Conferences - BARS Conferences

Report on the BARS (British Association of Retinal Screeners) Conference 2009 - Newcastle

The 9th Annual BARS conference took place in Newcastle on 1st and 2nd October. It was the largest yet with over 300 delegates. There was a strong presence from the NSPDR (National Screening Programme for Diabetic Retinopathy) following a significant number of programmes having undergone or anticipating their first External Quality Assurance (EQA) visit There was also a look to the future of screening with the use of OCT (optical coherence tomography) to help define the presence of referable diabetic maculopathy, automated analysis of retinal images being introduced into the Scottish programme and the development of better medical strategies for the management of diabetes and treatment of diabetic retinopathy.

Dr Rod Harvey, clinical lead for the Scottish programme gave the opening presentation on 'Diabetic Retinal Screening in Scotland'. There are significant differences between the Scottish and English programmes. The population of Scotland is 5,144,200 mostly in the central belt (Glasgow and Edinburgh) so there are significant issues of remoteness and equitable access for the whole population. There are 14 individual health boards, a single national screening software, flexible delivery, policies determined by a DRS collaborative and sharing of a national database of eligible patients which tracks patients as they move around. If a GP registers a patient as diabetic they automatically go onto the database and are registered for screening. As someone who's screening programme has had significant problems accessing this data from local GPs I can recognise the huge advantages of this database to the screening programmes. Another major difference from the English scheme is that they take a single field macular centred image and mydriasis is used only if the image quality without dilation is inadequate. The rationale for these protocols is that there is very little difference in sensitivity if patients are dilated where necessary. It is also argued that uptake is greater without compulsory mydriasis as patients generally do not like the eyedrops. Grading is entirely ‘feature based' so the grader reports the presence or absence of the various features of retinopathy and the software calculates the final grade dependant upon the features reported and the outcome is determined automatically. There are 3 levels of graders. Level 1 graders only final grade the ‘normals', level 2 graders only grade ‘non-referable retinopathy' and level 3 (usually ophthalmologists) graders grade all the referable retinopathy. They have 3 M grades – M0 is no maculopathy, M1 is observable (but not referable) and M2 is referable maculopathy. M1 is defined as exudates between 1 to 2 Disc Diameters (DD) from the fovea and they are reviewed at a 6 month interval, and M2 is exudates within 1 DD of the fovea and is referable. This is a pragmatic approach and does provide more flexibility in management of these patients than the English programme. For Quality Assurance of the grading, 500 image pairs from Levels 1 and 2 are randomly regraded by Level 3 graders. External QA is currently under development. At present the invitation rate is 93.2% (target 100%) and the attendance rate is 77% (target 80%). The Scottish programme is looking to replace Level 1 grading with automated analysis using the Medylitix software as a means of making significant financial savings for the programme.

Catherine Egan from Moorfields Eye Hospital gave a presentation entitled 'Needles in a Haystack – highlights in a screening day' in which she looked a few case studies of the more unusual conditions that present themselves in a screening programme. Case 1 was Macular Telangetasia. Despite often having good to reasonable VA these patients are functionally quite visually impaired due to a small field loss adjacent to the fovea. In appearance there is a paramacular grey region and research has shown that there is no macula pigment at the fovea as this has been redistributed to about 5° away from the fovea. There are often inner retinal crystals which look like exudates. It can be mistaken for wet ARMD in the late stages. It does not respond to macular laser. Case 2 was MIDD (Maternally Inherited Diabetes & Deafness). These cases often show bilaterally symmetrical paramacular atrophy with preservation of the central macular region. It is strongly associated with neurosensory hearing loss and nephropathy. There is a family history passed down the maternal line (via the mitochondrial DNA). It may also show as a pattern dystrophy in the paramacular region. It is important to refer these patients because of the transmission of the condition. Case 3 demonstrated the need for careful grading inpeople with type 1 diabetes. In some cases there may be minimal microaneurysms and dot and blot haemorrhages but careful examination of the horizontal meridians temporal to the macula and nasal to the disc may reveal areas of 'creeping IRMAs'. These patients often have profound macular ischaemia (which does not respond to macular laser or ante VEGF treatments) and may present to the programme with a vitreous haemorrhage if this is not detected earlier.

Richard Dewhirst gave a short presentation on the 'BARS Forum' to encourage BARS members and anyone associated with retinal screening to use the facility. He pointed out that the more people who use it, the greater the benefits and that it has the potential to be the one area where anyone participating in a screening programme gets to share their experience and influence the development of screening in the UK. The BARS council will monitor the views expressed and it could help develop council policy.

Fionna O'Leary, the lead of the ENSPDR, gave a presentation on 'EQA – Where are we now'. 48 out of the 92 Programmes have had full External Quality Assurance visits. Greater understanding of systematic screening and clinical governance has led to an increase in reported Serious Incidents (SIs) and greater understanding of risk has led to several programmes pausing their screening programmes in order to regroup and start again more securely. She made the point that reporting serious incidents is not necessarily all bad in as much as it shows good clinical governance and can lead to significant improvements.  Nearly all SIs were rooted in incomplete commissioning of the pathway, inadequate funding, inadequate staffing levels and the lack of an effective high level public health led Programme Board to oversee performance against national standards which would enable early warning and correction. Examples included grading backlogs of up to 2 years in the desperate rush to meet screening targets, large gaps in the cohort identification, rushed grading leading to missed sight threatening pathology and patients lost to ophthalmology resulting in sight impairment. There were also IT security issues, loss of data, poorly maintained equipment, incorrectly managed exclusions, and lack of failsafe strategies compounded by over configurable software or inappropriate categories and automatic movement of patients into various states leading to patients not being recalled to screening. Standardisation of categories within the software is essential to prevent inadvertent errors and software suppliers have been asked to remodel software to prevent error, and increase ability to flag categories of patients as active but requiring special attention. Serious Incidents are leading to a marked increase in investment and improved quality of working and early identification and correction of error.  Incident reporting is a good thing! The NSPDR has responded by developing EARS - a new Annual reporting system which will have up to 40 management reports available to programmes and will provide the ability to lodge annual reports monthly in order to monitor progress more frequently. A GP2DRS pilot has been very successful and demonstrates clearly that even with thorough manual updating to the database hundreds if not thousands of patients are likely to be found. National procurement of this system is now underway. There is also an on-line EQA grading test set which can provide detailed reporting on over and under-grading for each grader. Repeated use of the test sets, backed up by feedback and additional training in MDTs, has been shown to produce marked improvement in grading accuracy. Finally Fionna listed the personal qualities such as courage, dedication and enthusiasm that she has seen in abundance in the various programmes and said she was proud to be associated with a world leading screening programme. There followed short presentations from a few programmes who had declared SIs following their EQA visits demonstrating the positive outcomes.

'Tales from the Front Line'.  Tony Nicholls from the Norfolk & Norwich University Hospital DRSS described the process of the EQA visit from the pre-visit questionnaires and audits through to the final report. The positives to come out of EQA were an increased profile with the commissioning PCT which identified them as an under resourced programme, the recommendation to appoint a Failsafe Officer, Audits that were useful in identifying delays in ophthalmology, confirmed retinopathy prevalence to be in line with National expectations and an IT solution for outcome reporting from ophthalmology. The negatives were the huge investment of time spent on preparation, questionnaires and audits in an already busy schedule, possibly too much emphasis on relatively minor issues and failure to recognise recent improvements and a delay in receiving the final report although an interim report was forthcoming.
Paul Robey, the acting programme manager for Wrightington, Wigan and Leigh DRSS spoke about the serious failings of this programme including serious underfunding, lack of continuity in clinical leadership and general management, absent failsafe procedures and lack of dedicated space and how these had resulted in the EQA panel suspending the programme because of its findings which included serious under grading at both grader and ophthalmologist levels (new vessels were missed in 4 out of 21 cases). Peer reviewers disagreed with almost 50% of referable DR reviewed! Inadequate software and rushed grading was highlighted. The programme was paused to allow retraining of the grading staff, there was a look back exercise to risk assess those who needed to be recalled as a priority and there were weekly SUI (serious untoward incident) meetings at executive level. This resulted in funding being effectively doubled with further capital investment of about £100K for new equipment and software. The programme re-launched on  1st June 2009 with quarterly high level programme management board meetings, active consultant clinical leadership, protected time for weekly MDT's and CPD, EQA on line grading tests , City & Guilds training in place and new protocols and processes. This programme recognised that none of this would have been achieved without the interventions following the EQA visit.

Waltham Forest DRSS described their experience. The EQA visit highlighted the inadequate office space and lack of a separate grading room and 4 SIs were declared relating to the screening to laser interval. There was incomplete commissioning of the service initially so that there was a backlog of over 600 patients waiting for slit lamp examination and grading. The EQA interim report has resulted in moves to increase the size and suitability of the space for screening, commissioning of a slit lamp service with Whipps Cross Hospital Ophthalmology, a monthly validation process with ophthalmology and formal adoption of policies and protocols.

Wendy Newton, lead optometrist for Cambridgeshire DRSS described a CVI (Certificate of Visual Impairment) audit that was carried out prior to their EQA visit. The goal of the NSPDR is to reduce the risk of sight loss amongst people with diabetes. The objectives of the audit were to obtain the rates of visual impairment registration due to Diabetic Retinopathy in the population covered by the programme, to compare this rate to previous audit in 2004 and to compare these with National Standards. Standard 1 of the Service Objectives and Quality Assurance Standards: Release 6.5, June 2009 quotes National rates in 1990/91 of SSi  9.5 per million per annum and Si  9.3 per million per annum ( where SSi is VA of 6/60 or worse in the better eye and Si is VA of 6/18 or worse in the better eye ).  The minimum standard is a 10% reduction within 5 years of start of screening programme and the achievable standard is quoted as a 40% reduction within 5 years of start of screening programme. Other data for comparison purposes was found in Bunce C, Wormald R. Leading causes of certification for blindness and partial sight in England & Wales. BMC Public Health. 2006 Mar 8; 6:58. Which reports SSi of 13.6 per million per year in 1999/2000 and Si of 24.8 per million per year in 1999/2000.  The results showed that of the 315 CVi forms recorded in the 12 months of the audit, 21 were predominantly due to diabetic retinopathy (SSi 7 cases and Si 14 cases) This translates to rates of SSi of 12.6 per million per year and Si of 25.2 per million per year. This is an increase on the Rates from 1990/91 but it is thought that this may be due to increased reporting and also these are small numbers so a minor annual variation has a major effect. The rates were comparable to those from 1999/2000. Recommendations to come out from this audit showed that a good CVi audit is a large undertaking, a 5 year gap between audits would appear to be a good interval as numbers are very small, the CVi Database should be held at the eye department, it is important to record postcode of both patient and GP to ensure correct boundaries for the DRS Scheme and it might in future be possible to use the national CVi database, held at Moorfields, for verification of local and national figures.

The second session of the conference began with a presentation by Steve Aldington, Education & Research Development Manager for the NSPDR on the 'DRS grading criteria'. He began with a history of the various grading schemes for diabetic retinopathy which began with the Hammersmith Scheme in 1966, through the Airlie House scheme in 1968 which built on the work from the Diabetic Retinopathy Study (DRS) (1981), the Diabetic Vitrectomy Study (1985) and the Early Treatment DR Study (ETDRS) (1987/91). In 1985 there was the Eurodiab scheme followed in 2003 by the International scheme and the NSPDR grading. The last 3 schemes group lesions by severity and risk of progression.
The evidence from the EQA visits and the annual reports shows excellent on-going programme development with highly motivated and skilled staff and increasingly high coverage. However there are a range of non-standard R and M grades, a wide variety of interpretations of the grading classifications and a wide variety of data forms.
He clarified the grading criteria as required by the NSPDR

Retinopathy – The R level

    1. R0 – no diabetic retinopathy detected
    2. R1 – presence of any DR but not R2 or R3
    3. R2 – presence of IRMA, Venous Beading, loops, multiple deep round haemorrhages (±Cotton Wool Spots)
    4. R3 – presence of proliferative or advanced DR

Every record must map to one of these 4 grades OR be regarded as unclassifiable, unobtainable or ungradable.

Maculopathy – The M level

    1. M0 – no referable maculopathy detected
    2. M1 – presence of referable maculopathy:
      1. Exudate within 1DD of centre of fovea
      2. Circinate or group of exudate within the macula (2DD from centre of fovea)
      3. Thickening within 1DD of centre of fovea
      4. MA or haemorrhage within 1DD of centre of fovea ONLY if associated with a best VA of < 6/12

Every record must map to one of these 2 grades

Photocoagulation – The P level

    1. P0 – no scars of photocoagulation detected
    2. P1 – presence of photocoagulation scars:
      1. Evidence of focal/grid laser to macula
      2. Evidence of peripheral scatter laser

Every record must map to one of these 2 grades

Note: R0 M1 is not a valid grade as maculopathy is inherently diabetic retinopathy.

He also gave examples of the various ‘inconsistencies' and made the point that these were NOT of the minimum standard dataset. There may be very good reasons for these inconsistencies such as local referral protocols, a legacy from previous systems, customised software and imprecise definitions. Some of the ‘definitions' open to interpretation included the macula, ‘within 1 DD of the centre of the fovea', ‘multiple deep, round or blot haemorrhages', ‘circinate or group of exudates', IRMA and venous loop.

There is currently a research project looking to define ‘multiple deep, round or blot haemorrhages', ‘circinate or group of exudates', IRMA using an expert reference group of ophthalmologists from 6 UK centres. This has revealed the degree of divergence present in defining these terms. The point was made that the use of non-standard data makes inter programme comparisons impossible, retinopathy incidence and prevalence figures inaccurate, reduces the overall sample size of the data groups and lessens the potential for a national evidence base. However many of these issues will not be resolved until precise definitions are available.

Mr Martin Harris, consultant ophthalmologist for the Barnet scheme gave a talk on OCT in DRS – A Look to the Future. Optical Coherence Tomography is effectively optical ultrasound. Its advantages include live subsurface images at high resolution, direct imaging of tissue morphology, minimal preparation of the subject and no ionising radiation. It allows cross sectional visualisation, quantitative measurements, a high axial resolution of < 10mm and good reproducibility.
The ETDRS classification of maculopathy is: Retinal thickening at or within 500mm of the centre of the macula and/or hard exudates at or within 500mm of the centre of the macula, if associated with thickening of the adjacent retina and/or a zone or zones of retinal thickening 1 disc area in size at least part of which is within 1 disc area of the centre of the macula. Digital photography cannot show retinal thickening so various surrogate markers are used in the NSPDR classification. Barnet DRSS have carried out an M1 study to see if OCT can refine referable maculopathy. A consecutive series of diabetics screened as having M1 then had 3D OCT scans on both eyes. This measured the central retinal thickness and recorded site specific retinal thickness and observed the type of thickening and the vitreo-retinal interface. A comparison was then made on the clinical decision based on screening data from both the OCT and the single colour fundus photograph. The results of this small study ( 62 patients – 123 eyes) showed that only 29% of patients screened as M1 had OCT morphology changes. OCT clearly demonstrate the disparity between M1 at screening and clinically significant macular Oedema. In conclusion he stated that a definition of clinically significant maculopathy needs to be redefined in terms of OCT imaging and as the cost of FD-OCT falls, screening for maculopathy by OCT will become feasible.

Sue Cobbold from the Suffolk DRS spoke about 'Screening in Prisons'. The Prison Service Instruction, Feb 2002 states that prisoners should have access to the same range and quality of services that the general public receives from the National Health Service. PCTs are responsible for the delivery of this care so retinal screening is a compulsory activity. It can be very challenging to provide this service. Prisoners are amongst high risk groups for retinopathy, referable retinopathy is more prevalent, the prison population is very fluid and most sentences are about six months duration combined with a lack of continuity in medical records. The practical aspects of undertaking these visits were discussed and the need for good communications was paramount. Also taking a basic minimum of kit was recommended. However annual visits don't guarantee annual screening for eligible patients. More frequent visits are often impractical and it is impossible for prisoners to attend venues outside the prison. Slit lamp biomicroscopy performed by a trained, accredited screener who visits each prison more frequently was suggested as an alternative approach.

Dr Kevin Shotliff, consultant diabetologist for the Chelsea & Westminster Hospital presented the Results of the BARS Survey.  The method was a postal questionnaire survey of the 220 UK members of BARS (175 ‘Graders'). The survey asked for information about

  • Length of standard am and pm screening sessions
  • Number of people screened per session
  • Just image collection / photos or grading as well
  • If fixed or mobile
  • If graded as stand alone activity how many images graded per session

68 individuals replied (31% of those contacted)

  • 55 reported fixed site work
      • Average 14.4 patients per session
      • Range 10 – 26 patients / session
      • 12.7 / session if included grading
      • 16.4 / session if just image collection
  • 30 reported mobile screening work
      • Average travel time of 48 minutes
      • Average 15.7 patients per session
      • Range 11 – 30 patients / session
      • 14.8 / session if included grading
      • 17.1 / session if just image collection
  • Standard morning session
      • Average 3 hours 23 minutes fixed
      • Average 3 hours 14 minutes mobile
  • Standard afternoon session
      • Average 3 hours 5 minutes fixed
      • Average 2 hours 44 minutes mobile
  • Stand alone grading
      • Average 39.3 patients per session
      • Ranged from 20 – 75 patients per session

This equates to an average of 12 patient image sets graded per hour or 1 patient graded every 5 minutes. However there was a range of 1 patient graded between 2½ to 10 minutes. It was hoped that the survey would promote a discussion on what was considered good practice in terms of allowing the graders to grade to national standards. There may be a follow on survey soon.        

There were then 3 short presentations on career development and professional registration followed by a panel discussion. Terry Johnson, the Chair of the 'Voluntary Registration Council' (VRC) spoke about the history of the VRC which was set up following meetings in the summer 2003 between the Department of Health and several professions indicating their intention to proceed to formal regulation. Registration indicates that the profession is regulated and follows a professional code of conduct. The Health Professions Council (HPC) has stipulated that a profession needs to run a voluntary register for two years before formal registration. The effects on retinal screening as a profession are a rationalisation of standards with mandatory training, education and qualification. The granting of professional status means that retinal screeners will be on the same level as Biomedical Scientists, Physiotherapists, Radiographers, Nurses regarding regulation and this will have an impact on Agenda for Change and consequently salaries. For information on the regulation process and how to register details are available on

Christopher Mody, Chair of AHPO (Association of Health Professions in Ophthalmology) gave a talk entitled 'Modernising Scientific Careers (MSC) & Extending Professional Regulation'. Building on previous work around National Occupational Standards for Healthcare Scientists, a draft training and career framework was developed and this lead on to the MSC programme. This describes the process to raise the profile of healthcare science workers (as separate from clinical staff) to define career pathways, qualifications and regulation. Its core principles include patient safety, standards and quality underpinned by assessment, an explicit training/academic career structure, recognition of previous experience, training and qualifications through an equivalence mechanism and supporting more flexible and adaptable training arrangements to create a more flexible and adaptable workforce in the face of changing technology, care settings and care pathways. There are three levels: Healthcare Science Assistant (HCSA), Healthcare Scientist Practitioner (HCSP) and Healthcare Scientist (HCS). There are two career pathways: Healthcare Science Practitioners (HCSP) and Healthcare Scientists (HCS). Each has its own development and learning pathway, supported by an explicit training programme with clear routes to progression. This will enable the development of both a regulated practitioner workforce, as well as a senior scientific workforce. He recommended people register with the VRC, develop a PDP and start to acquire CPD points and additional qualifications.

Grant Duncan then gave a presentation on 'Continuous Professional Development' and how BARS intends to help support this. CPD is the means by which members of a profession maintain, improve and broaden their skills and knowledge. This can be through a variety of routes including the City & Guilds certificate / diploma, the various MSc programmes, attending conferences like BARS, book and journal study, Multi-disciplinary team meetings (MDTs), clinical review sessions and participating in professional bodies such as the BARS council. A points system seems to be recognised as the best way of running the scheme and these are NOT based on formal examination. BARS plans to support the development of CPD with an on-line CDP diary, modelled on those of various Riyal Colleges and with access to help and advice and the provision of workshops. The BARS conference will also be a good way to earn CPD points.

There followed a panel discussion which touched on some controversial points such as what would happen if a degree became the entry level qualification and various issues of contention between the VRC, HPC and MSC. It was clear that a lot more work needs to go into the whole area of regulation and career development.

The conference ended with the TOPCON lecture delivered this year by Professor Paul Dodson (Aston University & Birmingham Heartlands Hospital) on the 'Medical management of diabetic retinopathy: any progress?' The goal is to prevent diabetic blindness now. The strategy is the provision of retinal screening, improved diabetic care, improved treatment uptake and new treatments for diabetes and diabetic retinopathy. A comparison of retinopathy levels between Birmingham and the Leopard Programme in Addis Ababa Ethiopia showed significantly higher levels of retinopathy, especially sight threatening retinopathy (referral rates: Birmingham : 5 – 9 %  Addis Ababa: 51% with 20% at a level which would be registered as blind in the UK) indicating that proper management of diabetes and retinopathy has a huge impact on reducing the risk of progression to blindness.

The targets for medical management are HbA1c <7% (USA <6.5%), Blood Pressure < 130/80 mmHg and blood lipids - Cholesterol < 4, and LDL < 2 mmol/l . Data from the UKPDS showed that lowering the blood pressure had a bigger effect than lowering the blood sugars. (a drop in HbA1c of 1% reduced DR by 25% whereas a drop of BP of 10% reduced the risk of DR by 37%). However there is a legacy effect with a “glycaemic memory” meaning that control of blood sugars in the early years after diagnosis has an impact in future development of DR whereas the effect of blood pressure had no such memory. The risk of a sudden tightening of controls was highlighted as this can also lead to progression.

Diabetes care is moving into the community so patients with sight threatening maculopathy may not be under specialist diabetes care. Patients questioned in the screening environment about their diabetes care are often quite vague about it. Maybe there should be a fast track for these patients to see a diabetologist. The under lying cause will determine outcomes to some degree as ischaemic maculopathy is very difficult to treat. Recent trials have given new evidence about medical management of DR.

CARDS (Collaborative Atorvastatin Diabetes Study) has shown that less laser is used with the patients on statins to lower cholesterol, although further trials with higher doses are needed.  The FIELD(Fenofibrate Intervention and Event Lowering in Diabetes) study has also shown that macular oedema is reduced and there is less need for macular laser and also pan retinal laser for proliferative changes. TheDIRECT(Diabetes Retinopathy Candesartan Trials) study looked at whether Candesartan reduces the progression of DR and the incidence of macular oedema in type 1 and type 2 diabetes. The evidence points to a significant effect in type 2 diabetes. A secondary aspect of the RASS (Renin-Angiotensin System) study was to look at the effects of RAS inhibition on retinal lesions in type 1 diabetes using enalapril (an ACE inhibitor) and losartan (angiotensin II receptor blocker ARB). Evidence would indicate that ACE inhibitors have little effect on DR progression in type 1 diabetes. The DRS Study Group 1 showed that ruboxistaurin mesylate reduced the risk of sustained moderate vision loss by 40 percent when compared to placebo in patients with moderate to severe non-proliferative diabetic retinopathy. It works by limiting the overactivation of protein kinase C beta (PKC beta), a naturally occurring enzyme that has been linked to the development of diabetic retinopathy.

In summary the importance of using evidence based medicine to refine the treatments given to patients was demonstrated. The overall strategy is the use of digital photography and OCT for screening, use of intra vitreal steroids and Lucentis and lowering of the HbA1c by 2-3% , lowering the blood pressure (the lower the better) and lowering the blood lipids combined with new medical therapies using candesartan, fenofibrate and ruboxistaurin in appropriate cases. He suggested a working group be set up to look at the risk management of patients.

Report by Peter Mitchell
Senior Optometrist
Hackney Diabetes Centre
Homerton University Hospital