An Innovative Australian Outreach Model of Diabetic Retinopathy Screening in Remote Communities

Optimizing the OCTA layer fusion option for deep learning classification of diabetic retinopathy

The purpose of this study is to evaluate layer fusion options for deep learning classification of optical coherence tomography (OCT) angiography (OCTA) images. A convolutional neural network (CNN) end-to-end classifier was utilized to classify OCTA images from healthy control subjects and diabetic patients with no retinopathy (NoDR) and non-proliferative diabetic retinopathy (NPDR). For each eye, three en-face OCTA images were acquired from the superficial capillary plexus (SCP), deep capillary plexus (DCP), and choriocapillaris (CC) layers. The performances of the CNN classifier with individual layer inputs and multi-layer fusion architectures, including early-fusion, intermediate-fusion, and late-fusion, were quantitatively compared. For individual layer inputs, the superficial OCTA was observed to have the best performance, with 87.25% accuracy, 78.26% sensitivity, and 90.10% specificity, to differentiate control, NoDR, and NPDR. For multi-layer fusion options, the best option is the intermediate-fusion architecture, which achieved 92.65% accuracy, 87.01% sensitivity, and 94.37% specificity. To interpret the deep learning performance, the Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to identify spatial characteristics for OCTA classification. Comparative analysis indicates that the layer data fusion options can affect the performance of deep learning classification, and the intermediate-fusion approach is optimal for OCTA classification of DR.

Primary eye health services for older adults as a component of universal health coverage: a scoping review of evidence from high income countries

In pursuit of Universal Health Coverage (UHC) for eye health, countries must strengthen services for older adults, who experience the highest prevalence of eye conditions. This scoping review narratively summarised (i) primary eye health services for older adults in eleven high-income countries/territories (from government websites), and (ii) the evidence that eye health services reduced vision impairment and/or provided UHC (access, quality, equity, or financial protection) (from a systematic literature search). We identified 76 services, commonly comprehensive eye examinations ± refractive error correction. Of 102 included publications reporting UHC outcomes, there was no evidence to support vision screening in the absence of follow-up care. Included studies tended to report the UHC dimensions of access (n=70), equity (n=47), and/or quality (n=39), and rarely reported financial protection (n=5). Insufficient access for population subgroups was common; several examples of horizontal and vertical integration of eye health services within the health system were described.

Funding

This work was funded by Blind Low Vision New Zealand for Eye Health Aotearoa.

Diabetic Macular Oedema Guidelines: An Australian Perspective

The number of people living with diabetes is expected to rise to 578 million by 2030 and to 700 million by 2045, exacting a severe socioeconomic burden on healthcare systems around the globe. This is also reflected in the increasing numbers of people with ocular complications of diabetes (namely, diabetic macular oedema (DMO) and diabetic retinopathy (DR)). In one study examining the global prevalence of DR, 35% of people with diabetes had some form of DR, 7% had PDR, 7% had DMO, and 10% were affected by these vision-threatening stages. In many regions of the world (Australia included), DR is one of the top three leading causes of vision loss amongst working age adults (20-74 years). In the management of DMO, the landmark ETDRS study demonstrated that moderate visual loss, defined as doubling of the visual angle, can be reduced by 50% or more by focal/grid laser photocoagulation. However, over the last 20 years, antivascular endothelial growth factor (VEGF) and corticosteroid therapies have emerged as alternative options for the management of DMO and provided patients with choices that have higher chances of improving vision than laser alone. In Australia, since the 2008 NHMRC guidelines, there have been significant developments in both the treatment options and treatment schedules for DMO. This working group was therefore assembled to review and address the current management options available in Australia.

Considerations for Training and Workforce Development to Enhance Rural and Remote Ophthalmology Practise in Australia: A Scoping Review

Australia has one of the lowest per capita numbers of ophthalmologists among OECD countries, and they predominantly practise in metropolitan centres of the country. Increasing the size and distribution of the ophthalmology workforce is of critical importance. The objective of this review was to investigate the context of rural ophthalmology training and practise in Australia and how they relate to future ophthalmology workforce development. This scoping review was informed by Arksey and O’Malley’s framework and the methodology described by Coloqhuon et al. The search yielded 428 articles, of which 261 were screened for eligibility. Following the screening, a total of 75 articles were included in the study. Themes identified relating to rural ophthalmology training and practise included: Indigenous eye health; access and utilisation of ophthalmology-related services; service delivery models for ophthalmic care; ophthalmology workforce demographics; and ophthalmology workforce education and training for rural and remote practise. With an anticipated undersupply and maldistribution of ophthalmologists in the coming decade, efforts to improve training must focus on how to build a sizeable, fit-for-purpose workforce to address eye health needs across Australia. More research focusing on ophthalmology workforce distribution is needed to help identify evidence-based solutions for workforce maldistribution. Several strategies to better prepare the future ophthalmology workforce for rural practise were identified, including incorporating telehealth into ophthalmology training settings; collaborating with other health workers, especially optometrists and specialist nurses in eyecare delivery; and exposing trainees to more patients of Indigenous background.

Distribution and Location Stability of the Australian Ophthalmology Workforce: 2014-2019

OBJECTIVE

To investigate the ophthalmology workforce distribution and location stability using Modified Monash Model category of remoteness.

METHODS

Whole of ophthalmologist workforce analysis using Australian Health Practitioner Registration Agency (AHPRA) data. Modified Monash Model (MMM) category was mapped to postcode of primary work location over a six-year period (2014 to 2019). MMM stability was investigated using survival analysis and competing risks regression.

DESIGN

Retrospective cohort study.

SETTING

Australia.

PARTICIPANTS

Ophthalmologists registered with AHPRA.

MAIN OUTCOME MEASURES

Retention within MMM category of primary work location.

RESULTS

A total of 948 ophthalmologists were identified (767 males, 181 females). Survival estimates indicate 84% of ophthalmologists remained working in MMM1, while 79% of ophthalmologists working in MMM2-MMM7remained in these regions during the six-year period.

CONCLUSION

The Australian ophthalmology workforce shows a high level of location stability and is concentrated in metropolitan areas of Australia. Investment in policy initiatives designed to train, recruit and retain ophthalmologists in regional, rural and remote areas is needed to improve workforce distribution outside of metropolitan areas.

Evaluation of the initial implementation of a nationwide diabetic retinopathy screening programme in primary care: a multimethod study

OBJECTIVES

The Australian Government funded a nationwide diabetic retinopathy screening programme to improve visual outcomes for people with diabetes. This study examined the benefits and barriers of the programme, image interpretation pathways and assessed the characteristics of people who had their fundus photos graded by a telereading service which was available as a part of the programme.

DESIGN

Multimethod: survey and retrospective review of referral forms.

SETTING

Twenty-two primary healthcare facilities from urban, regional, rural and remote areas of Australia, and one telereading service operated by a referral-only eye clinic in metropolitan Sydney, Australia.

PARTICIPANTS

Twenty-seven primary healthcare workers out of 110 contacted completed a survey, and 145 patient referrals were reviewed.

RESULTS

Manifest qualitative content analysis showed that primary healthcare workers reported that the benefits of the screening programme included improved patient outcomes and increased awareness and knowledge of diabetic retinopathy. Barriers related to staffing issues and limited referral pathways. Image grading was performed by a variety of primary healthcare workers, with one responder indicating the utilisation of a diabetic retinopathy reading service. Of the people with fundus photos graded by the reading service, 26.2% were reported to have diabetes. Overall, 12.3% of eyes were diagnosed with diabetic retinopathy. Photo quality was rated as excellent in 46.2% of photos. Referral to an optometrist for diabetic retinopathy was recommended in 4.1% of cases, and to an ophthalmologist in 6.9% of cases.

CONCLUSIONS

This nationwide diabetic retinopathy screening programme was perceived to increase access to diabetic retinopathy screening in regional, rural and remote areas of Australia. The telereading service has diagnosed diabetic retinopathy and other ocular pathologies in images it has received. Key barriers, such as access to ophthalmologists and optometrists, must be overcome to improve visual outcomes.

A Review of Advancements and Evidence Gaps in Diabetic Retinopathy Screening Models

Diabetic retinopathy (DR) is a microvascular complication of diabetes with a prevalence of ~35%, and is one of the leading causes of visual impairment in people of working age in most developed countries. The earliest stage of DR, non-proliferative DR (NPDR), may progress to sight-threatening DR (STDR). Thus, early detection of DR and active regular screening of patients with diabetes are necessary for earlier intervention to prevent sight loss. While some countries offer systematic DR screening, most nations are reliant on opportunistic screening or do not offer any screening owing to limited healthcare resources and infrastructure. Currently, retinal imaging approaches for DR screening include those with and without mydriasis, imaging in single or multiple fields, and the use of conventional or ultra-wide-field imaging. Advances in telescreening and automated detection facilitate screening in previously hard-to-reach communities. Despite the heterogeneity in approaches to fit local needs, an evidence base must be created for each model to inform practice. In this review, we appraise different aspects of DR screening, including technological advances, identify evidence gaps, and propose several studies to improve DR screening globally, with a view to identifying patients with moderate-to-severe NPDR who would benefit if a convenient treatment option to delay progression to STDR became available.

Transfer Learning for Automated OCTA Detection of Diabetic Retinopathy

Purpose

To test the feasibility of using deep learning for optical coherence tomography angiography (OCTA) detection of diabetic retinopathy.

Methods

A deep-learning convolutional neural network (CNN) architecture, VGG16, was employed for this study. A transfer learning process was implemented to retrain the CNN for robust OCTA classification. One dataset, consisting of images of 32 healthy eyes, 75 eyes with diabetic retinopathy (DR), and 24 eyes with diabetes but no DR (NoDR), was used for training and cross-validation. A second dataset consisting of 20 NoDR and 26 DR eyes was used for external validation. To demonstrate the feasibility of using artificial intelligence (AI) screening of DR in clinical environments, the CNN was incorporated into a graphical user interface (GUI) platform.

Results

With the last nine layers retrained, the CNN architecture achieved the best performance for automated OCTA classification. The cross-validation accuracy of the retrained classifier for differentiating among healthy, NoDR, and DR eyes was 87.27%, with 83.76% sensitivity and 90.82% specificity. The AUC metrics for binary classification of healthy, NoDR, and DR eyes were 0.97, 0.98, and 0.97, respectively. The GUI platform enabled easy validation of the method for AI screening of DR in a clinical environment.

Conclusions

With a transfer learning process for retraining, a CNN can be used for robust OCTA classification of healthy, NoDR, and DR eyes. The AI-based OCTA classification platform may provide a practical solution to reducing the burden of experienced ophthalmologists with regard to mass screening of DR patients.

Translational Relevance

Deep-learning-based OCTA classification can alleviate the need for manual graders and improve DR screening efficiency.

Nurse-led diabetic retinopathy screening: a pilot study to evaluate a new approach to vision care for Canadian Aboriginal peoples

Diabetic retinopathy is the most common cause of new cases of blindness and is pandemic among Aboriginal people around the world. To reduce health inequities, accessible vision screening among these high-risk populations is essential. To assess cardio-metabolic co-morbidities associated with type 2 diabetes and the use of a portable fundus camera as a novel approach for convenient, earlier and more accessible vision screening for Aboriginal peoples living with type 2 diabetes in northern and remote Canadian communities. This quantitative pilot study screened participants diagnosed with type 2 diabetes for commonly associated cardio-metabolic co-morbidities using anthropometrical measurements, blood pressure and a A1c (HbA1c) blood glucose test, followed by vision exams conducted first by a trained nurse and then by an ophthalmologist to screen for signs of retinopathy using fundus photography. Large numbers of the participants presented with overweight/obese (84.8%), pre-hypertension/hypertension (69.7%) and an elevated A1C (78.8%). Inter-rater reliability demonstrated substantial agreement between vision exam judgements made by the nurse and ophthalmologist (k = .67). Nurse-led vision screening in remote or northern communities can improve the standard of care by extending access to health services, lowering the costs to families by reducing travel expenses and preventing vision loss in a family member.

Diabetic Retinopathy in the Asia-Pacific

Diabetic retinopathy (DR), the most common complication of diabetes mellitus, is the leading cause of new cases of blindness in middle-aged and elderly in the Asia-Pacific. It has been estimated that 51% of all those with blindness due to DR globally (n = 424,400) and 56% of those with visual impairment due to DR (2.1 million) come from the Asia-Pacific. Prevalence of DR among those with diabetes ranged from 10% in India to 43% in Indonesia within the Asia-Pacific. Awareness of DR among persons with diabetes ranged from 28% to 84%. Most common modifiable risk factors for DR in the Asia-Pacific were hyperglycemia, blood pressure, dyslipidemia, and obesity. Implementation of systematic screening programs for DR and advancement in telemedicine screening methods have increased patient coverage and cost-effectiveness, though there are still numerous factors impeding screening uptake in the low-middle income regions of the Asia-Pacific. Management and treatment of DR in the Asia-Pacific is mainly limited to traditional laser retinopexy, but it is suboptimal despite new clinical approaches such as use of intravitreal anti.vascular endothelial growth factor and steroids due to limited resources. Further research and data are required to structure a more cost-effective public healthcare program and more awareness-building initiatives to increase the effectiveness of DR screening programs.

What do patients with diabetes and providers think of an innovative Australian model of remote diabetic retinopathy screening? A qualitative study

BACKGROUND

Diabetic retinopathy (DR) is the commonest cause of preventable blindness in working age populations, but up to 98% of visual loss secondary to DR can be prevented with early detection and treatment. In 2012, an innovative outreach DR screening model was implemented in remote communities in a state of Australia. The aim of this study was to explore the acceptability of this unique DR screening model to patients, health professionals and other key stakeholders.

METHODS

This descriptive qualitative study used semi-structured interviews with patients opportunistically recruited whilst attending DR screening, and purposefully selected health care professionals either working within or impacted by the programme. Interviews were audiotaped, transcribed and analysed using NVIVO. An iterative process of thematic analysis was used following the principles of grounded theory.

RESULTS

Interviews were conducted with fourteen patients with diabetes living in three remote communities and nine health professionals or key stakeholders. Nine key themes emerged during interviews with health professionals, key stakeholders and patients: i) improved patient access to DR screening; ii) efficiency, financial implications and sustainability; iii) quality and safety; iv) multi-disciplinary diabetes care; v) training and education; vi) operational elements of service delivery; vii) communication, information sharing and linkages; viii) coordination and integration of the service and ix) suggested improvements to service delivery.

CONCLUSIONS

The Remote Outreach DR Screening Service is highly acceptable to patients and health professionals. Challenges have primarily been encountered in communication and coordination of the service and further development in these areas could improve the programme's impact and sustainability in remote communities. The service is applicable to other remote communities nationally and potentially internationally.

Operational Components of Telemedicine Programs for Diabetic Retinopathy

Diabetic retinopathy is a leading cause of new-onset vision loss worldwide. Treatments supported by large clinical trials are effective in preserving vision, but many persons do not receive timely diagnosis and treatment of diabetic retinopathy, which is typically asymptomatic when most treatable. Telemedicine evaluation to identify diabetic retinopathy has the potential to improve access to care and improve outcomes, but incomplete implementation of published standards creates a risk to program utility and sustainability. In a prior article, we reviewed the literature regarding the impact of imaging device, number and size of retinal images, pupil dilation, type of image grader, and diagnostic accuracy on telemedicine assessment for diabetic retinopathy. This article reviews the literature regarding the impact of automated image grading, cost effectiveness, program standards, and quality assurance (QA) on telemedicine assessment of diabetic retinopathy. Telemedicine assessment of diabetic retinopathy has the potential to preserve vision, but greater attention to development and implementation of standards is needed to better realize its potential.

Clinical Components of Telemedicine Programs for Diabetic Retinopathy

Diabetic retinopathy is a leading cause of new-onset vision loss worldwide. Treatments supported by large clinical trials are effective in preserving vision, but many persons do not receive timely diagnosis and treatment of diabetic retinopathy, which is typically asymptomatic when most treatable. Telemedicine evaluation to identify diabetic retinopathy has the potential to improve access to care, but there are no universal standards regarding camera choice or protocol for ocular telemedicine. We review the literature regarding the impact of imaging device, number and size of retinal images, pupil dilation, type of image grader, and diagnostic accuracy on telemedicine assessment for diabetic retinopathy. Telemedicine assessment of diabetic retinopathy has the potential to preserve vision, but further development of telemedicine specific technology and standardization of operations are needed to better realize its potential.