Lessons from the UK National DR Screening
Program for Areas with Limited Resources
Prof. Peter Scanlon MD FRCOphth FRCP DCH
Programme Director
English National Programme
UK Population
Scotland -population 5.2 million
Northern Ireland population 1.8
million
England population 51.9
million
Wales- population
3.0 million
2.6 million with
diabetes
English National DR Screening Programme
Large Telemedicine Programme
2011-12
81 centres
2.6 million with diabetes
2.4 million offered
1.9 million actually screened
Increase 121,000 in 12 months
Cost approx 80 million US dollars
Screening
What are the risks?
Are the risks changing?
Basis of the ENSPDR Grading Criteria
Key points –
refer at 11.3% risk of developing proliferative in 12 months
R1 = background
= mild NPDR
= do not refer
R2 = pre-proliferative = moderate to severe NPDR
= refer
R3 = proliferative
= refer
ETDRS final
Retinopathy Severity
Scale
ETDRS
(Final)
Grade
Lesions
No apparent
retinopathy
10
14, 15
DR absent
DR questionable
Mild NPDR
20
35
a
b
c
d
e
Micro aneurysms only
One or more of the following:
Venous loops > definite in 1 field
SE, IRMA, or VB questionable
Retinal haemorrhages present
HE > definite in 1 field
SE > definite in 1 field
43a
b
H/Ma moderate in 4-5 fields or severe in 1 field or
IRMA definite in 1-3 fields
Level 41 = 11.3%
Moderately severe
NPDR
47
a
b
c
d
Both level 43 characteristics –
H/Ma moderate in 4-5 fields or severe in 1 field and
IRMA definite in 1-3 fields
or any one of the following:
IRMA in 4-5 fields
HMA severe in 2-3 fields
VB definite in 1 field
Level 45 = 20.7%
Severe NPDR
53
a
b
c
d
One or more of the following:
> 2 of the 3 level 47 characteristics
H/Ma severe in 4-5 fields
IRMA > moderate in 1 field
VB > definite in 2-3 fields
Level 51 = 44.2%
Level 55 = 54.8%
61a
b
FPD or FPE present with NVD absent or
NVE = definite
Moderate NPDR
Mild PDR
Risk of progression
to PDR in 1 year
(ETDRS Interim)
ETDRS follow up intervals
English Screening Programme levels3
R0
Currently screen
Annually
1 year
Level 30 = 6.2%
4-6 months
R1
Screen annually
Background
microaneurysm(s)
Retinal haemorrhage(s)
any exudate
venous loop
3-6 months
4 months
R2
Refer to ophthalmologist
Pre-proliferative
venous beading
or reduplication
intraretinal microvascular abnormality (IRMA)
multiple blot haemorrhages
3 months
R3 Proliferative
Urgent referral to ophthalmologist
Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology 1991; 98:823-33.
WESDR: Twenty-Five Year Progression Of
Retinopathy In Patients With T1DM
Year of diagnosis of diabetes
Better glycaemic control and to a
70
lesser extent BP control may be
1960–69
60
beneficial in reducing incidence of
improvement of DR
Reduction in prevalence of PDR in
more recently diagnosed cohorts
possible benefit of recent
1970–74
1975–80
PDR Prevalence (%)
PDR and increasing odds of
1922–59
50
40
30
20
10
changes in management of
diabetes
0
0–4
5–9 10–14 15–19 20–24 25–29 30–34 35+
Duration of Diabetes (years)
WESDR; Wisconsin Epidemiologic Study of Diabetic Retinopathy
Klein, R et al. Ophthalmol. 2008; 115:1859–1868
Maculopathy
Key points –
M0 = No maculopathy
M1 = Maculopathy
•M0
No
maculopathy
• M1
Maculopathy
•exudate within 1 disc
diameter (DD) of the centre
of the fovea
29% thickening
(Birmingham P Dodson
personal communication)
•group of exudates within
the macula
Did not find CSMO if <1DD
Mr N Dhingra, Wakefield
UK
•any microaneurysm or
haemorrhage within 1DD of
the centre of the fovea only
if associated with a best VA
of  6/12 (if no stereo)
14% thickening
(Birmingham P Dodson
personal communication)
What are the risks of developing
Clinically Significant Macular Oedema
from 2D photographic markers?
OCT photographic clinics for screen test positive maculopathy
Standardising the grading of retinopathy
In 2009/10, English DESP introduced:
Monthly QA test sets for all (1500+) graders in 86 local screening sites
Sets of 30 (Yr 1) or 20 (Yr 2) cases / month, weighted to DR+ cases
Year 1: Up to 12 blocks in numerical order
Year 2: Up to 12 ‘monthly’ sets
Accessed via the internet
Accessed via the internet at their own
place of work or from home
Not referred:
11/157 (7.0%)
Exact:
235/300 (79%)
Exact agreement with R + M grade: Yr 1
Number of Users completing all cumulative blocks -
Mean proportion (%)
agreement with system grade
1301 1278 1235 1112 1003
948
901
827
731
664
564
461
100
90
80
70
60
50
Mean (SD)
40
30
67.1
64.3
65.5
1
2
3
62.3
64.3
61.9
62.5
4
5
6
7
68.5
69.7
66.4
70.5
73.9
8
9
10
11
12
20
10
0
Block number Yr 1 (2009-10)
Trend: p<0.001
Exact agreement with R + M grade: Yr 2
Number of Users completing set in month -
Mean proportion (%)
agreement with system grade
780
886
957
865
940
882
(0)
1016
896
1011 931
952
100
90
80
70
60
50
40
Mean (SD)
72.6
30
67.4
74.2
63.0
62.1
59.2
June
July
August
74.2
70.3
68.3
71.6
75.3
Nov
Dec
Jan
Feb
March
20
10
0
April
May
Sept
Oct
Monthly sets Yr 2 (2011-12)
N.B. No test was presented in October 2011
Trend: p<0.01
Agreement against system & peers
April, Screen no.7
‘System’ grade
Grader 23 -
Agreement against system & peers
April, Screen no.8
R0
R1
R2
R3
‘System’ grade
100%
75%
50%
25%
Grader 23 -
0%
National
My Programme
Screening Programme – why does it matter?
Generic QA Themes & Objectives
Theme
Objective
1. Identify cohort
To maximise offer of screening to all eligible population
2. Inform
To maximise informed choice throughout screening programme
3. Invite
In those who want screening, to facilitate uptake in eligible population
4. Test
To maximise accuracy of screening test
5. Minimising harm
To minimise potential harms from screening
6. Diagnose
To ensure accurate diagnosis
7. Intervene/ Treat
To ensure high quality and timely intervention
8. Outcome
To optimise public health and individual outcomes in target population
9. Staff
To ensure that whole screening programme is provided by a trained and competent
workforce
10. Commissioning and governance
To ensure effective commissioning and good governance of the screening programme
11. User experience/
patient journey
To ensure a high quality journey throughout the screening process
12. Equality
To ensure that screening programmes fulfil their requirements to reduce health
inequalities
Diabetic Retinopathy Screening
How to Start
Buy a Fundus Camera?
Step 1. Manoeuvring around the
politics of funding
Many different levels
Who is going to provided funding to support
Is this going to be run by Public Health Physicians or by
Specialists – Diabetologist? Or Ophthalmologist?
A Champion is needed who has some skills in diplomacy
Budgets need to be ring fenced
Politics change from when a service is getting off the ground
to when it is up and running
European Experience - barriers
Public awareness
Patient compliance
Lack of funding for equipment, training, education
Collaboration between ophthalmologists and diabetologists
Lack of engagement of private providers of eye care
Lack of systematic process, competency, registers, data
Political instability
Access to laser treatment remained poor in a few countries. Some perverse financial
incentives were reported causing for example intravitreal bevacizumab or
triamcinolone being given even when laser is available.
Step 2: Are Assessment and Treatment facilities
available?
Adequate number of lasers and ophthalmologists to
treat
If not - Contract with an organisation that can provide
treatment
Step 3. Identify cohort for invitation
and call - recall
Diabetes Register
How do you record patient details?
If literacy levels are low the patient surname may be
spelled differently at each visit
Is there a National ID number?
Are births and deaths recorded in the population?
Step 4. How are you going to invite them?
In those who want screening, to facilitate uptake in eligible population
Letter?
Word of mouth?
Etc………..
Step 5. How are you going to inform the patients and
maximise uptake?
To maximise informed choice throughout the screening programme
1. Educating the population - this is not a diagnostic test – some patients with
sight threatening diabetic retinopathy will be missed.
2. Patient education, engagement with patient organisations,
3. Appropriate exclusion criteria
e.g. those already under ophthalmology, terminally ill etc..
Step 6. Establish an IT infrastructure
Preferably as simple as possible
Need reliable power supply
An inexpensive joined up solution for
administration of call recall, screening,
grading and audit is an urgent requirement.
Make sure images attached to patient details
Who is going to support that IT iinfrastructure?
How is it going to be backed up?
How are you going to ensure confidentiality of patient data?
Step 7 - Purchase a Camera
Minimum camera specification
Most of the modern non-mydriatic digital cameras meet
a good quality specification
What relationship is there with the camera manufacturer
for technical support in your area?
Image sizes of cameras and recommend compression at
source
Output
resolution
H
V
Output
Resolution in
millions of
pixels
integral
1360
1024
1.39
4.2MB
348KB
209KB
Kowa Non-Myd alpha
integral
1600
1216
1.95
5.8MB
486KB
292KB
Topcon NW100
integral
1792
1184
2.12.
6.4MB
530KB
318KB
Topcon NW6
Nikon D1H
2000
1312
2.62
7.9MB
656KB
394KB
Nikon D1x
3008
1960
5.90
17.7MB
1.47MB
884KB
Nikon D1x
2000
1312
2.62
7.9MB
656KB
394KB
Canon EOS 10D
3072
2048
6.29
18.9MB
1.57MB
944KB
Canon EOS 10D
2048
1360
2.79
8.4MB
696KB
418KB
D30
2160
1440
3.11
9.3MB
778KB
467KB
Camera
Back
Nidek NM-1000
Canon CR6/DGi
Uncompressed
File size in MB
12:1
Compression
in KB
20:1
Compression
in KB
Step 8 - The test and grading images –
Choices for programmes
1.Mydriasis or non-mydriasis?
2.The number of fields
3.The grading referral criteria
4.Viewing the images for grading
The test – mydriasis, selective mydriasis or not?
Clear protocols need to be in place
80.0%
Unassessable image patients for mydriatic (Total 133/3611 =3.7%)
Unassessable image patients for non-mydriatic (Total 711/3604 =19.7%)
70.0%
68.2%
60.0%
50.0%
40.1%
40.0%
36.4%
30.0%
26.2%
20.0%
16.5%
11.3%
11.0%
10.0%
5.6%
5.3%
6.3%
4.8%
2.0%
0.0%
0.0%
0.0%
1.4%
1.2%
16-19
20-29
30-39
40-49
0.8%
50-59
Age Group
60-69
70-79
80-89
90+
The test – number of fields and positioning
The Grading Referral Criteria
Retinopathy progresses with increasing ischaemia
R grade
Leaks occur in the macular area
M grade
The treated patient is more difficult to grade
Recommend R0M0, R1M0, R1M1 etc…. So that every eye has at least an R and M grade
This makes it much easier to compare between programmes
Recommendations on Viewing Images
1. Screen resolution
2. Display 60% of the image at once on the grading screen
Management of patients with ungradable images
Clear protocols need to be
in place
Step 9.Employ and train a competent workforce
To ensure that whole screening programme is provided by a trained and competent
workforce
1. Staff accreditation
2. Evidence of ongoing CPD and EQA test sets
Step 10. introduce some Quality Assurance
1.
Reduce the probability of error and risk
2.
Ensure that errors are dealt with competently and sensitively
3.
Help professionals and organisations improve year on year
4.
Set and keep under review national standards;
5.
Manage these processes.
1. What would I do with 150k USD recurring?
• Start with a pilot project
• Check that assessment and laser treatment facilities in place
• Liaise with local patient groups, ophthalmologists and diabetologists
• Write protocols and decide on patient pathways for screen positive
and ungradable images
• Make sure adequate power supply to screening and grading locations
• Employ someone with IT skills
• Choose software, hardware and back up facilities
2. What would I do with 150k USD recurring?
• Decide on grading form that refers at the agreed level of risk
• Train non medical graders
• Buy a camera
• Provide patients with appropriate education
• Invite cohort for screening
• Photograph eyes
• Send image to central grading where possible.
It is worth doing despite all the obstacles and
organisational difficulties!
Thank you for listening