Paper 6

“The Role of OCT (Optical Coherence Tomography) within the

Diabetic Retinal Screening Service in identifying referable
Diabetic Macular Oedema and looking at correlation for Ethnicity,
Postcode, Age and Gender: A Clinical Audit”

Fiona Heggie
Student ID: S1232902
MSc in Diabetes Care Management
May 2015

0
Statement of Originality
This Clinical Audit Project is my own original work and has not been submitted
elsewhere in fulfilment of this or any other award.

1
Acknowledgements

I would like to take this opportunity to thank the following people for their support and
assistance in the completion of this written work:-

Dr Xinhua Shu – Project Supervisor, Glasgow Caledonian University.

Dr Sonia Zachariah – Clinical Supervisor and Associate Specialist in Ophthalmology,

Diabetic Retinal Screening Service, NHS Greater Glasgow & Clyde.

Dr Jon Godwin – For Statistical and Data Analysis assistance.

My Family – Gordon, Duncan and Euan for their unconditional support in my long
academic journey.

2
List of Contents

Page 1 Statement of Originality

Page 2 Acknowledgements
Page 3 List of Contents
Page 4 Summary and Title
Page 5 Section 1: Introduction
1.1: Epidemiology
1.2: Diabetic Retinopathy and Diabetic Macular Oedema
1.3: Optical Coherence Tomography
Page 17 Section 2: Methods
Page 20 Section 3: Results of the Audit population
3.1: Results in relation to OCT clinic outcomes
3.2: Results in relation to Gender category
3.3: Results in relation to Age category
3.4: Results in relation to audit populatioin and Diabetes type
3.5: Results in relation to the audit Postcode areas
3.6: Results in relation to audit population and SIMD categories
3.7: Results in relation to audit population and Ethnicity
Page 33 Discussion
Page 36 References
Page 39 Appendix 1 – OCT Outcome Crosstabulations
Page 45 Appendix 2 – Correlation categories Crosstabulations

3
SUMMARY AND TITLE

The Role of OCT (Optical Coherence Tomography) within the Diabetic Retinal
Screening Service in identifying referable Diabetic Macular Oedema and looking at
correlation for Ethnicity, Postcode, Age and Gender: A Clinical Audit.
This clinical audit aimed to evaluate the effectiveness of an OCT clinic within the
diabetic retinal screening service, in detecting referable diabetic macular oedema. OCT
is normally a tool reserved for use within the ophthalmology clinic. It is routinely used to
diagnose a multitude of ophthalmic conditions and assist in monitoring treatment
interventions. The diabetic retinopathy screening service was established in 2004 with
the aim of detecting sight threatening diabetic retinopathy to enable timely treatment
interventions. Patients who have signs of referable diabetic retinopathy features are
referred to the hospital eye service for assessment. The DRS Service found that
patients who they referred to the eye clinic were seen and often discharged back to the
screening service without treatment. The DRS Service of NHS Greater Glasgow and
Clyde currently has a 4% referral rate to ophthalmology. During the period of January
2013 to December 2014 the service had a cumulative total of 3,884 ophthalmology
referrals. 725 of these referrals were appointed to the two OCT clinics at the screening
sites in the Southside of Glasgow. The audit found that by using OCT within the
screening service to detect diabetic macular oedema, the amount of ophthalmology
referrals were reduced to just fewer than 14% of the audit population. As part of the
audit no correlation was found with age, gender, ethnicity and postcode. Correlation
was found with SIMD and clinic, Age and Diabetes Type, as well as OCT outcome and
R & L diabetic maculopathy features, and OCT Outcome and R & L Visual acuities.

4
INTRODUCTION
1.1: Epidemiology
The incidence of Diabetes Mellitus Worldwide is set to increase into epidemic
proportions. There are currently 387 million people worldwide with Diabetes, of which
46% are undiagnosed according to the International Diabetes Federation (IDF) 2014.
Of these 52 million people reside in Europe. The IDF estimate that by the year 2035
there will be 592 million people with Diabetes. This is a projected increase of 22.4%. In
the United Kingdom (UK) there are currently 2,974,950 diabetes cases with a
Prevalence of 6.57%. In Scotland alone 268,154 people were diagnosed with Diabetes
in 2013 according to the Scottish Diabetes Survey monitoring group (2014),
representing 5% of the Scottish population. In the same survey it was found that NHS
Greater Glasgow & Clyde, with a population of 1,217,025 people had a total of 61,647
people registered with Diabetes Mellitus which equates to a crude prevalence 5.07%.
Of this total 10.5% are people with Type 1 Diabetes and 88.3% have Type 2 Diabetes.

Diabetes Mellitus is a long term chronic condition which can affect every system of the
Human body, and occurs when there isn’t enough insulin produced (Type 1) or it isn’t
used effectively (Type 2). Without effective insulin hormone production or usage,
Glucose cannot enter the cells to be used for energy. Increased levels of glucose
circulate in the bloodstream attaching to the red blood cells, which can also attach to the
vessel walls causing damage over periods of time.

Diabetes is no longer considered a disease that affects rich countries. It is also

increasing in developing countries. According to the IDF in 2013 the rate of people with
diabetes worldwide who were undiagnosed in low income groups was 29.3%, in
middle income groups it was 35.1% and it was 36.6% in high income groups. In the
united kingdom alone there are 834,000 people who are undiagnosed according to the I
DF 2014 update. Effective management of the condition is essential in order to prevent
complications and mortality. Complications of diabetes include cardiovascular and
cerebrovascular disease (the circulation), retinopathy (eyes), nephropathy (kidneys)
5
and neuropathy (feet). It is therefore imperative, that both the person with diabetes and

their care providers work closely to reduce the risks of developing progressive diabetes
complications and prevent early mortality. In 2013 diabetes accounted for 5.1 million
deaths world wide. In Europe, 1 in 10 deaths could be attributed to diabetes and 147
billion US Dollars were spent on diabetes healthcare provision. The total spend in
Europe per person was between 3,000-6,500 US Dollars (IDF 2013).

Diabetic Retinopathy is a complication of Diabetes Mellitus, the prevalence of which is

suggested to range from 10% in Norway to 61% in South Africa. The figure for
Retinopathy in newly diagnosed diabetics range from 1.5% in African Americans in
the United States of America (USA) to 31% in China. The IDF suggest that the
Prevalence of diabetic retinopathy is higher overall in developing countries but there is a
lack of data. Within Europe the prevalence of diabetic retinopathy in newly diagnosed
people ranges from 6 – 11%. In people known to have Type 2 diabetes the figures
range to between 10.1 – 50.7%. The increased prevalence in low to middle income
countries (although limited data is available) could be due to social determinants to
health including poverty, poor nutrition and poor access to healthcare. It is a disease
that affects the blood vessels supplying the retina at the back of the eye characterised
by micro vascular occlusion and leakage. Diabetic Retinopathy affects patients with all
types of diabetes regardless of age. According to Davis (1992) “the prevalence of
retinopathy and in particular proliferative diabetic retinopathy (PDR) is greater in Type 1
diabetes and is closely related to the duration of diabetes. Prevalence of PDR in Type 1
patients who were < 30yrs of age at diagnosis rose from 2% in those with 5-10 yrs
diabetes duration to 50% in those with >20yrs diabetes duration................after 15yrs
duration of diabetes the prevalence of macular oedema was 12-20% with less variation
by diabetes type. Ribeiro et al stated in 2011 that approximately 25% of patients with
Type 1 diabetes will develop retinopathy within 5 years of diagnosis which increases to
60% after 10 years, with 80% having retinopathy after 15 years duration. Type 2
diabetes is the most common type of diabetes with >50% of individuals who have PDR
and 80% of those with macular oedema belonging to this group”. He also states that
the onset of vision threatening Retinopathy is rare in children pre puberty regardless of
the diabetes duration. This correlates with the age of initiating screening for diabetic
6
retinopathy which is 12 years. MacKenzie et al, (2011) state that 75% of people with
diabetes have retinopathy 20yrs after diagnosis. These figures would suggest a further
potential increasing burden on Health care services.

Many of the signs of Retinopathy can be modified by early detection and review of the
diabetes risk factors. Review of a person’s glycemic control, blood pressure and lipid
status can help slow, and in some cases halt the progression of retinal disease.
However there are diabetic risk factors which cannot be modified. These include family
history, duration of diabetes and ethnicity.

As an integral part of diabetes care, Diabetic Retinopathy Eye Screening is an essential

tool and can be argued has been effective in reducing the rate of severe diabetes retinal
disease (IDF 2013). For many years eye screening methods varied from not only health
board to health board, but also from location to location. Standards of care varied and it
was felt by service providers and healthcare professionals that care should be
standardised for the service user regardless of their point of entry.

Eye care was identified as one of the first seven building blocks of diabetes care in the
Scottish Diabetes Framework in 2002. In 2002 the Health Technology Board for
Scotland (HTBS) published their report on the organisation of services for Diabetic
Retinal Screening as early detection of sight threatening eye disease can enable timely
treatment intervention to prevent sight loss. A year later, the Scottish Executive
published their recommendations for the Implementation of Diabetic Retinopathy
Screening Services based on the HTBS report. The report recommended Digital Retinal
Photography as the most efficient and cost effective method of screening with the target
for implementation as 2006. Implementation of the digital retinal photography service
would vary from health board to health board in their organisational process, but with
the aim of either fixed or mobile screening sites or a combination of both depending on
location. The process would also vary from area to area with the personnel performing
the screening as there was no restrictions. However the HTBS recommended that all
staff participating in the screening process must be effectively trained and accredited.
The process of screening was established and the image grading scheme was

7
developed with the primary aim being “The detection of Referable, potentially sight
threatening retinopathy so that it can be treated”. This is the basis of the National
Diabetic Retinopathy Screening Programme. Gibson (2014) states that the introduction
Of a national Diabetic Screening Service within the four nations has possibly had the
greatest effect on the management of Diabetic Retinopathy in the last 25 years. The
resulting effect on hospital eye clinic services will be the next imminent challenge in
managing the resulting workload effectively.

In a study carried out by Olson et al (2003), single field digital photography with
automated grading, followed by manual grading when disease was detected, was the
most effective screening method. The study looked at the sensitivity and specificity in a
variety of screening methods including slit lamp bio microscopy, digital photography,
conventional photography (using 35mm film) and slit lamp performed by an
Ophthalmologist. In a study carried out by Vujosevic et al in 2009, the suggestion was
“that a single, central 45 degree field image is sufficient to determine the absence or
presence of DR and DMO, but not for grading it”. They also stated that grading of non
mydriatic fundus photographs should be considered the new standard in DR screening.

In implementing the Diabetic Retinal Screening Service (DRS) within NHS Greater
Glasgow & Clyde, the cameras used are Canon Mk 2 Fundus cameras with 50 and 60D
Digital backs. The camera operators are nurses and clinical science technicians whilst
the Level 1 & 2 graders are nurses, optometrists and experienced screeners. The DRS
camera operators and grading staff have been trained in the Diabetic Retinopathy
Screening (DRS) National Qualifications for their role through the City & Guilds Institute,
London. The graders are subject to regular internal and external Quality Assurance
processes. The grading of the digital images is carried out by feature recognition and is
a three tier process. In first level grading the images are checked for presence or
absence of disease. Initially this was done manually but is now done by a combination
of both autograding technology and manually. Images with disease are then checked
by a Level 2 grader. The Level 2 grader determines the extent of the pathology that is
present, its location, and potential outcome. Images with no, mild or moderate
retinopathy (R0, R1 and R2) and no or observable maculopathy results (M0 and M1

8
respectively) are graded at L2, generating a recall period of 6 months or 12 months.
Images with severe background or proliferative retinopathy (R3 or R4) or referable
maculopathy (M2) are advanced to the next level. The final level 3 grading is carried
out by an Ophthalmologist who arbitrates the final outcome for these patients. The table
below shows the National grading protocol.
According to the Scottish Diabetic Retinopathy Grading scheme (2007) noted in table 1,
Retinopathy grades of R3 or R4, and a Maculopathy grade of M2 are all referable
grades. The retinal Images are also graded according to a 5 point grading scale for
image quality (with a grade 1 image being the best quality and grade 5 image being the
worst). Images are checked firstly for signs of retinopathy with a grade (R) being
assigned, then graded specifically for signs of maculopathy and assigned a grade (M).
A person may not have a referable Retinopathy grade e.g. R1 or R 2, but if they
have a maculopathy grade of M2 then they will be referred to Ophthalmology, as per the
DRS protocol. The Maculopathy grade, if more severe, supersedes the Retinopathy
grade as features have been detected in an area where the central vision could be
affected.

From the Diabetic Retinal Screening (DRS) service in NHS Greater Glasgow and Clyde,
the current referral rate to ophthalmology is 4%. The Referral rate varied from 3 – 6%
before steadying consistently at 4% which is comparative to the national average. The
cumulative total for referrals to ophthalmology from Jan 2013 – Dec 2014 (the audit
period) for the DRS Service was 3,884 (local service data). 725 of these referrals were
appointed between two OCT clinics based on their M2 grading for maculopathy. This
equates to 18.6% of the cumulative referrals. With the current referral rate to
ophthalmology at 4%, this has a burdening effect on the hospital eye care services. This
view is further supported by Manjunath et al (2015) who suggests that the increased
workload in ophthalmology clinics generated by the screening programmes has been
considerable.

Patients referred to the hospital ophthalmology service require to be assessed by an

Ophthalmologist (eye specialist) to determine the need for treatment.

9
Table 1: Retinopathy grading features and outcomes
Retinopathy Category Features Present Outcome
Grade

R0 No Retinopathy None Rescreen 12 months

R1 Mild Background The presence of at least Rescreen 12 months

Diabetic Retinopathy one of the following –
(BDR)
Dot / Blot haemorrhages,
microaneurysms, hard
exudates, cotton wool
spots, superficial / flame
haemorrhages
R2 Observable BDR 4 or more blot Rescreen 6/12 months
haemorrhages in one hemi-
field only (inferior or (or refer to ophthalmology
superior hemi-field) if this is not feasible)

R3 Referable BDR Any of the following Refer to Ophthalmology

features :-
These patients may be kept
4 or more blots in both under surveillance
hemi-fields, Venous
Beading or IRMA
R4 Proliferative Diabetic Any of the following Refer to Ophthalmology
Retinopathy (PDR) features:-
These patients are likely to
Active new vessels at disc receive laser treatment or
or elsewhere, Vitreous another intervention
Haemorrhage
M0 No Maculopathy No features present ≤ 2 Rescreen 12 months
Disc Diameters from the
centre of fovea sufficient to
qualify for M1 or M2
M1 Observable Lesions are specified below Rescreen 6 months (or refer
Maculopathy within a radius of ≥1 but ≤ to ophthalmology if this is
2 DD from centre of fovea not feasible)

M2 Referable Maculopthy Lesions as specified below Refer to Ophthalmology

≤1 DD of centre of fovea
These patients may be kept
Any Blot haemorrhages, under surveillance and will
Any hard exudates not necessarily receive
immediate laser treatment
R6 Not adequately Retina not sufficiently Technical Failure Pt
visualised visible for assessment requires alternative
screening method
Adapted from Scottish Diabetic Retinopathy Grading Scheme 2007 v1.1.
10
Initial audits in the DRS Service of NHS Greater Glasgow and Clyde concluded that not
all M2 referrals to ophthalmology, based on the grading protocol, required treatment and
were after the first visit discharged to the DRS Service. In order to streamline these
referrals from the DRS, an OCT clinic was piloted. This clinic was established as an
intermediary clinic between the DRS Service and the Hospital Eye Service.

The initial pilot prior to January 2013 involved patients identified as M2, having both
eyes scanned using the OCT 3D scanner and a concurrent clinical examination by slit
lamp bio microscopy. The purpose of the slit lamp examination was to establish if the
clinical findings correlated with the findings and interpretation of the OCT scan. Results
of the pilot showed not just a close correlation between the clinical examination and
OCT, but a significant reduction in the referrals in to the hospital eye clinic. Based on
these results it was decided that the clinic could progress effectively as a virtual one
without the slit lamp examination.

1.2: Diabetic Retinopathy and Diabetic Macular Oedema

Diabetic retinopathy is one of the most common causes of blindness in people of a

working age not only in Scotland and the UK, but worldwide. It is a progressive

disease, which, if left untreated will affect the patients central vision and can lead to

visual loss. In its early stages diabetic retinopathy is symptomless for the diabetic

individual. It is only when the macula (area of central vision) starts to become affected

that the individual may notice changes in their vision.

Davis (1992) describes 5 fundamental stages in the natural course of diabetic

retinopathy. He describes the first pathological change as being the formation of retinal
capillary microaneurysms which appear as hypercellular sacs on the capillary wall. Their
walls can become thickened and as they do so, they becomes less transparent. This is
then followed by excessive vascular permeability, which is due to changes in the retinal
- blood barrier. With increased numbers of microaneurysms retinal changes within the
capillaries occur. These include capillary dilation and capillary non perfusion. In these
cases the eyes have usually progressed to developing retinal haemorrhages or hard

11
exudates. Hard exudates can be dispersed in the fundus but are more usually found in
partial or complete rings (circinates of hard exudate) which indicate the area of
thickened, oedematous retina that usually surrounds one or more microanuerysms.
The small molecules and lipids are dispersed within the oedematous areas at the
edges and become reabsorbed across the walls of surrounding normal capilliaries.

Diabetic macular oedema, along with proliferative retinopathy, is a common cause of

sight threatening retinopathy. There are two causes of visual loss from diabetic
retinopathy. These are proliferative retinopathy and macular oedema. Macular oedema
is more common than proliferative retinopathy and is the leading cause of moderate
visual loss in people with diabetes. The posterior pole is the most common area of
macular oedema. Macular oedema according to Nisic et al (2014) is defined as “the
accumulation of liquids and other substances (lipids and proteins), which leads to the
thickening of retinal tissue and its functional decrease”. According to Alkuraya et al
(2005) the incidence and prevalence of macular oedema increases with the duration of
diabetes and increasing severity of concurrent retinopathy. This is reinforced in many
reviews and studies including the Health Technology Assessment carried out by Olson
et al (2013). They describe a figure of 60% of people with type 2 diabetes having
retinopathy 20 yrs following onset. This updates the findings of Davis (1992) who
suggested that “after 15 years duration of diabetes, the prevalence of macular oedema
is 12 – 20%, and approximately 80% of these people had type 2 diabetes” more than 20
years earlier. Type 2 diabetes is often diagnosed several years later following onset
with as many as 40% of people with type 2 diabetes having signs of retinopathy at
diagnosis. Delcourt et al (2009) state that diabetic macular odema account for 17% of
all the diagnosed cases of diabetic retinopathy in eight population based studies that
they reviewed.

Visual Acuity in patients with Diabetic Macular Oedema is not only determined by the
amount of retinal swelling and the consequential changes but also by the neuro
ophthalmological pathologies, types and amount of previous treatments according to
Boltz et al (2014). The aim of diabetic retinopathy screening is to detect possible sight
threatening changes and initiate timely treatment. OCT has been found by Buabbud et

12
al (2010) to be more sensitive to the presence of diabetic macular oedema than clinical
examination is.

1.3: Optical Coherence Tomography (OCT)

Optical Coherence Tomography has been in use initially in the measurement of retinal
thickening, since the first development of time domain scanners in 1991. OCT is a non
invasive, in vivo ophthalmic technique providing multiple cross-sectional A- scan images
of the retina. It has repeatedly proved its value as a diagnostic tool for assessing retinal
disease by providing qualitative and quantative analysis of the retina. Visualisation of
the layers of the retinal is important in the assessment and management of retinal
disease and this has been made easier by the replicablility function of the OCT
scanners. According to Jaffe and Capriole (2004) these include macular holes, macular
cysts, vitreomacular traction, subretinal fluid, pigment epithelial detachment and
choroidal neovascularisation. Multi rapid A-scans are taken from light reflected from the
retina by low coherence interferometry, against the time delay of the same light beam
reflected onto a reference mirror. The length of time of scan acquisition could be a
problem due to patient movement altering the image quality. Technology continued to
improve and the new generation spectral domain scanners were introduced. These
devices have the ability to acquire data up to 50 times faster than the time domain
scanners, with speeds of between 18-40,000 scans per second. Comparison of the
various spectral domain OCT devices available was originally carried out by Gabriele et
al (2008) and listed by Sakata et al (2009) in their review paper. Image quality and
resolution improved allowing quantative reliable data. Schimel et al (2011) also states
that it is an important tool in monitoring the response to treatment of diabetic macular
oedema.

Below (Fig 1) is an OCT image showing the retinal layers, while Fig 2 and Fig
3 show a normal OCT scan and OCT scan with macular oedema respectively.

13
A Normal OCT Image clearly showing the individual layers of the retina

Fig 1. ( image accessed on 23/04/15 @ www.retinareference.com)

The following images show a normal OCT in colour (Fig 2) and a scan showing macula
oedema (Fig 3).

Colour OCT scan Image showing no abnormality

Fig 2. (Image accessed @ www.illinoisretinainstitute.com on 23/04/15)

OCT scan image showing macular oedema

Fig 3. (image accessed @ www.imagearcade.com on 23/04/15)

Manjurath et al (2015) conducted a study using a combination of clinical examination,

14
retinal images, wide field photographic imaging and OCT to determine the method that
detects the maximum amount of diabetic pathology and agreement of the various
methods. The aim was to evaluate if wide field imaging and OCT could improve the
management of patients in the hospital eye service. They found that this method could
be used to assess and manage referrals from DRS. Other studies and reviews have
used stereoscopic photography (Ahmed et al, 2006 and Delcourt et al, 2009 ) , slit lamp
bio-microscopy (Alkuraya and Abu el-Asrar, 2005) and two field digital photography
(Mackenzie et al, 2011) in conjunction with OCT.

With the prevalence of diabetes increasing it would therefore seem natural that the
more methods available to help make diabetes services more efficient and effective with
referrals to the hospital eye clinic will be of benefit to the patient. It has been
suggested by Manjunath et al, that only 10% of referred patients to the eye clinic
actually require treatment.

In order to control and reduce the incidence of blindness occurring in people with
diabetes, it is important to educate both patient and health care professionals. We all
have a responsibility to help reduce and slow the development and progression of
diabetic retinopathy. There is much that the person with diabetes can do to reduce their
risk of developing diabetic retinopathy. Diabetes care is a partnership between the
health care professional and the patient. The cost of Prevention is better than the
overall costs to cure the disease. The risk factors to developing diabetic retinopathy are
increased hyperglycemia and hypertension.

Fletcher & Chong (2008) state that intensive glucose control reduces the incidence of
microvascular complications by 25% in 10 years and a 1% reduction in HbA1c leads to
a 24% reduction in the 14 year incidence of diabetic macular oedema. It is also stated
that by maintaining HbA1c at <7%, progression of retinopathy is haulted. Intensive
blood pressure control reduces the risk by 47%, of a 3 line vision loss in people with
Type 2 diabetes. These are simple measures that can be adopted through patient
education that can have an effect on retinopathy progression without the need for
ophthalmological intervention. How we chose to reduce hyperglycemia is also a

15
consideration. If the tightening of diabetes control is too rapid and intensive we run the
risk of the retinopathy getting worse. Therefore it is important to employ a controlled
approach. The method of tightening diabetes control must also be considered as some
medications have been associated with macular oedema.

16
METHODS
Advice was obtained from Research and Development staff at NHS Greater Glasgow
and Clyde, and after using the Health Research Authority (HRA) decision tool it was
determined that ethical approval was not required. A comprehensive, retrospective
audit of clinic appointment data was carried out for OCT clinics at two sites in the south
of Glasgow. The data review parameters were restricted to the period outlined
01/01/13 to 31/12/14 for both clinics. A total of 1,185 appointment episodes
were recorded which were further refined by removing episodes of non attenders,
review appointments for patients seen before the above period and review
appointments for non diabetic pathology. Following refinement, this yielded a total of
n= 725 appropriate appointment episodes. Anonymised data was then recorded for
each of these episodes. Data fields recorded were clinic location, appointment date,
date of diabetes diagnosis, date of first recording of Retinopathy on DRS database,
diabetes type and Retinopathy referral grade. Additional fields were also recorded
including individual diabetes retinopathy features present in both eyes, visual acuity for
right & left eye, gender, age, ethnicity, postcode area, presence of diabetic macular
oedema right & left eye, OCT examination outcome, any subsequent OCT review dates
with outcomes and dates of eye clinic attendance.

The screening images were taken with Canon Non Mydriatic Fundus Cameras and

Canon EOS digital camera bodies as part of the national diabetic retinopathy screening

programme. The images taken were single field, 45 degree standard macular field

images of each eye. The images were assessed by the photographer in real time for

image quality and clarity, with mydriatics being used for any images which had a clarity

problem. The images were graded post examination for diabetic retinopathy and

maculopathy according to the DRS 2007 grading criteria. Grading review station

monitors were 19 inches, high resolution and definition cathode ray tube (CRT)

computer screens to aid detection of retinal pathology. Level 3 grading was carried out

by any of three DRS ophthalmologists. Vetting and selection of the patients for the OCT

17
clinics was carried out by a single ophthalmologist, who would also interpret the OCT

scans and arbitrate outcomes from this clinic ensuring consistency. The ophthalmologist

works both in the screening programme and in the Secondary care ophthalmology

clinic, at both of the clinic locations. By having a role within DRS, and the secondary

care clinics, they were in a position to see the burdening effect of the referrals from

DRS into secondary care services and the management of these.

The OCT scanner used was a Topcon 3D OCT 2000 spectral domain scanner which
has the capacity to take 18,000 scans per second. The clinics were carried out weekly
on a Wednesday afternoon in clinic 1 and fortnightly on a Friday morning in clinic 2 (due
to access restrictions). The patient had their ophthalmic history and visual acuity
recorded at the clinic appointment followed by their OCT scan of each eye using the 3D
retinal scan setting. The length of time for each scan was 5 minutes. Pupil dilation was
carried out only for patients who required it due to pupil size and clarity of the scan.
Patients who required pupil dilation had to wait on the mydriasis taking effect (a
minimum of 15 minutes). Review of the completed scans was carried out on a weekly
basis by the ophthalmologist on a wednesday and thursday afternoon. The scans were
reviewed on the Topcon OCT Viewer programme which could be accessed from
networked, designated review stations at DRS grading centre and at both clinic sites.
The ophthalmologist also had access to all the relevant information of DRS retinopathy
screening episode including retinal images, visual acuity recordings and outcome
results whilst reviewing the scans. The ophthalmologist interpreted the OCT scans
and determined the outcome based on the OCT and photographic findings, combined
with the visual acuity assessments and screening history.

There were several outcomes possible for patients who attended the OCT clinic for their
scan. Depending on the findings, they could either be recalled for a 6/12 review within
the OCT clinic, referred to the hospital Eye clinic for assessment / treatment or be
recalled for review in DRS service in 6/12 or 12/12 intervals. The DRS reviews were
either Photographic reviews or slit lamp reviews. All administrative duties for
the OCT clinics were carried out by one member of the administration staff. Patients

18
were advised of their scan results by letter. Patients who cancelled (CAN) their OCT
appointment were given a new one depending on the reason. Patients were contacted
by the admin person a few days prior to the appointment as a reminder. The importance
of attending for the scan was stressed at this time. Patients who did not attend (DNA)
their appointment were sent another appointment. If the patient had two appointments
that they did not attend, they were either referred into the eye clinic by the
ophthalmologist or discharged back to the screening service for a 6/12 review
depending on the original M2 photography grading.

19
RESULTS
The audit included a total of 725 patients across the two clinics. Particular emphasis of
the audit was to examine the effectiveness of using an OCT scanning clinic to
streamline referrals to the hospital eye service, as well as looking at the possibility of
any correlation patterns with gender, age, postcode and ethnicity.

3.1: Optical Coherence Tomography (OCT) Outcomes

To determine the effectiveness of the clinic we used the OCT clinic outcomes as a
measurement. There were 4 possible outcomes for the clinic. These were 6/12 OCT
review, referral to the Eye Clinic, DRS 6/12 review and DRS 12/12 review. The
outcomes were examined for both clinics and there was no significant pattern found
(Pearson chi square test 2p=0.659, df 6).
The bar chart (Fig 4) below shows the distribution of outcomes per clinic.

OCT scan outcome in relation to clinic distribution

Fig 4.

20
The outcomes were similarly distributed across both clinics as noted in table 2 below.

Table 2: OCT Outcome result figures for clinic 1 and 2.

DRS DRS DRS EC Ref OCT OCT Ref EC Totals
12/12 12/12 6/12 6/12 6/12
S/L DNAx2
Clinic 1 127 1 90 62 214 1 1 496
Clinic 2 64 0 43 36 86 0 0 229

totals 191 1 133 98 300 1 1 725

OCT Outcome and diabetic features Right & left eye were examined. There was similar
distribution of outcomes across both sets of features. The results were Right diabetic
maculopathy features (Pearson chi square 2p=0.446) and Left diabetic maculopathy
features (Pearson chi square 2p=0.018). The largest groups for maculopathy features of
both eyes were no features, exudates and blot. Exudate was the largest group for Eye
Clinic (EC) referral in both eye feature groups (33 for Right and 31 for Left respectively).

OCT outcomes and Visual acuity for right and left eye was examined for correlation. A
significant correlation was found for both eyes. Pearson chi square 2p=<0.001 was the
result for Right and Left visual acuity. The notable visual acuities for EC referral for both
eyes were 6/12 (RE 10 & LE 12), 6/9 ( RE 33 & LE 32) and 6/6 ( RE 36 & LE 26).

OCT outcome was also examined with SIMD (Pearson Chi Square test 2p= 0.746). The
outcome categories were distributed across all 5 categories with 6/12 OCT review being
the largest group for each SIMD. The results of OCT outcome with Age (Pearson chi
square test 2p=0.997). The Age group 61-70 years had the largest EC referral group
(29) and DRS 6/12 (36). Individually age 61 years had the largest group for both EC
referral (7) and DRS 6/12 review (11). There were more males than females in each of
the OCT outcome categories correlated with gender (Pearson chi square test 2p=0.624)
and referral grade (Pearson chi square 2p=0.935) were not significant. .The result of
OCT outcome with Ethnicity (Person Chi Square test 2p=0.041) was significant. White
Scottish was the largest ethnic group (n=400) and had the largest numbers in each of
the outcome groups. The above crosstabulations can be found in appendix 1.

21
Data was also obtained for 6/12 OCT clinic 1st review (n=300) which is shown in Fig 5
chart below.

Proportion of OCT Outcomes for both clinics in the whole audit

population

Fig 5.

The OCT Review crosstabulation is shown in table 9 of Appendix 1. The main focus of
the 1st OCT review outcome is the amount of referrals to the Eye clinic (n=48) and
also for follow up within the DRS OCT clinic (n=60), and patients returned to the DRS
screening clinic for 6 month review (n=74) and 12 month review (n=43). 2nd OCT 6/12
review (n=60) although not all data for 2nd review is complete as some appointments
were awaited .

22
3.2: Gender
The gender of the audit group consisted of 414 males and 311 females equating to
57.1% and 42.9% respectively as shown in fig 6 below.

Gender distribution for both clinics within the audit population

Fig 6.

The bar chart below at fig 7 shows the gender distribution across the two clinic sites and
no association was found (2p = 0.747, Fisher’s exact test).
Audit population Gender distribution across clinics 1 and 2.

Fig 7.

23
The gender distribution across diabetes type (type 1 or 2) was also examined. This is
shown in table 3 and the bar chart below (fig 8). There was no association found
between gender and diabetes type (2p = 0.4, Fisher’s exact test)
.
Table 3: Audit population‘s Gender distribution across Diabetes type

Gender * Diab_Type_1_or_2 Crosstabulation

Count

Diab_Type_1_or_2

Type 1 Type 2 Total

Gender Female 60 251 311

Male 92 322 414

Total 152 573 725

Audit population’s Gender and Diabetes type bar chart

Fig 8.

The proportions were:-

Type 1 = 60 females (39.4%) and 92 males (60.5%)
Type 2 = 251 females (43.8%) and 322 males (56.1%)

A summary of gender and clinic distribution in relation to age has been collated below
In table 4 into age groups for ease of reference.
24
Table 4: Audit population gender and clinic distribution in relation to age groupings

Age 20-30 31-40yrs 41-50yrs 51-60yrs 61-70yrs 71-80yrs 81-90yrs 91-93yrs

yrs
F = 7 27 44 69 69 58 34 3
M = 19 34 70 122 99 56 14 0
C1 = 17 45 82 129 109 74 37 3
C2 = 9 16 32 62 59 40 11 0
Codes: F = Females, M = Males, C1 = Clinic 1 and C2 = Clinic 2

There was however a significant gender difference found in ages. The Linear-by-Linear
trend Chi-Square test result was 2p = <0.001. The chi-square test and bar chart are
highlighted in table 5 and Fig 7 respectively.

Table 5: Gender and Age Chi-Square Test

Asymp. Sig. (2-

Value df sided)

Pearson Chi-Square 102.390a 72 .011

Likelihood Ratio 115.336 72 .001

Linear-by-Linear Association 16.459 1 .000

N of Valid Cases 725

Audit population Gender distribution across Age

Fig 9.
25
3.3: Age
Using the Kolmogorov-Smirnov hypothesis test, the distribution of age was found to be
normal across both clinics. Fig.10 below shows the age distribution across the audit
population, while table 6 shows the mean age in each clinic being 58.33yrs in clinic 1
and 58.48yrs in clinic 2.

Table 6: Audit population Mean age distribution to each clinic

Clinic N Mean Std. Deviation Std. Error Mean
Age 1 496 58.33 14.837 .666
2 229 58.48 14.138 .934

A t-test showed that the mean ages of clinic 1 and 2 were not significantly different. The
mean age of the whole audit population was 58.38years as shown in Figure 10

Age distribution histogram for total audit population

Fig 10.

Using the Mann-Whitney U test the distribution of age across the two clinic categories

was not significant as shown in table 4, as they were the same.

26
Table 7: Age distribution and clinic hypothesis test

Histogram showing the distribution of Age across clinics 1 and 2

Fig 11.

3.4: Diabetes Type

The IDF found that in 2013 the greatest number of people with diabetes were within the
ages of 40 to 59 years old. The pattern across the audit population was comparable to
this data. In Europe in 2013, 37% of the diabetes population were over the age of 50
yearswhich is expected to rise to over 44% by 2035.

According to the 2013 Scottish Diabetes Survey, the distribution of diabetes nationally
by diabetes type is 10.9% for Type 1 diabetes and 88.2% for Type 2. NHS Greater
Glasgow and Clyde’s distribution figures are comparative with the national figures.
However the pattern for this sample population differed at 20.9% having Type 1 and
27
79.1% with Type 2. The mean age differed according to diabetes type. The results
were Age and Type 1, n=152, Means 44.41, Std Dev 12.992, with Age and Type 2,
n=573, Std Dev 12.649 respectively. This was confirmed by an independent T-Test. The
Chi- Square test for trend in age and diabetes type was very significant (2p= <0.001).
The histogram in Fig 12 shows the distribution of age across diabetes type, however
this would be expected due to age at diagnosis.

Distribution of age and diabetes type histogram

Fig 12.

There was no effect found for age and ethnicity.

3.5: Postcode
There was a significant pattern for postcode and diabetes type. Pearson Chi-Square
test was 2p = 0.022. The crosstabulation between postcode and diabetes type (table
15) and Chi square test result (table 16) can be found in appendix 2. There was a

28
total of 27 postcode areas within the audit population. The distribution of Type 2
diabetes was spread across all but 2 of the postcode areas. However there were 8
postcodes that did not have patients with Type 1 diabetes, none of which overlapped
with Type 2.

Diabetes type distribution across audit population Postcode area

Fig 13.

There was however no effect for postcode in relation to age.

3.6: Scottish Index of Multiple Deprevation (SIMD)

Due to the large numbers of postcodes involved in the audit it was felt that it may be
more relevant to look at the SIMD categories. SIMD categories range from Category 1

(the most deprived area), to Category 5 being the least deprived. In clinic 1 the audit

population was 496 people. Of these 215 were found to be in deprivation category

3 followed by 148 in category 4, 75 in category 2, with 32 in category 1 and 26 in

deprivation category 5, as shown in Fig 14.

29
Clinic 1 & 2 distribution in relation to SIMD categories

Fig 14.

The results were different in clinic 2 with a total of 229. The largest group of 135 was in

category 2, followed by 53 in category 1, 18 in category 3, 13 in category 4 and 10 in

category 5. Clinic 1 dominated SIMD categories 3, 4, and 5 with 53.6% of the audit

population, and Clinic 2 having 40.6% in categories 1 and 2. Table 8 shows the chi-

square test for SIMD and clinic.

Table 8: Chi-Square Test for SIMD and Clinic for whole audit population

SIMD and Clinics Value df Asymp. Sig. (2-sided)

Linear-by-Linear Association 146.189 1 .000

N of Valid Cases 725

Chi-square (1) = 146.1, 2p < 0.001 as shown by the Linear-by-Linear association test.

There was no significant difference between gender and SIMD. The Chi-Square

(hetrogenity) test picked up a difference (Pearson Chi-Square 2p = <0.001, df 4).

However the Linear-by-Linear test for trend did not. Please see appendix. More males

were detected in categories 1,3, 4 and 5 while in SIMD category 2 females were

30
dominant. Table 18 in appendix 2 shows the chi square test results for the SIMD
categories and clinic population. Also in appendix 2 the crosstabulation for SIMD and

Clinic (table 19) and table 20, which shows the crosstabulation between SIMD and

Gender. There was no significant difference noted between SIMD and Age or for

SIMD and Ethnicity.

3.7: Ethnicity

One of the aims of the audit was to determine if there was any correlation to ethnicity.

The sample size for this was reduced due to 16 items of missing data. There was a

correlation in the ethnicity patterns across both clinics found which is highlighted

in Fig 15. Pearson chi-square test was significant (2p = 0.002). The

crosstabulation table is shown in appendix 1.

Ethnicity distribution for the audit population across clinics 1 & 2

Fig 15.

31
There was however no significant difference found for ethnicity and gender.

The audit encompassed 27 postcode areas, two of whom had the largest number of
patients from the audit population. These were G41 and G52 with 85 people each. G41
was found to have the most diverse ethnic grouping with 6 groups (white Scottish /
british / other), pakistani, indian, asian, chinese and bangladeshi. Table 17, in
appendix 2 shows the Crosstabulation for ethnicity and clinic distribution.

32
DISCUSSION

OCT has been revolutionising the diagnosis and management of ophthalmic diseases

since its introduction in the early 90’s. The practicalities of OCT mean that fluorescein

angiography may no longer routinely be the first mode of examination for retinal

vasculiture. The non invasive technology of the OCT which provides thousands of high

quality scans per second, is not only convenient for professionals but also for patients

alike. High quality scans providing exceptional detail of the retinal layers and

vasculiture can be achieved in less than 5 minutes. Virgili et al (2015) carried out a

comprehensive review of existing studies using OCT and state that the availabilty and

decreasing cost of OCTs is making them an attractive means of improving the

detection of diabetic macular oedema. These studies have all debated the use of OCT

in conjunction with other modes of clinical examination and in some cases they have

suggested that OCT is more sensitive. Stereoscopic retinal fundal photography has

been use in several studies where macular thickness was measured by OCT. This

audit has reviewed the use of OCT detecting diabetic macular oedema following

maculopathy feature recognition on single field fundus photographs. The audit results

suggest that OCT is effective in detecting fluid accumulation and its used in conjunction

with retinal photography review, visual acuity recordings and patient ophthalmic history

can forge a role within the diabetic retinal screening service. In 2000, Newsom et al

found that digital colour photography was highly sensitive in detecting retinopathy but

was not sensitive in detecting clinically significant diabetic maculopathy. They did

however suggest that a two stage screening process could be implemented for patients

with suspicious maculopathy features. Delcourt et al, 2009 also found that screening

programmes using non mydriatic fundal photography was effective in identifying both

early and advanced diabetic retinopathy. Effective screening for retinopathy combined

with OCT use for patients with macular pathology as part of a two stage process could
33
therefore prove to be an efficient and effective use of resources. This two stage process

could effectively reduce the increasing demands on the hospital eye care services.

Some clinicians argue that OCT could become the new gold standard for assessing

macular oedema, while others (Goebel et al, 2002 and Sandhu et al, 2005) feel that it

should be used in conjunction with fluorescein angiography and biomicroscopy.

However Puzyeyeva et al (2011), argue that even high resolution spectral domain OCT

scanner have their limitations. Lesions like major retinal vessels, exudates or

haemorrhages are hyper reflective. This results in shadowing of the underlying

structures and loss of detail. Virgili et al reviewed 10 studies, only 3 of which reported

accuracy in OCT detection of diabetic macular oedema.

In 2009, the Cost of OCT usage was examined in an analysis carried out by the Ontario

Health Technology (OHT) Advisory Committee. The OHT assessment found that an

estimation of the cost for OCT examination amounted to $26.67 (average fee) as

compaired with the cost of fluorescein angiography at $46.35 including interpretation.

The costs were considerably higher when physicians were performing the examinations

($225.90 for 1 biomicroscopy test, 2 fluorescein tests and 4 OCT tests) for diabetic

macular oedema. The hospital costs were reduced by $138 per patient when moving

from the traditional 3 flourescein tests to 2 when using OCT for diagnosis and

monitoring. A cost saving of $2.6 million was implied for diabetic macular oedema.

Olson et al, 2013 also carried out a health technology assessment into the economic

value of photographic screening for detecting diabetic macular oedema by OCT. This

was a multicentre study which found that costs were reduced from £985 per incidence

of macular oedema being detected using a higher specificity, simple manual

grading system, to £528 when coupled with OCT in the screening pathway. Olson

concluded that by introducing OCT into the pathway results in “cost savings without

reducing health benefits”. The author would suggest that by introducing OCT as a

second stage to the screening process, that a reduction in the amount of referrals to the

34
hospital eye service would occur as the referrals would be dealt with more

appropriately. Patients with signs of maculopathy and subsequent intraretinal fluid

following OCT, will be appropriately referred to specialist eye care services. There

would be a reduction in the appointment demands of the specialist clinics from DRS

inappropriate referrals. The hospital services often involve a lengthy visit with numerous

tests to assist in diagnosis. Introduction of OCT scanning within the screening pathway

will assist the diagnosis process, and for the patient reduce the demands on their time

as it is generally a quick examination. The OCT scan does not require patients to have

dilation and a retinal image is also taken at the same time. Reproducibility of the scan

is high as the exact area can be duplicated by the newer advanced spectral domain

scanners. With the Glasgow audit we identified 725 patients with M2 grades for the

OCT clinics. These patients represented a fifth of our cumulative total of 3,887 referrals

in the two year period over both clinics. By introducing OCT as an interim step within the

screening process, we were able to reduce the number of referrals to 99. This has cost

saving implications to the health service with regards to appointments, in an

already burdened service where clinics are often over subscribed and waiting times are

long ( in some cases being from 1 to 3 hours).

As previously mentioned, physician time and costs are high. There are currently

shortages in the provision of ophthalmologist services. Therefore introduction of OCT

scanning as part of the DRS service, carried out by suitably trained screening staff will

have a cost benefit. Costs for a technician completing the scans will be less than

nursing or optometry staff costs.

OCT clinics can be set up within the national screening database in the same way as

screening clinics. However additional technological provision would be required for

storage of the OCT scan. Within Glasgow our current appointments are every 10

minutes. Appointment times within the hospital eye clinics can vary depending on the

type of patient (if they are a new patient appointments can take longer than if it is a

35
review patient).

REFERENCES

"Optical coherence tomography for age-related macular degeneration and diabetic

macular edema: an evidence-based analysis", 2009, Ontario health technology
assessment series, vol. 9, no. 13, pp. 1-22.

Ahmed, J., Ward, T.P., Bursell, S.E., Aiello, L.M., Cavallerano, J.D. & Vigersky, R.A.
2006, "The sensitivity and specificity of nonmydriatic digital stereoscopic retinal
imaging in detecting diabetic retinopathy", Diabetes care, vol. 29, no. 10, pp. 2205-
2209.

Alkuraya, H, Kangave, D and Abu El-Asrar, A.M. 2005. “The correlation between OCT
features and severity of retinopathy, macular thickness and VA in diabetic macular
oedema”. Int. Ophthalmology. 26:Pg 93-99

Bolz, M., Lammer, J., Deak, G., Pollreisz, A., Mitsch, C., Scholda, C., Kundi, M.,
Schmidt-Erfurth, U. & Diabetic Retinopathy Research Group Vienna 2014, "SAVE:
a grading protocol for clinically significant diabetic macular oedema based on
optical coherence tomography and fluorescein angiography", The British journal of
ophthalmology, vol. 98, no. 12, pp. 1612-1617.

Bresnick, G.H., Mukamel, D.B., Dickinson, J.C. & Cole, D.R. 2000, "A screening
approach to the surveillance of patients with diabetes for the presence of vision-
threatening retinopathy", Ophthalmology, vol. 107, no. 1, pp. 19-24.

Browning, D J, Glassman, A R, Aiello, L P, Bressler, N M, Bressler, S B, Danis, R P,

Davis, M D, Ferris, F L, Huang, S S, Kaiser, P K, Kollman, C, Sadda, S, Scott, I U,
Qin, H for the Diabetic Retinopathy Clinical Research Network Study Group
2008. “Optical coherence tomography measurements and analysis methods in OCT
studies of Diabetic Macular Oedema”. American Academy of Ophthalmology, 115
(8). Pg 1366 - 1371

Buabbud, J.C., Al-latayfeh, M.M. & Sun, J.K. 2010, "Optical coherence tomography
imaging for diabetic retinopathy and macular edema", Current diabetes reports, vol.
10, no. 4, pp. 264-269.

Davis, M.D. 1992. “Diabetic Retinopathy: A clinical overview.” Diabetes Care. Vol 15.
December 1992. No 12. Pg 1844 - 1874

Delcourt, C., Massin, P. & Rosilio, M. 2009, "Epidemiology of diabetic retinopathy:

expected vs reported prevalence of cases in the French population", Diabetes &
metabolism, vol. 35, no. 6, pp. 431-438.

Facey, K, Cummins E, Macpherson K, Morris A, Reay L and Slattery J 2002.

“Organisation of services for diabetic retinopathy screening”. Health technology
Assessment Report 1. Health Technology Board for Scotland (HTBS). Glasgow.

Fletcher, E and Chong, V.2008. “Diabetic macular oedema – is micropulse laser

treatment the way forward?” Ophthalmology International. Vol 3. No 1. Pg 19 – 22.
36
Gella, L., Raman, R., Rani, P.K. & Sharma, T. 2014, "Spectral domain optical
coherence tomography characteristics in diabetic retinopathy", Oman journal of
ophthalmology, vol. 7, no. 3, pp. 126-129.

Gibson, J.M. 2014, "25th RCOphth Congress, President's Session paper: 25 years of
progress in medical retina", Eye (London, England), vol. 28, no. 9, pp. 1041-1052.

Goebel, W. & Franke, R. 2006, "Retinal thickness in diabetic retinopathy: comparison of

optical coherence tomography, the retinal thickness analyzer, and fundus
photography", Retina (Philadelphia, Pa.), vol. 26, no. 1, pp. 49-57.

International Diabetes Federation (IDF) 2013. IDF Diabetes Atlas.:6th Edition. Brussels,
Belgium.

International diabetes Federation (IDF) 2014. Update to Diabetes Atlas: 6th edition.
Www. idf.org/diabetesatlas/update-2014. Accessed 31/03/15.

Jaffe, G and Caprioli, J. 2004. “OCT to detect and manage retinal disease and
glaucoma”. American Journal ofOphthalmology. January2004. Vol 137. Issue 1. Pg
156 -169

Maalej, A., Turki, W., Hadj Alouane, B., Rannen, R., Laabidi, H. & Gabsi, S. 2009,
"[Prognosis factors in diabetic macular edema: an OCT study]", Journal français
d'ophtalmologie, vol. 32, no. 2, pp. 117-125.

Mackenzie, S., Schmermer, C., Charnley, A., Sim, D., Vikas, T., Dumskyj, M., Nussey,
S. & Egan, C. 2011, "SDOCT imaging to identify macular pathology in patients
diagnosed with diabetic maculopathy by a digital photographic retinal screening
programme", PloS one, vol. 6, no. 5, pp. e14811.

Manjunath, V., Papastavrou, V., Steel, D.H., Menon, G., Taylor, R., Peto, T. & Talks, J.
2015, "Wide-field imaging and OCT vs clinical evaluation of patients referred from
diabetic retinopathy screening", Eye (London, England), vol. 29, no. 3, pp. 416-423.

Nisic, F., Turkovic, S., Mavija, M., Jovanovic, N. & Alimanovic, E.H. 2014, "Correlation
Between the Findings of Optical Coherent Retinal Tomography (OCT), Stereo
Biomicroscopic Images from Fundus of an Eye and Values from Visual Acuity of
Diabetic Macular Edema", Acta informatica medica : AIM : journal of the Society for
Medical Informatics of Bosnia & Herzegovina : casopis Drustva za medicinsku
informatiku BiH, vol. 22, no. 4, pp. 232-236.

Olson, J.A., Strachan, F.M., Hipwell, J.H., Goatman, K.A., McHardy, K.C., Forrester,
J.V. & Sharp, P.F. 2003, "A comparative evaluation of digital imaging, retinal
photography and optometrist examination in screening for diabetic retinopathy",
Diabetic medicine : a journal of the British Diabetic Association, vol. 20, no. 7, pp.
528-534.

Olson, J., Sharp, P., Goatman, K., Prescott, G., Scotland, G., Fleming, A., Philip, S.,
Santiago, C., Borooah, S., Broadbent, D., Chong, V., Dodson, P., Harding, S.,
Leese, G., Styles, C., Swa, K. & Wharton, H. 2013, "Improving the economic value
of photographic screening for optical coherence tomography-detectable macular
oedema: a prospective, multicentre, UK study", Health technology assessment
(Winchester, England), vol. 17, no. 51, pp. 1-142.
37
Puzyeyeva, O., Lam, W.C., Flanagan, J.G., Brent, M.H., Devenyi, R.G., Mandelcorn,
M.S., Wong, T. & Hudson, C. 2011, "High-resolution optical coherence tomography
retinal imaging: a case series illustrating potential and limitations", Journal of
ophthalmology, vol. 2011, pp. 764183.

Riberio, M L., Figueira, J & Cunha-Vaz, J. 2011. “Management of Diabetic

Retinopathy”. Ophthalmology International, Summer 2011. Vol. 6, No. 2, pp 38 -
43.

Sakata, L.M., Deleon-Ortega, J., Sakata, V. & Girkin, C.A. 2009, "Optical coherence
tomography of the retina and optic nerve - a review", Clinical & experimental
ophthalmology, vol. 37, no. 1, pp. 90-99.

Schimel A M, Fisher Y L and Flynn H W Jnr. 2011. “Optical Coherence Tomography in

the diagnosis and management of Diabetic Macular Edema: Time-Domain versus
Spectral Domain”. Ophthalmic Surgery,Lasers and Imaging. Vol 42. no 4 (suppl),
SPg 41- 55

Scottish Diabetes Survey 2013. Scottish Diabetes Survey Monitory Group. NHS
Scotland. Edinburgh.

Scottish Index of Multiple Deprivation(SIMD). 2012. The Scottish Government.

www.gov.scot/Topic/Statistics/SIMD/SIMDInteractive

Virgili, G., Menchini, F., Casazza, G., Hogg, R., Das, R.R., Wang, X. & Michelessi, M.
2015, "Optical coherence tomography (OCT) for detection of macular oedema in
patients with diabetic retinopathy", The Cochrane database of systematic reviews,
vol. 1, pp. CD008081.

Vujosevic, S., Benetti, E., Massignan, F., Pilotto, E., Varano, M., Cavarzeran, F.,
Avogaro, A. & Midena, E. 2009, "Screening for diabetic retinopathy: 1 and 3
nonmydriatic 45-degree digital fundus photographs vs 7 standard early treatment
diabetic retinopathy study fields", American Journal of Ophthalmology, vol. 148, no.
1, pp. 111-118.

38
Appendix 1

OCT 1st review and Outcome crosstabulation

Table 9: Showing OCT 1st Review Outcome

Cumulative
Frequency Percent Valid Percent Percent

Valid 440 60.7 60.7 60.7

6/12 OCT 1 .1 .1 60.8

await OCT appointment 35 4.8 4.8 65.7

await OCT appt 10 1.4 1.4 67.0

CANC 2 .3 .3 67.3

CANC (pt unwell) 1 .1 .1 67.4

DNA 12 1.7 1.7 69.1

Drs 12/12 1 .1 .1 69.2

DRS 12/12 42 5.8 5.8 75.0

DRS 6/12 74 10.2 10.2 85.2

OCT 6/12 59 8.1 8.1 93.4

Ref EC 48 6.6 6.6 100.0

Total 725 100.0 100.0

39
OCT outcome and Ethnicity crosstabulation

Table 10a: Showing Ethnicity and OCT_outcome Crosstabulation

OCT_outcome

DRS 12/12 DRS 12/12 S/L DRS 6/12 EC Ref OCT 6/12

Ethnicity Arabian 1 0 0 0 0

Asian 4 0 3 3 21

Bangladeshi 1 0 2 2 5

Black African 1 0 0 0 2

Black Caribbean 1 0 0 0 1

Chinese 3 1 1 1 8

Indian 13 0 15 2 25

Mixed Origin 3 0 4 1 3

Other White British 2 0 1 2 5

Other 1 0 1 0 3

Other Black 0 0 1 0 0

Other White 6 0 5 2 17

Pakistani 7 0 9 9 13

Punjabi 0 0 0 0 2

White 1 0 2 2 0

White British 25 0 16 10 37

White Irish 1 0 0 0 1

White Scottish 115 0 70 62 152

Total 185 1 130 96 295

40
Table 10b: Showing Ethnicity and OCT outcome Crosstabulation

OCT_outcome

OCT 6/12 DNA x2 Ref EC total

Ethnicity Arabian 0 0 1

Asian 0 0 31

Bangladeshi 0 0 10

Black African 0 0 3

Black Caribbean 0 0 2

Chinese 0 0 14

Indian 0 0 55

Mixed Origin 0 0 11

Other White British 0 0 10

Other 0 0 5

Other Black 0 0 1

Other White 0 1 31

Pakistani 0 0 38

Punjabi 0 0 2

White 0 0 5

White British 0 0 88

White Irish 0 0 2

White Scottish 1 0 400

Total 1 1 709

41
OCT Outcome distribution across SIMD category

Fig 16.
The OCT outcomes have the following reviews – DRS (Diabetic Retinal Screening) 6 or
12 month review, EC (Eye Clinic) and OCT (OCT clinic within DRS) 6 month review.

Table 11a: SIMD category and OCT outcome Crosstabulation

OCT_outcome

OCT 6/12 DNA

DRS 12/12 DRS 12/12 S/L DRS 6/12 EC Ref OCT 6/12 x2

SIMD 1 25 0 15 13 31 1

2 51 0 40 32 86 0

3 61 1 42 28 101 0

4 40 0 31 23 67 0

5 14 0 5 2 15 0

Total 191 1 133 98 300 1

42
Table 11b: SIMD category and OCT outcome Crosstabulation continued

OCT_outcome

Ref EC Total

SIMD 1 0 85

2 1 210

3 0 233

4 0 161

5 0 36

Total 1 725

OCT Outcome and Retinopathy Referal grade crosstabulations

Table 12a: Retinopathy Referral Grade and OCT outcome Crosstabulation

OCT_outcome

DRS 12/12 DRS 12/12 S/L DRS 6/12 EC Ref OCT 6/12

Referral_Grade R0 M2 1 0 0 0 0

R1 M0 0 0 0 0 1

R1 M2 184 1 131 89 286

R2 M2 6 0 2 8 12

R3 M2 0 0 0 1 1

Total 191 1 133 98 300

43
12b: Retinopathy Referral Grade and OCT outcome Crosstabulation continued

Count

OCT_outcome

OCT 6/12 DNA x2 Ref EC

Referral_Grade R0 M2 0 0 1

R1 M0 0 0 1

R1 M2 1 1 693

R2 M2 0 0 28

R3 M2 0 0 2

Total 1 1 725

OCT outcome and Gender crosstab

Table 13a: Gender and OCT outcome Crosstabulation

OCT_outcome

DRS 12/12 DRS 12/12 S/L DRS 6/12 EC Ref OCT 6/12

Gender Female 83 1 60 38 128

Male 108 0 73 60 172

Total 191 1 133 98 300

Table 13b: Gender and OCT outcome Crosstabulation continued

OCT_outcome

OCT 6/12 DNA x2 Ref EC

Gender Female 0 1 311

Male 1 0 414

Total 1 1 725

44
Appendix 2: Postcode area and Diabetes type crosstabulation

Table 14: Post code area and Diabetes Type 1 or 2 Crosstabulation

Diab_Type_1_or_2

Type 1 Type 2 Total

Post_code_area G14 0 1 1

G20 0 1 1

G3 0 2 2

G4 1 0 1

G40 0 3 3

G41 11 76 87

G42 16 43 59

G43 4 32 36

G44 16 51 67

G45 7 16 23

G46 12 35 47

G5 1 7 8

G51 9 50 59

G52 14 71 85

G53 11 45 56

G62 0 1 1

G72 8 29 37

G73 16 42 58

G74 0 2 2

G75 0 1 1

G76 16 24 40

G77 4 32 36

G78 1 1 2

45
 PA1 2 0 2

PA2 0 3 3

PA4 2 2 4

PA8 1 3 4

Total 152 573 725

Table 15: Chi-Square test for postcode area and Diabetes Type

Value df Asymp. Sig. (2-sided)

Pearson Chi-Square 42.391a 26 .022

Likelihood Ratio 42.424 26 .022

N of Valid Cases 725

Ethnicity and Clinic crosstabulation.

Table 16: Chi-Square Test for Ethnicity and Clinic Distribution

Value df Asymp. Sig. (2-sided)

Pearson Chi-Square 38.383a 17 .002

Likelihood Ratio 46.577 17 .000

Linear-by-Linear Association 16.191 1 .000

N of Valid Cases 709

a. 20 cells (55.6%) have expected count less than 5. The minimum expected count is .31.

46
SIMD and Clinic Crosstabulation

Table 17: Showing Chi-Square Test for SIMD categories and Clinic

Value df Asymp. Sig. (2-sided)

Pearson Chi-Square 243.961a 4 .000

Likelihood Ratio 258.405 4 .000

Linear-by-Linear Association 146.189 1 .000

N of Valid Cases 725

a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 11.37.

Table 18: Showing SIMD categories and Clinic Crosstabulation

Clinic

1 2 Total

SIMD 1 32 53 85

2 75 135 210

3 215 18 233

4 148 13 161

5 26 10 36

Total 496 229 725

47
SIMD and Gender crosstabulation

Table 19: Showing SIMD categories and Gender categories

Crosstabulation

Gender

Female Male Total

SIMD 1 24 61 85

2 111 99 210

3 87 146 233

4 75 86 161

5 14 22 36

Total 311 414 725

Table 20: Chi-Square Test for SIMD and Clinic

Value df Asymp. Sig. (2-sided)

Pearson Chi-Square 20.032a 4 .000

Likelihood Ratio 20.310 4 .000

Linear-by-Linear Association .246 1 .620

N of Valid Cases 725

a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 15.44.

48