates. An observation in favour of this scenario is that the numbers of blindness certification for 9 10 optic atrophy have also increased over the last decade, in tandem with those for inherited retinal disease, 11 12 while those for other non-inherited conditions such as glaucoma have remained fairly constant. Hereditary 13 14 15 retinal diseases occurFor more frequently peer in communities review with a higher rate only of consanguinity, and it is 16 17 conceivable that increased rates of immigration from countries where consanguinity is more prevalent may 18 19 have contributed to these findings, though at this stage this remains speculative. Another possibility is that 20 21 diagnostic transfer or misclassification may have occurred, for example where some cases of hereditary 22 23 retinal disorders may have been mislabeled as 'degeneration of macula and posterior pole'. In order to 24 25 explain the increase in hereditary retinal disorders, this would have had to occur preferentially in 1999- 26 27 28 2000 versus 2009-2010. However, misclassification of diabetic retinopathy/maculopathy as hereditary 29 30 retinal disorders is unlikely to occur given how different the conditions are, and so would not explain the 31 32 absolute reduction in the number of certifications for blindness due to diabetes. 33 34 http://bmjopen.bmj.com/ 35 36 37 38 These findings have implications for clinical care and research budget allocation. A prolonged focus on 39 40 41 prevention and treatment of diabetic eye disease has likely contributed to the decline in blindness

42 on September 27, 2021 by guest. Protected copyright. 43 certifications from this disorder amongst working age adults, and the rate is expected to decline further 44 45 with the recent National Institute for Clinical Excellence (NICE) approval of ranibizumab for treatment of 46 47 diabetic maculopathy. Now that hereditary retinal diseases comprise the leading cause of blindness 48 49 certifications in working age adults, an increased focus on clinical management of these conditions (e.g. 50 51 with low vision aids, visual rehabilitation) and greater allocation of research funding to study these 52 53 54 disorders may be appropriate. Funding bodies may need to re-asses their funding priorities. 55 56 57 58 Strengths and Limitations 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Data from BD8 and CVI registrations represent some of the best available epidemiological data on sight 3 4 impairment in England and Wales and is regarded as a major public health indicator.5 Strengths of the data 5 6 include nationwide coverage, collection of uniform data fields with pre-specified tick boxes, and clear and 7 8 consistent definitions of sight impairment. 9 10 11 12 Limitations that should be highlighted include that fact that blindness certifications are not equivalent to 13

14 3;4 15 blindness rates. ThisFor has been peerdiscussed previously review and it is estimated only that up to 53% of eligible patients 16 7 17 may not be registered blind despite consultation with an ophthalmologist. However, arguments have been 18 19 advanced that in time, most patients eligible to be registered will in fact do so,7 and studies have shown a 20 21 major increase in registration rates with increasing clinic visits.4 22 23 24 25 Another caveat to these results is that some patients who are certified blind may not always satisfy all of 26

27 3 28 the official criteria, with one study suggesting an inappropriate blindness certification rate of 23%. Such 29 30 inappropriate certifications may inflate the numbers somewhat but it should be borne in mind that the aim 31 32 of certification is not to identify persons meeting rigid clinical criteria but to identify and count those with 33 34 significant visual impairment who may benefit from state assistance. Indeed, current guidelines for http://bmjopen.bmj.com/ 35 36 completion of CVI forms state that the criteria should be interpreted in the context of the patient’s 37 38 functional status rather than as strict cut-offs. Another precaution when interpreting these results is that 39 40 41 the figures for the two time periods were collated from different forms. The differences in these forms are

42 on September 27, 2021 by guest. Protected copyright. 7 43 discussed elsewhere and one of the main complications in comparing temporal trends is the increase in 44 45 the number of forms where a main cause has not been identified. In the 1999-2000 dataset, which was 46 47 derived from BD8, approximately 4% of forms had ‘multiple pathology’; in the 2009-2010 dataset derived 48 49 from CVI, this had increased to 14%. This raised the possibility that diabetic retinopathy/maculopathy may 50 51 have been under-reported for the 2009-2010 period. We attempted to address this by examining the 52 53 54 contributory causes in those without a main cause recorded and using these in place of the missing main 55 56 cause; this analysis resulted in only small changes to the percentage of blindness due to each cause and did 57 58 not change the overall ranking of the top six causes. This suggests that the rate of under-reporting of main 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 causes was similar for most categories and not responsible for the shift in the leading causes of blindness 3 4 certifications. 5 6 7 8 In summary, this report found three main causes were responsible for half of all blindness certifications 9 10 among working age adults in England and Wales from 2009-2010 - hereditary retinal disorders (20.2%), 11 12 followed by diabetic retinopathy/maculopathy (14.4%) and optic atrophy (14.1%). This marks the first time 13 14 15 in almost five decadesFor that diabetic peer retinopathy/maculopathy review is no longer only the leading cause of blindness in 16 17 working age adults, and may be related to the introduction of nationwide public health measures in 18 19 England and Wales. The results have implications for resource allocations for clinical care delivery and 20 21 research. 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Table 1. Numbers of working age adults (age 16-64) with severe sight impairment (blindness) in England 3 4 and Wales: certifications 2009-2010. The ‘Main cause’ column lists the number of certifications with the 5 6 corresponding diagnosis; the ‘Contributory cause’ column lists the contributory causes in certifications from 7 8 the ‘Multiple pathology’ and ‘No Information on main cause’ categories. 9 10 ICD-9 Codes Diagnosis Main cause Contributory Combined 11 12 (% Total) cause (% Total) 13 14 15 For peer review only(% Total) 16 17 18 19 362.7 Hereditary retinal disorders 354 (20.2) 29 (6.6) 383 (20.0) 20 21 362/34000 Diabetic retinopathy/maculopathy 253 (14.4) 56 (12.8) 309 (16.2) 22 23 24 377.1 Optic atrophy 248 (14.1) 46 (10.5) 294 (15.4) 25 26 365 Glaucoma 104 (5.9) 60 (13.7) 164 (8.6) 27 28 29 743-760 Congenital abnormalities of the eye 89 (5.1) 32 (7.3) 121 (6.3) 30 31 377.7 Disorders of visual cortex 72 (4.1) 24 (5.5) 96 (5.0) 32 33 430-438 Cerebrovascular disease 56 (3.2) 21 (4.8) 77 (4.0) 34 http://bmjopen.bmj.com/ 35 36 362.5 Degeneration of macula and posterior pole 52 (3.0) 14 (3.2) 66 (3.5) 37 38 360.2 Myopiap 49 (2.8) 23 (5.2) 72 (3.8) 39 40 370-371 Corneal disorders 45 (2.6) 34 (7.7) 79 (4.1) 41

1 2 Figure legends 3 4 Figure 1. Main causes of severe sight impairment (blindness) in England and Wales in working age adults 5 6 (age 16-64): certifications 2009-2010. 7 8 Figure 2. 10-year change in causes of severe sight impairment (blindness) in England and Wales in working 9 age adults (age 16-64): certifications 1999-2000 and 2009-2010. 10 11 12

13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Sources of Funding: 3 This study was supported by grants from the RNIB and the National Institute for Health Research 4 Biomedical Research Centre at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of 5 Ophthalmology. The data captured by CVI are DH copyright and this work was made possible by 6 collaboration with the Royal College of Ophthalmologists. The views expressed in this report are those of 7 the authors and do not necessarily reflect those of the funding bodies. MM is a recipient of a Foundation 8 Fighting Blindness Career Development Award. 9 10 Contributorship Statement: 11 All authors (GL, MM, CB) have contributed to 1) conception and design, acquisition of data, and analysis 12 and interpretation of data; 2) drafting the article and revising it critically for important intellectual content; 13 14 and 3) given final approval of the version to be published. 15 For peer review only 16 Financial Disclosure(s): 17 The authors have no proprietary or commercial interest in any materials discussed in this article. 18 19 Conflict of interest: 20 The authors have read and understood the BMJ Group policy on declaration of interests and declare no 21 conflict of interest. All authors have completed the Unified Competing Interest form at 22 www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare that 23 (1) No authors had support from any commercial companies for the submitted work; (2) GL received a 24 fellowship from the Royal Australia New Zealand College of Ophthalmologists (RANZCO) Eye 25 Foundation/Novartis scholarship for unrelated work during the period this report was written. No other 26 27 authors have relationships with any companies that might have an interest in the submitted work in the 28 previous 3 years; (3) No authors, their spouses, partners, or children have financial relationships that may 29 be relevant to the submitted work; and (4) No authors have no non-financial interests that may be relevant 30 to the submitted work. 31 32 Data sharing: 33 Technical appendix, statistical code, and dataset available from the corresponding author or at Dryad 34 repository, who will provide a permanent, citable and open access home for the datase. http://bmjopen.bmj.com/ 35 36 Licensing 37 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all 38 authors, a worldwide licence to the Publishers and its licensees in perpetuity, in all forms, formats and 39 40 media (whether known now or created in the future), to i) publish, reproduce, distribute, display and store 41 the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include within collections and create summaries, extracts and/or, abstracts of the Contribution, iii) create any other 42 on September 27, 2021 by guest. Protected copyright. 43 derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the 44 inclusion of electronic links from the Contribution to third party material where-ever it may be located; 45 and, vi) licence any third party to do any or all of the above. 46 47 48 49 50 51

52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Reference List 3 4 (1) Sorsby A. The causes of blindness in England 1948-50. 1953. London, HMSO. 5 6 (2) Sorsby A. The incidence and causes of blindness in England and Wales 1948-1962. 1-78. 1966. 7 London, HM Stationery Office. Reports on Public Health and Medical Subjects No. 114. 8 9 10 (3) Barry RJ, Murray PI. Unregistered visual impairment: is registration a failing system? Br J 11 Ophthalmol 2005; 89(8):995-998. 12 13 (4) Robinson R, Deutsch J, Jones HS, et al. Unrecognised and unregistered visual impairment. Br J 14 Ophthalmol 1994; 78(10):736-740. 15 For peer review only 16 (5) UK Department of Health. Healthy Lives, Healthy People: Improving Outcomes and Supporting 17 18 Transparency. 2012. 19 20 21 (6) Bunce C, Wormald R. Causes of blind certifications in England and Wales: April 1999-March 2000. 22 Eye (Lond) 2008; 22(7):905-911. 23 24 (7) Bunce C, Wormald R. Leading causes of certification for blindness and partial sight in England & 25 Wales. BMC Public Health 2006; 6:58. 26 27 (8) Bunce C, Xing W, Wormald R. Causes of blind and partial sight certifications in England and Wales: 28 April 2007-March 2008. Eye (Lond) 2010; 24(11):1692-1699. 29 30 (9) Sorsby A. The incidence and causes of blindness tn England and Wales 1963-1968. 1972. London, 31 HM Stationary Office. DHSS reports on public health and medical subjects No 128. 32 33 34 (10) Centers for Disease Control and Prevention. http://www cdc http://bmjopen.bmj.com/ 35 gov/visionhealth/basic_information/eye_disorders htm [ 2013 [cited 2013 July 20]; 36 37 (11) Lusignan S, Sismanidis C, Carey IM, et al. Trends in the prevalence and management of diagnosed 38 type 2 diabetes 1994-2001 in England and Wales. BMC Fam Pract 2005; 6(1):13. 39 40 41 (12) Khunti K, Goyder E, Baker R. Collation and comparison of multi-practice audit data: prevalence and treatment of known diabetes mellitus. Br J Gen Pract 1999; 49(442):375-379. 42 on September 27, 2021 by guest. Protected copyright. 43 44 (13) Holman N, Forouhi NG, Goyder E, et al. The Association of Public Health Observatories (APHO) 45 Diabetes Prevalence Model: estimates of total diabetes prevalence for England, 2010-2030. Diabet 46 Med 2011; 28(5):575-582. 47 48 (14) Forouhi NG, Merrick D, Goyder E, et al. Diabetes prevalence in England, 2001--estimates from an 49 epidemiological model. Diabet Med 2006; 23(2):189-197. 50 51 (15) NHS Diabetic Eye Screening Programme. http://diabeticeye screening nhs uk/about [ 2013 52 53 (16) Diabetic Retinopathy Screening Service Wales. http://www cardiffandvaleuhb wales nhs uk/drssw [ 54 2013 55 56 (17) Calvert M, Shankar A, McManus RJ, et al. Effect of the quality and outcomes framework on diabetes 57 care in the United Kingdom: retrospective cohort study. BMJ 2009; 338:b1870. 58 59 60 15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 (18) Campbell SM, Reeves D, Kontopantelis E, et al. Effects of pay for performance on the quality of 3 primary care in England. N Engl J Med 2009; 361(4):368-378. 4 5 (19) Millett C, Saxena S, Ng A, et al. Socio-economic status, ethnicity and diabetes management: an 6 analysis of time trends using the health survey for England. J Public Health (Oxf) 2007; 29(4):413- 7 419. 8 9 (20) Oluwatowoju I, Abu E, Wild SH, et al. Improvements in glycaemic control and cholesterol 10 concentrations associated with the Quality and Outcomes Framework: a regional 2-year audit of 11 diabetes care in the UK. Diabet Med 2010; 27(3):354-359. 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Changes in the Leading Causes of Blind Certifications 3 4 5 in Working Age Adults in England and Wales 2009/2010 versus 1999/2000 6 7 – With Diabetic Retinopathy Now No Longer the Leading Cause. 8 9 10 Short title: Blindness in England and Wales 2009/2010 11 12 Gerald Liew, Michel Michaelides*, Catey Bunce* 13 14 15 Moorfields Eye HospitalFor NHS Foundationpeer Trust, reviewLondon. EC1V 2PD, UK. only 16 Gerald Liew, Medical Retina Research Fellow; 17 Michel Michaelides, Consultant Ophthalmologist. 18 19 20 Centre for Vision Research, Westmead Millenium Institute, University of Sydney, Sydney, Australia. 21 Gerald Liew, Hononary Senior Lecturer 22 23 Department of Genetics, University College London, Institute of Ophthalmology, Bath Street, London EC1V 24 9EL, UK; 25 Michel Michaelides, Clinical Senior Lecturer 26 27 National Institute for Health Research Biomedical Research Centre at Moorfields Eye Hospital NHS 28 Foundation Trust and UCL Institute of Ophthalmology. 29 Catey Bunce, Senior Statistician 30

31 32 *Both authors contributed equally to the work and should be considered joint senior and corresponding 33 authors.

34 http://bmjopen.bmj.com/ 35 Financial Disclosure(s): 36 The authors have no proprietary or commercial interest in any materials discussed in this article. 37 38 Conflict of interest: 39 The authors have read and understood the BMJ Group policy on declaration of interests and declare no 40 conflict of interest. All authors have completed the Unified Competing Interest form at 41 www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare that 42 (1) No authors had support from any commercial companies for the submitted work; (2) GL received a on September 27, 2021 by guest. Protected copyright. 43 fellowship from the Royal Australia New Zealand College of Ophthalmologists (RANZCO) Eye 44 45 Foundation/Novartis scholarship for unrelated work during the period this report was written. No other 46 authors have relationships with any companies that might have an interest in the submitted work in the 47 previous 3 years; (3) No authors, their spouses, partners, or children have financial relationships that may 48 be relevant to the submitted work; and (4) No authors have no non-financial interests that may be relevant 49 to the submitted work. 50 51 Sources of Funding: 52 This study was supported by grants from the RNIB and the National Institute for Health Research 53 Biomedical Research Centre at Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of 54 Ophthalmology. The data captured by CVI are DH copyright and this work was made possible by 55 collaboration with the Royal College of Ophthalmologists. The views expressed in this report are those of 56 the authors and do not necessarily reflect those of the funding bodies. MM is a recipient of a Foundation 57 58 Fighting Blindness Career Development Award. 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Licensing 3 The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all 4 authors, a worldwide licence to the Publishers and its licensees in perpetuity, in all forms, formats and 5 media (whether known now or created in the future), to i) publish, reproduce, distribute, display and store 6 the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include 7 within collections and create summaries, extracts and/or, abstracts of the Contribution, iii) create any other 8 derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the 9 inclusion of electronic links from the Contribution to third party material where-ever it may be located; 10 and, vi) licence any third party to do any or all of the above. 11 12 13 14 15 Corresponding authors:For peer review only 16 C. Bunce, Research and Development, Moorfields Eye Hospital, City Road, London EC1V2PD, UK 17 Tel: +442075662820 18 Fax: +442086086925 19 Email: [email protected] 20 21 M. Michaelides, Department of Genetics, University College London, Institute of Ophthalmology, Bath 22 Street, London EC1V 9EL, UK 23 Tel: +442076086864 24 Email: [email protected] 25 26 Word count: 2367 27 28 29 Abstract: 233 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 Abstract 4 5 Objectives. To report on the causes of blindness certifications in England and Wales in working age adults 6 7 8 (16-64 years) in 2009-2010; and to compare these with figures from 1999-2000. 9 10 Design. Analysis of the national database of blindness certificates of vision impairment (CVIs) received by 11 12 the Certifications Office. 13 14 Setting and Participants. Working age (16-64 years) population of England and Wales. 15 For peer review only 16 Main outcome measures. Number and cause of blindness certifications. 17 18 Results. The Certifications Office received 1756 CVIs for blindness from persons aged between 16 and 64 19

20 st st 21 inclusive between 1 April 2009 to 31 March 2010. The main causes of blindness certifications were 22 23 hereditary retinal disorders (354 certifications comprising 20.2% of the total), diabetic 24 25 retinopathy/maculopathy (253 persons, 14.4%), and optic atrophy (248 persons, 14.1%). Together these 26 27 three leading causes accounted for almost 50% of all blindness certifications. Between 1st April 1999 to 31st 28 29 March 2000, the leading causes of blindness certification were diabetic retinopathy/maculopathy (17.7%), 30 31 hereditary retinal disorders (15.8%) and optic atrophy (10.1%). 32 33

34 Conclusions. For the first time in at least five decades, diabetic retinopathy/maculopathy is no longer the http://bmjopen.bmj.com/ 35 36 leading cause of certifiable blindness amongst working age adults in England and Wales, having been 37 38 overtaken by inherited retinal disorders. This change may be related to factors including the introduction of 39 40 nationwide diabetic retinopathy screening programmes in England and Wales, and improved glycaemic 41

42 control. Inherited retinal disease now representing the commonest cause of certification in the working age on September 27, 2021 by guest. Protected copyright. 43 44 population has both clinical and research implications, including with respect to the provision of 45 46 47 care/resources in the NHS and the allocation of research funding. 48 49 50 51 240 words 52 53 54 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 “Article Summary” 4 5 Article Focus. 6 7 - Blindness certifications are an important public health indicator in England and Wales. 8 9 - Every year, data on the number of persons certified blind is collected and the main causes listed. 10 11 - In 1999-2000, the leading causes of blindness certifications were diabetic retinopathy/maculopathy, 12 13 14 hereditary retinal disorders and optic atrophy. We report results for 2009-2010. 15 For peer review only 16 17 18 Key messages. 19 20 - We report that for the first time in over five decades, diabetic retinopathy/maculopathy is no longer the 21 22 leading cause of blindness in working age adults, having been overtaken by hereditary retinal disorders. 23 24 - This decline in blindness certifications from diabetic retinopathy/maculopathy may be related to the 25 26 27 introduction of nationwide public health measures in England and Wales. 28 29 - The results have implications for resource allocations for clinical care delivery and research. 30 31 32 33 Strengths and Limitations

34 http://bmjopen.bmj.com/ 35 - Strengths of the data include nationwide coverage, collection of uniform data fields with pre-specified tick 36 37 boxes, and clear and consistent definitions of sight impairment. 38 39 40 - Limitations include comparisons across two slightly different data collection forms. 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 4 Introduction 5 6 England and Wales have collected data on the number of blind people in the country since 1851.1;2 From 7 8 1930 until 2003, the causes of blindness have also been collected through use of a designated certificate, 9 10 the BD8, which was completed by the examining ophthalmologist.2 The BD8 was replaced in England in 11 12 September 2005 by the Certificate of Vision Impairment (CVI), and in Wales in April 2007 by the equivalent 13 14 15 form, the CVI-W.For One copy of thepeer CVI and CVI-W review is then sent to the Certifications only Office, London, for 16 17 anonymised epidemiological analysis. The Certifications Office is currently funded by the RNIB (Royal 18 19 National Institute for the Blind) and operates under the auspices of the Royal College of Ophthalmologists 20 21 with CVI data under Crown copyright, meaning the copyright is owned by the British Government. 22 23 24 25 Although not compulsory, certification allows patients to be registered (i.e. placed on an official local 26 27 28 council register) as either severely sight impaired (blind) or sight impaired (partially sighted), which then 29 30 permits access to certain state benefits and social service provisions. There is hence an incentive for 31 32 patients to be certified and counted, providing an estimate, albeit imperfect,3;4 of the causes and burden of 33 34 blindness in England and Wales. The importance of these data is highlighted by the fact that certification http://bmjopen.bmj.com/ 35 36 figures have recently been adopted in 2012 as an indicator for preventable sight loss and included in the 37 38 Public Health Outcomes Framework.5 In this report we present the findings from an analysis of working age 39

40 6 41 blindness certifications from 2009-2010 and compare these with figures from 1999-2000. Major changes in

42 on September 27, 2021 by guest. Protected copyright. 43 leading causes of blindness certification in this age group have occurred over this period; subsequent 44 45 publications will report on findings in persons of other age groups. 46 47 48 49 Methods 50 51 Details of data entry, collections and transmission have been reported previously.6-8 Between 1st April 2009 52

53 st st st 54 to 31 March 2010 data were obtained from CVI forms, whereas between 1 April 1999 to 31 March 2000, 55 56 data were obtained from BD8 forms. With regards to CVI forms, data were transcribed from these paper- 57 58 based forms into a database at the Certification Office. Part C of the CVI form collects information on the 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 causes of visual loss and requires the completing ophthalmologist to select from a list of common 3 4 diagnoses. More than one cause of visual loss can be selected in which case the main cause should be 5 6 highlighted using either an asterisk or circle. Guidelines are provided to assist with this process when the 7 8 main cause is not evident. ‘Multiple’ causes are used where the ophthalmologist completing the form has 9 10 not indicated a single main cause. Possible reasons for this selection include differing causes in the two 11 12 eyes, or more than one cause within one eye and the ophthalmologist is unable to determine which 13 14 15 contributes the mostFor to the visual peer loss. Causes forreview visual impairment were only coded using International 16 6-8 17 Classification of Diseases (ICD)-9 codes and grouped into disease categories as in previous reports. 18 19 20 21 For 1999-2000, data were extracted from paper based BD8 forms in a similar fashion.6 Part 5 of the BD8 22 23 form has 16 fields for medical information including one for the main cause of visual loss and five for each 24 25 eye for contributory causes. For records where the main cause was not filled, this field was imputed 26

27 8 28 wherever possible using an algorithm published previously. In short this algorithm brought forward as the 29 30 main cause the contributing cause if any were so listed; if more than one contributory cause was listed, the 31 32 patient was listed as having ‘multiple causes’. The main difference between data collected from the CVI and 33 34 BD8 forms is the number of certifications with ‘multiple causes’ of visual impairment which was higher in http://bmjopen.bmj.com/ 35 36 the CVI forms.7 We have previously minimized bias from this source by re-examining the certificates with 37 38 ‘multiple causes’, extracting the contributory causes listed and tabulating them with the main causes. 7 We 39 40 41 follow the same procedure in this report.

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 Blindness was defined according to criteria previously described on the BD8 and CVI forms as either best 46 47 corrected visual acuity in the better eye of either (1) worse than 3/60, or (2) worse than 6/60 with severely 48 49 constricted visual fields, or (3) better than 6/60 with severely constricted visual fields particularly the 50 51 inferior field. 52 53 54 55 56 Analyses 57 58 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Data regarding the main cause of visual impairment were extracted from the forms and grouped into 3 4 disease categories as previously described.6-8 For CVI data, those with ‘multiple causes’ or ‘no information 5 6 on main cause’ of visual impairment had contributory causes extracted and listed, and later combined with 7 8 the main causes. Pie and bar charts were used to graphically illustrate the distribution of the main causes of 9 10 visual impairment. Proportions of blindness registrations due to each cause are presented rather than 11 12 adjusted incidence rates in order to indicate the relative contribution of each condition to the pool of vision 13 14 15 impairment. CurrentlyFor almost everypeer grant application review for diabetic retinopathy only projects commences with a 16 17 statement that diabetic retinopathy is the most common cause of visual loss in the working age population 18 19 – we present proportions of blindness registrations to determine if this statement remains valid. Χ2 tests 20 21 were performed to test differences in proportions. 22 23 24 25 Results 26

27 st st 28 For the period 1 April 2009 to 31 March 2010, the Certifications Office received 1756 CVIs for blindness 29 30 from persons aged between 16 and 64 inclusive. This compares with 1637 BD8 forms for blindness received 31 32 between 1st April 1999 and 31st March 2000, details of which have been reported previously.8 Table 1 33 34 shows the number of persons certified blind for each of the disease categories. Hereditary retinal http://bmjopen.bmj.com/ 35 36 disorders, including Stargardt Disease and Retinitis Pigmentosa, formed the largest category with 354 37 38 certifications comprising 20.2% of the total. Diabetic retinopathy/maculopathy was the second largest 39 40 41 cause of certifiable blindness with 253 persons (14.4%), followed by optic atrophy with 248 persons

42 on September 27, 2021 by guest. Protected copyright. 43 (14.1%). Together these 3 leading causes accounted for almost 50% of all blindness certifications in the 44 45 working age group. Glaucoma was responsible for 104 (5.9%) blindness certifications, followed by 46 47 congenital abnormalities of the eye which included congenital cataracts and retinopathy of prematurity (89 48 49 certifications, 5.1%). Multiple pathologies were listed for 242 persons (13.8%) and no information on the 50 51 main cause was listed for 42 persons (2.4%). When these categories were examined for contributory 52 53 54 causes, the most common contributory causes recorded were glaucoma (60 persons), diabetic 55 56 retinopathy/maculopathy (56 persons) and optic atrophy (46 persons). Combining the main and 57 58 contributory causes resulted in small changes in the overall proportion of certifications due to specific 59 60 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 causes, but did not change the relative rankings of the top six causes of blindness (Table 1). ‘Other 3 4 conditions’ comprised 150 certifications (8.5%) of which the most common were malignant neoplasms of 5 6 the brain and nervous system (27 persons, 1.5%) and retinal detachments (24 persons, 1.4%). Figure 1 7 8 shows the distribution of the causes of blindness certifications graphically in a pie chart. 9 10 11 12 We next compared the main causes of blindness certifications from 1999-2000 (n=1637) with the figures 13 14 15 above from 2009-2010For (n=1756). peer The results are review shown graphically in Figureonly 2. From 1999-2000, the 16 17 leading cause of blindness certification was diabetic retinopathy/maculopathy which accounted for 290 18 19 certifications (17.7%). By 2009-2010, this figure had decreased to 253 (14.4%), and diabetic 20 21 retinopathy/maculopathy was now the second leading cause of blindness certification. This difference was 22 23 statistically significant (P=0.009). In contrast, hereditary retinal disorders which were the second leading 24 25 cause of blindness certification in 1999-2000 accounting for 258 cases (15.8% of total), had increased to 26 27 28 354 cases (20.2%) by 2009-2010 and have now become the leading cause of certifiable blindness in the 29 30 working age group in England and Wales. Optic atrophy remained the third leading cause in 1999-2000 and 31 32 2009-2010 with an increase from 165 cases (10.1%) to 248 cases (14.1%) respectively. A notable finding 33 34 was that degeneration of the macula and posterior pole, which accounted for 7.7% of blindness registration http://bmjopen.bmj.com/ 35 36 in 1999-2000, had dropped in percentage terms and now accounted for only 3.0% by 2009-2010. Other 37 38 causes of blindness registration remained roughly similar for the two time periods. 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 Discussion 44 45 This report provides updated estimates on the causes of certifiable blindness in England and Wales in 46 47 working age adults. Three main diseases were responsible for half of all certifications – hereditary retinal 48 49 disorders (20.2%), diabetic retinopathy/maculopathy (14.4%) and optic atrophy (14.1%). A marked change 50 51 has occurred in the relative importance of these main causes of blindness certifications since the last major 52

53 6 54 analysis in 1999-2000, with diabetic retinopathy/maculopathy now no longer the leading cause of 55 9 56 blindness in working age adults. Since at least 1963, diabetic retinopathy/maculopathy has been the 57 58 leading cause of blindness among working age adults in England and Wales; a similar situation exists in 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 10 2 other developed countries such as the United States. Over the past decade the proportion of 3 4 certifications for hereditary retinal disorders has slowly risen,8 while that for diabetic 5 6 retinopathy/maculopathy has reduced, resulting in the two conditions swapping rankings. 7 8 9 10 This report is not designed to identify the reasons behind these changes or estimate the incidence of 11 12 blindness from diabetes. The prevalence of diabetes in the UK is not known with certainty, with several 13

14 11-14 15 limited general practiceFor surveys peer suggesting a range review of between 1-2% of onlythe general population ; we are 16 17 thus not able to estimate the incidence of blindness from diabetes. Nonetheless, available data suggest the 18 19 prevalence of diabetes in England and Wales has increased over the period in question11;13, which would be 20 21 expected to lead to increased rates of blindness if other factors remained constant. In this context we 22 23 speculate that several intervening public health developments may have contributed to the reduction in 24 25 both absolute and proportional rates of registrable blindness from diabetes amongst working age adults. 26 27 28 Between 2003-2008 both England and Wales introduced nationwide diabetic retinopathy screening 29 30 services with the aim of reducing the incidence of blindness from diabetic eye disease. These are known as 31 32 the NHS Diabetic Eye Screening Program (England)15 and Diabetic Retinopathy Screening Service Wales 33 34 (DRSSW)16, and these programmes annually screen almost 2 million and 150,000 patients with diabetes, http://bmjopen.bmj.com/ 35 36 respectively. Concurrent with these screening programmes, in 2004 the Quality and Outcomes 37 38 Framework17 was introduced to incentivise general practitioners in the United Kingdom to improve primary 39 40 41 care management of several conditions including diabetes. Several studies have documented an

42 on September 27, 2021 by guest. Protected copyright. 18;19 43 improvement in the quality of care for diabetes since this was introduced, and the effort may have 44 45 contributed to the improvement in glycaemic control documented since the late 1990s.17;20 The decline in 46 47 both the absolute number and relative proportion of blindness certifications due to diabetic 48 49 retinopathy/maculopathy amongst working age adults since introduction of these public health measures 50 51 may be an indicator of their effectiveness. Nonetheless this remains speculative at present, and such 52 53 54 explanations should be read with caution. 55 56 57 58 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 Whether increased numbers of certification for inherited eye diseases reflects improved certification of 3 4 existing sight impairment, or a true increase in incidence of these disorders is unclear. The progress made 5 6 over the last decade in molecular genetics/diagnostics and the increasing avenues of research/clinical trials 7 8 for inherited retinal disease with widespread media coverage may plausibly have resulted in higher 9 10 community awareness and increased clinic visits and thereby registration,4 without a true increase in 11 12 incidence rates. An observation in favour of this scenario is that the numbers of blindness certification for 13 14 15 optic atrophy haveFor also increased peer over the last decade,review in tandem with thoseonly for inherited retinal disease, 16 17 while those for other non-inherited conditions such as glaucoma have remained fairly constant. Hereditary 18 19 retinal diseases occur more frequently in communities with a higher rate of consanguinity, and it is 20 21 conceivable that increased rates of immigration from countries where consanguinity is more prevalent may 22 23 have contributed to these findings, though at this stage this remains speculative. Another possibility is that 24 25 diagnostic transfer or misclassification may have occurred, for example where some cases of hereditary 26 27 28 retinal disorders may have been mislabeled as 'degeneration of macula and posterior pole'. In order to 29 30 explain the increase in hereditary retinal disorders, this would have had to occur preferentially in 1999- 31 32 2000 versus 2009-2010. However, misclassification of diabetic retinopathy/maculopathy as hereditary 33 34 retinal disorders is unlikely to occur given how different the conditions are, and so would not explain the http://bmjopen.bmj.com/ 35 36 absolute reduction in the number of certifications for blindness due to diabetes. 37 38

39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 These findings have implications for clinical care and research budget allocation. A prolonged focus on 44 45 prevention and treatment of diabetic eye disease has likely contributed to the decline in blindness 46 47 certifications from this disorder amongst working age adults, and the rate is expected to decline further 48 49 with the recent National Institute for Clinical Excellence (NICE) approval of ranibizumab for treatment of 50 51 diabetic maculopathy. Now that hereditary retinal diseases comprise the leading cause of blindness 52 53 54 certifications in working age adults, an increased focus on clinical management of these conditions (e.g. 55 56 with low vision aids, visual rehabilitation) and greater allocation of research funding to study these 57 58 disorders may be appropriate. Funding bodies may need to re-asses their funding priorities. 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 4 Strengths and Limitations 5 6 Data from BD8 and CVI registrations represent some of the best available epidemiological data on sight 7 8 impairment in England and Wales and is regarded as a major public health indicator.5 Strengths of the data 9 10 include nationwide coverage, collection of uniform data fields with pre-specified tick boxes, and clear and 11 12 consistent definitions of sight impairment. 13 14 15 For peer review only 16 17 Limitations that should be highlighted include that fact that blindness certifications are not equivalent to 18 19 blindness rates. This has been discussed previously3;4 and it is estimated that up to 53% of eligible patients 20 21 may not be registered blind despite consultation with an ophthalmologist.7 However, arguments have been 22 23 advanced that in time, most patients eligible to be registered will in fact do so,7 and studies have shown a 24 25 major increase in registration rates with increasing clinic visits.4 26 27 28 29 30 Another caveat to these results is that some patients who are certified blind may not always satisfy all of 31 32 the official criteria, with one study suggesting an inappropriate blindness certification rate of 23%.3 Such 33 34 inappropriate certifications may inflate the numbers somewhat but it should be borne in mind that the aim http://bmjopen.bmj.com/ 35 36 of certification is not to identify persons meeting rigid clinical criteria but to identify and count those with 37 38 significant visual impairment who may benefit from state assistance. Indeed, current guidelines for 39 40 41 completion of CVI forms state that the criteria should be interpreted in the context of the patient’s

42 on September 27, 2021 by guest. Protected copyright. 43 functional status rather than as strict cut-offs. Another precaution when interpreting these results is that 44 45 the figures for the two time periods were collated from different forms. The differences in these forms are 46 47 discussed elsewhere7 and one of the main complications in comparing temporal trends is the increase in 48 49 the number of forms where a main cause has not been identified. In the 1999-2000 dataset, which was 50 51 derived from BD8, approximately 4% of forms had ‘multiple pathology’; in the 2009-2010 dataset derived 52 53 54 from CVI, this had increased to 14%. This raised the possibility that diabetic retinopathy/maculopathy may 55 56 have been under-reported for the 2009-2010 period. We attempted to address this by examining the 57 58 contributory causes in those without a main cause recorded and using these in place of the missing main 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 cause; this analysis resulted in only small changes to the percentage of blindness due to each cause and did 3 4 not change the overall ranking of the top six causes. This suggests that the rate of under-reporting of main 5 6 causes was similar for most categories and not responsible for the shift in the leading causes of blindness 7 8 certifications. 9 10 11 12 In summary, this report found three main causes were responsible for half of all blindness certifications 13 14 15 among working ageFor adults in Englandpeer and Wales review from 2009-2010 - hereditary only retinal disorders (20.2%), 16 17 followed by diabetic retinopathy/maculopathy (14.4%) and optic atrophy (14.1%). This marks the first time 18 19 in almost five decades that diabetic retinopathy/maculopathy is no longer the leading cause of blindness in 20 21 working age adults, and may be related to the introduction of nationwide public health measures in 22 23 England and Wales. The results have implications for resource allocations for clinical care delivery and 24 25 research. 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 Figure 1. Main causes of severe sight impairment (blindness) in England and Wales in working age adults 4 5 (age 16-64): certifications 2009-2010. 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 31 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 Figure 2. 10-year change in causes of severe sight impairment (blindness) in England and Wales in working 4 5 age adults (age 16-64): certifications 1999-2000 and 2009-2010. 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 Reference List 4 5 (1) Sorsby A. The causes of blindness in England 1948-50. 1953. London, HMSO. 6 7 (2) Sorsby A. The incidence and causes of blindness in England and Wales 1948-1962. 1-78. 1966. 8 London, HM Stationery Office. Reports on Public Health and Medical Subjects No. 114. 9 10 11 (3) Barry RJ, Murray PI. Unregistered visual impairment: is registration a failing system? Br J 12 Ophthalmol 2005; 89(8):995-998. 13 14 (4) Robinson R, Deutsch J, Jones HS, Youngson-Reilly S, Hamlin DM, Dhurjon L et al. Unrecognised and 15 unregisteredFor visual impairment. peer Br J Ophthalmol review 1994; 78(10):736-740. only 16 17 18 (5) UK Department of Health. Healthy Lives, Healthy People: Improving Outcomes and Supporting 19 Transparency. 2012. 20 21 22 (6) Bunce C, Wormald R. Causes of blind certifications in England and Wales: April 1999-March 2000. 23 Eye (Lond) 2008; 22(7):905-911. 24 25 (7) Bunce C, Wormald R. Leading causes of certification for blindness and partial sight in England & 26 Wales. BMC Public Health 2006; 6:58. 27 28 (8) Bunce C, Xing W, Wormald R. Causes of blind and partial sight certifications in England and Wales: 29 April 2007-March 2008. Eye (Lond) 2010; 24(11):1692-1699. 30 31 (9) Sorsby A. The incidence and causes of blindness tn England and Wales 1963-1968. 1972. London, 32 HM Stationary Office. DHSS reports on public health and medical subjects No 128. 33 34 http://bmjopen.bmj.com/ 35 (10) Centers for Disease Control and Prevention. http://www cdc 36 gov/visionhealth/basic_information/eye_disorders htm [ 2013 [cited 2013 July 20]; 37 38 (11) Lusignan S, Sismanidis C, Carey IM, DeWilde S, Richards N, Cook DG. Trends in the prevalence and 39 40 management of diagnosed type 2 diabetes 1994-2001 in England and Wales. BMC Fam Pract 2005; 41 6(1):13.

42 on September 27, 2021 by guest. Protected copyright. 43 (12) Khunti K, Goyder E, Baker R. Collation and comparison of multi-practice audit data: prevalence and 44 treatment of known diabetes mellitus. Br J Gen Pract 1999; 49(442):375-379. 45 46 (13) Holman N, Forouhi NG, Goyder E, Wild SH. The Association of Public Health Observatories (APHO) 47 Diabetes Prevalence Model: estimates of total diabetes prevalence for England, 2010-2030. Diabet 48 Med 2011; 28(5):575-582. 49 50 (14) Forouhi NG, Merrick D, Goyder E, Ferguson BA, Abbas J, Lachowycz K et al. Diabetes prevalence in 51 England, 2001--estimates from an epidemiological model. Diabet Med 2006; 23(2):189-197. 52 53 (15) NHS Diabetic Eye Screening Programme. http://diabeticeye screening nhs uk/about [ 2013 54 55 (16) Diabetic Retinopathy Screening Service Wales. http://www cardiffandvaleuhb wales nhs uk/drssw [ 56 2013 57 58 59 60 16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 33 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 (17) Calvert M, Shankar A, McManus RJ, Lester H, Freemantle N. Effect of the quality and outcomes 3 framework on diabetes care in the United Kingdom: retrospective cohort study. BMJ 2009; 4 338:b1870. 5 6 (18) Campbell SM, Reeves D, Kontopantelis E, Sibbald B, Roland M. Effects of pay for performance on 7 the quality of primary care in England. N Engl J Med 2009; 361(4):368-378. 8 9 (19) Millett C, Saxena S, Ng A, Mainous A, III, Majeed A. Socio-economic status, ethnicity and diabetes 10 management: an analysis of time trends using the health survey for England. J Public Health (Oxf) 11 2007; 29(4):413-419. 12 13 (20) Oluwatowoju I, Abu E, Wild SH, Byrne CD. Improvements in glycaemic control and cholesterol 14 concentrations associated with the Quality and Outcomes Framework: a regional 2-year audit of 15 diabetes Forcare in the UK.peer Diabet Med 2010; review 27(3):354-359. only 16 17

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 90x109mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 35 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 http://bmjopen.bmj.com/ 35 36 106x90mm (300 x 300 DPI) 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 36 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

32 (b) Cohort study—For matched studies, give matching criteria and number of 33 exposed and unexposed

42 on September 27, 2021 by guest. Protected copyright. 43 Bias 9 Describe any efforts to address potential sources of bias 44 Y 45 Study size 10 Explain how the study size was arrived at 46 NA 47 48 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, 49 NA describe which groupings were chosen and why 50 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding 51 Y (b) Describe any methods used to examine subgroups and interactions 52 53 (c) Explain how missing data were addressed 54 (d) Cohort study—If applicable, explain how loss to follow-up was addressed 55 Case-control study—If applicable, explain how matching of cases and controls was 56 addressed 57 Cross-sectional study—If applicable, describe analytical methods taking account of 58 59 sampling strategy 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml1 Page 37 of 38 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from

1 2 (e) Describe any sensitivity analyses 3 Continued on next page 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 27, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml2 BMJ Open Page 38 of 38 BMJ Open: first published as 10.1136/bmjopen-2013-004015 on 12 February 2014. Downloaded from