Retinal photography for diabetic retinopathy screening in Indigenous primary health care: the Inala experience


Correspondence to:
Geoffrey Spurling, Discipline of General Practice, University of Queensland, Level 8, Health Sciences Building, Building 16/ 910, Royal Brisbane & Women's Hospital, Herston, Queesland 4029. Fax: (07) 3365 5130; e-mail:


Objective: We aimed to determine the impact of clinic based retinal photography on access to appropriate screening for diabetic retinopathy (DR).

Design, setting and participants: We opportunistically recruited patients undergoing their annual diabetic cycle of care over a two year period in the urban Indigenous primary health care clinic. Data were collected on retinal outcomes, health variables and referral patterns.

Main outcome measures: Access to appropriate screening and ophthalmic follow up, prevalence of DR, acceptability and feasibility of clinic-based retinal photography were the main outcome measures of this study.

Results: One hundred and thirty-two of a possible 147 patients consented to participate. 30% of participants had DR. Appropriate screening and ophthalmic follow up increased six fold, from 20 to 124 participants, following the introduction of the retinal camera. Most participants felt very positive about DR screening.

Conclusions: Primary care DR screening using retinal photography can improve access to DR screening for indigenous patients, reduce the burden on busy outpatient departments and should reduce visual loss. Policy-makers could contribute to screening sustainability by funding a medicare item-number for primary care based DR screening associated with the annual diabetic cycle of care. An upfront Practice Incentive Program (PIP) payment could offset set up costs.

While tight blood glucose control appears to be the holy grail of diabetes management, screening with appropriate and timely treatment is key to preventing visual loss from diabetic retinopathy (DR). DR is the leading cause of preventable blindness in Australians younger than 60 years, mainly from the development of diabetic macular oedema (DME) and the sequelae of advanced proliferative DR.1 However, the detection and timely treatment of DR with laser photocoagulation can prevent nearly all cases of severe vision loss and blindness.2 3 Thus, the 2008 National Health and Medical Research Council (NHMRC) guidelines for DR management recommend regular ocular review of patients with diabetes.1 However, many people with diabetes do not achieve this,4 demonstrating the need to develop systematic approaches to DR screening that will improve access and minimise preventable blindness by detecting DR before any risk of visual loss from more advanced retinopathy signs, particularly early DME.

Indigenous Australians have between two and four times the rate of diabetes compared with non-Indigenous people,5 and their access to health services is known to be problematic owing to affordability, transport, racism, cultural and language barriers.5 6 The recently released National Indigenous Eye Health Survey reports that of Indigenous people with diabetes, 36% have diabetic eye disease but only 20% of these have had a recent eye examination and only 37% had received the laser surgery they need.7 To improve access for our Indigenous patients with diabetes and to achieve the recommended yearly retinal screening,1 we introduced a program of retinal photography in our urban Indigenous primary health care service. Retinal photography has been successfully used in remote Indigenous primary care8 but has not been reported in urban Indigenous primary care. Therefore, we aimed to determine the impact of this screening program on access to appropriate screening and referrals, and secondly to determine the feasibility and acceptability of this service to our patients.


The Inala Indigenous Health Service (IIHS) is a Queensland Government primary health care facility situated in south-west Brisbane. The clinic has approximately 150 adult patients with diabetes who regularly attend for their diabetes annual cycle of care visits, at which time, diabetic eye complications are discussed and referrals made. There are three public ophthalmology outpatient clinics within 20km of the clinic, a bulk-billing ophthalmologist in a nearby suburb and bulk-billing optometrists in the shopping centre across the road. Waiting time for a public hospital ophthalmologist was approximately 6–12 months following referral. For private bulk-billing ophthalmologists the waiting time was no more than 4 weeks. Optometrists were available without a waiting time or referral letter.

We purchased a Canon CR-DGi digital non-mydriatic retinal camera costing approximately $40,000. We located the camera in the IIHS general treatment area, and hung special black-out curtains to ensure a dark room for maximum eye dilatation.

We opportunistically recruited adults with diabetes attending the health service for their ADC visits over a two-year period (1 October 2007 – 30 September 2009). After obtaining informed consent from patients, suitably trained practice nurses took single photographs of each retina. Dilatation drops were only used if the photo was judged by the general practitioner (GP) to be uninterpretable.

Two of the clinic GPs completed a four hour training program through The University of Queensland Discipline of General Practice (DGP) and accreditation assessment with the Royal Australian and New Zealand College of Ophthalmologist's (RANZCO) Queensland Faculty. As previously described, GPs’ diagnostic accuracy was confirmed by two independent and masked ophthalmologists double-reading the retinal photographs from the first 100 patients.9

The GPs read the retinal photographs, and using a purposefully designed recording form, diagnosed and graded the DR (according to the NHMRC guidelines1) diagnosed any non-diabetic retinal pathology, and developed appropriate referral plans. Age, gender, HbA1c, known duration of diabetes and blood pressure were collected during the patient's ADC visit and also recorded. Referrals made to specialist ophthalmologist outpatient clinics, and the outcome of these referrals, for the year preceding and following the IIHS based retinal photographs were extracted from the patient's medical record (Practix, version 1.34, build 1.43; iSoft, Sydney, NSW). Descriptive analyses were used to determine the characteristics of our participants, referrals made, and the outcomes of these referrals. Data were entered into an Excel database and graphs constructed using the same Excel software (Microsoft Excel, 2003; Microsoft, Redmond, Wash, USA).

Qualitative data collection and analysis

Semi-structured phone interviews were undertaken with eleven participants regarding their experience of retinal photography at IIHS. 10 Clinic staff selected participants for interview – all interviewees were well known to the clinic staff and were considered to be most likely to agree to participate in an interview with an unknown researcher. A clinic nurse contacted the selected participants to request permission for one of the researchers to contact them for an interview. Naomi Hansar (NH) conducted all interviews: 10 by telephone and one face-to-face at the IIHS. Interviews took approximately 15 minutes. Notes were taken during the interviews, key quotes were recorded verbatim and detailed summaries were written immediately afterwards. DA and NH coded the summaries, and identified and classified recurrent themes. 11

The local Inala Indigenous Elders were consulted and supported the project. Queensland Health's Princess Alexandra Human Ethics Research Committee approved the study.


Over the two years 1 October 2007 to 30 September 2009, 132 of the 147 patients who had ADCs (90%) participated in this project (88% of our regular patients with diabetes). Participants in this research were comparable in age and gender to non-participants. Demographic characteristics, previous referral and attendance patterns, participants’ health variables and clinic-based retinal photograph outcomes are presented in Table 1. The 30% of participants with at least one inadequate or missing photograph after initial photography was reduced to 6% following recall for pupil dilatation and/or repeat retinal photography. Retinal photography identified DR in 30% of participants and 64% with no evidence of DR.

Table 1.  Results for participants with diabetes undergoing retinal photography 2007–9 (n=132).
  1. Notes:

  2. Age, HbA1c and duration of disease are medians with range.

  3. Systolic blood pressure is mean with standard deviation.

  4. a) Percentage of those referred to this service'.

  5. b) Some patients had both moderate/severe non-proliferative DR and macular oedema or proliferative DR.

Demographic characteristics
  Age, years52 (24–78)
  Male60 (45%)
Health variables
  HbA1c (%)7.7 (4.8–14)
  Duration of Diabetes (years)6 (1–37)
  Systolic Blood Pressure (mmHg)130 (19.9)
Referral Pattern in year before photo
  No referral made97 (74)
  Optometry referral made19 (14)
  Private Ophthalmology referral made8 (5)
  Public Ophthalmology referral made8 (10)
Attendance pattern in year before photo
  No appointment attended112 (85)
  Optometry appointment attended13 (68)a
  Private Ophthalmology appointment attended6 (75)a
  Public Ophthalmology appointment attended1 (13)a
Clinic-based retinal photograph outcomes
  Patients with initially inadequate photos40 (30%)
  Patients receiving dilating drops10 (8%)
  Patients with missing or inadequate photos at end of study period (following re-photography with or without dilatation)8 (6%)
Clinical diagnoses
  No Diabetic Retinopathy (DR) detected85 (64%)
  Any DR detected39 (30%)
  – Mild DR23 (17%)
  – Moderate/ Severe DR16 (12%)
  – Macular Oedema or Proliferative DR7 (5%)b
Laser treatment following retinal photography2 (2%)

Receipt of appropriate DR screening and access to ophthalmic follow up where indicated before and after the introduction of the clinic based DR screening is summarised in Figure 1. Appropriate ophthalmic follow up in those with moderate or severe DR, diabetic maculopathy or proliferative DR increased from four participants to 12 out of 16. Two of these received laser treatment. Overall, appropriate screening and follow up increased more than six times from 20 to 124 out of 132 participants.

Figure 1.

Access to retinal screening and appropriate ophthalmic follow up (as required) one year before and one year after the introduction of clinic-based retinal photography screening.

Participant experience of retinal photography

Of the 11 participants (six males and five females) in the interviews, 10 were very positive about the experience, considering that it improved access to appropriate eye screening. Access was improved in two ways – it was convenient and it was done in an environment where they felt comfortable and safe. DR screening done at the clinic was convenient because the clinic is local and close to home for many, the screening was done while they are at the clinic for other aspects of their diabetes care (rather than having to make and attend an appointment at another site), there were not lengthy waiting times for appointments as there are at the specialist outpatient clinic, and there was easy and free parking. One participant summarised the convenience aspect when he stated “…having the photo done here (IIHS) means that even people who work can use the service because they might come in to get their blood done, get some tests done, and they can have their eyes done at the same time…”.

Although it is difficult to tease apart the attitudes to the DR screening at the Health Service vis-à-vis the attitudes to the Health Service itself, it was clear that all interviewees felt that the Inala Indigenous Health Service was a friendly and culturally safe environment where they were known and treated with respect. They considered that it was part of their community, and attendance at the Health Service had a social element, as one said “…they make you feel at home, it's a Murri thing – being around your own mob – all our people go there…you can sit and talk, have a yarn…you go somewhere else and you're sitting on your own…” For most interviewees, these attitudes about the Health Service itself were reflected in their attitudes towards DR screening being done in the Health Service. Furthermore, most interviewees considered that all Indigenous Health Services should offer this service because it had the potential to improve access for their people and decrease the negative consequences of diabetes.

However, one interviewee was very negative about the experience and was concerned about potential negative consequences of non-specialists providing eye care. The majority of interviewees were screened around the time of the introduction of the camera into the clinic when the nurses were developing skills and confidence in retinal photography. For this interviewee, the obvious inexperience of the practice staff in retinal photography reinforced the opinion that DR screening should be done by eye specialists. However, other interviewees were happy to be “guinea pigs” while the practice staff developed their technical skills in retinal photography.


On-site retinal photography increased screening rates with appropriate ophthalmic follow up in this urban Indigenous health service six fold and was well accepted by patients in a culturally sensitive primary health care setting. Twenty-four participants (18%) had DR or missing/inadequate photographs requiring specialist review. The remaining 82% of our patients had either no or mild DR, and therefore did not require referral. Participants with mild DR will be re-screened at the IIHS in 3 to 6 months and those with no DR will be re-screened in 12 months, as per the NHMRC guidelines.1 Thus, primary care based DR screening has the potential to substantially reduce outpatient workload and expense by eliminating the need for patients with no or mild DR to attend.

This study also identified that prior to the introduction of the clinic-based screening program, nearly 75% of our patients were not being referred for DR screening. The reasons for this are unclear, but are of concern. Some of our patients may have been instructed to see the local optometrist and therefore did not require a referral, but it would appear that this advice was not acted upon. Nevertheless, it would appear that we were not systematically referring patients for DR screening. It has previously been reported that social deprivation is strongly associated with poor attendance at retinal screening.12 This, combined with other factors that impede access to health care for Indigenous people reinforces the importance of a systematic approach to screening, and to providing health services such as DR screening in convenient and culturally safe environments.

This study has highlighted some interesting insights into the process of DR screening in an urban Indigenous health service. However, these data are only from one service and 142 patients, and from only two years, limiting the generalisability of this research. There are considerable costs in establishing retinal photography in primary care, and therefore it is likely to be most relevant in settings where rates of diabetes are high and/ or access to specialist services is inadequate. This is typically Indigenous or rural health services, but regional and outer-metropolitan areas also suffer from poor access to specialist services and could therefore benefit.

Clinicians and researchers often focus on glycaemic control in patients with diabetes. The value of tight glycaemic control has recently been questioned with a number of randomised controlled trials finding a surprising lack of impact on macro and micro vascular outcomes. 13 Perhaps therefore, our efforts are best focused on ensuring access to comprehensive complications screening, particularly for DR where timely and appropriate laser treatment can prevent nearly 100% of vision loss and blindness. This is of particular importance in Indigenous Australians because of the high prevalence of diabetes, and therefore the relative frequency of DR. According to the National Indigenous Eye Health Survey 2009, rates among Indigenous Australians for non-proliferative diabetic retinopathy were 25.45%, 2.5% for proliferative retinopathy and 8% diabetic macular oedema. 7 One previous study found similar results in a general Australian population with diabetes, with 24.5% having diabetic retinopathy, 2.1% having proliferative retinopathy and 3.3% with macular oedema.14

A significant area of concern identified in our study was the high number of patients identified with DR that were referred to public ophthalmology services but did not attend. During informal discussion, some of our referred patients reported receiving letters asking them to reply if they wished to remain on the waiting list for an ophthalmology appointment. This created confusion for the patients, and most did not reply. Improved referral pathways clearly need to be identified and implemented, in consultation with the hospital outpatient departments. A possible solution to ophthalmic access problems is a special on-site clinic with a visiting ophthalmologist. The IIHS will trial and evaluate this option in the near future.

DR screening using retinal cameras is sustainable over a long period of time in the remote Indigenous primary health care context and can successfully involve aboriginal health workers as well as practice nurses in taking photographs. 8 There is no reason why it is not equally sustainable and feasible in urban Indigenous primary health care settings. Primary care DR screening using retinal photography can improve access to DR screening for Indigenous patients, and reduce the burden on busy outpatient departments and should reduce visual loss. We support recommendations made by the National Indigenous Eye Health Survey to fund retinal photography so that health services can undertake screening sustainably. 7 Policy-makers could make the uptake of this screening more sustainable by funding a medicare item-number for primary care based DR screening associated with the annual diabetic cycle of care. An upfront Practice Incentive Program (PIP) payment could offset set up costs. Research across multiple sites using a cluster randomised controlled study design for evaluating retinal photography screening with a control group and clinically relevant outcomes such as vision would provide evidence about the generalisability of this approach.