Diagnostic Test Accuracy of Diabetic Retinopathy Screening by Physician Graders Using a Hand Held Non-Mydriatic Retinal Camera at a Tertiary Level Medical Clinic in a Lower-Middle Income Country Setting

doi:10.21203/rs.2.77/v1

Download PDF

Research Article

Diagnostic Test Accuracy of Diabetic Retinopathy Screening by Physician Graders Using a Hand Held Non-Mydriatic Retinal Camera at a Tertiary Level Medical Clinic in a Lower-Middle Income Country Setting

https://doi.org/10.21203/rs.2.77/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 06 Apr, 2019

Read the published version in BMC Ophthalmology →

You are reading this older preprint version

Read the latest preprint version →

Background

We identified that there was less evidence on diagnostic test accuracy (DTA) of diabetic retinopathy (DR) screening photographic studies that have used non-ophthalmologist human resources in low and middle-income country (LMIC) settings. This study is the first to assess the DTA of physician graders using hand held digital imaging in Sri Lanka. We aimed to assess DTA of DR screening using a nonmydriatic hand-held digital camera by trained general physicians in a non-ophthalmic setting. This modality will be useful for adaptation in similar settings.

Methods

This study is a validation of a screening intervention. We selected people with diabetes (PwDM) not previously screened or treated for DR, presenting at a tertiary medical clinic in Sri Lanka. Two-field retinal imaging was used to capture fundus images before and after pupil dilatation. The images were captured and graded by two trained, masked independent physician graders. The DTA of different levels of DR was assessed comparing physician’s grading with a retinologist’s clinical examination according to a locally adopted guideline.

Results

Seven hundred eligible PwDM were screened by physician graders. Their mean age was 60.8 years (SD ±10.08) and mean duration of DM was 9.9 years (SD ±8.09). Ungradable image proportion in non-mydriatic imaging was 43.4% (either eye-31.3%, both eyes 12.1%). This decreased to 12.8% (either eye-11.6%, both eyes-1.2%) following pupil dilatation. In comparison to detection of any level of DR, a referable level DR (moderate non-proliferative DR and levels above) showed a higher level of DTA. The sensitivity of the defined referable DR was 88.7% (95% CI 81.7-93.8%) for grader 1 (positive predictive value [PPV] 59.1%) and 92.5% (95% CI 86.4-96.5%) for grader 2 (PPV 68%), using mydriatic imaging, after including ungradable images as screen positives. The specificity was 94.9% (95% CI 93.6-96.0%) for grader 1 (negative predictive value [NPV] 99%) and 96.4% (95% CI 95.3-97.3%) for grader 2 (NPV 99.4%).

Conclusions

The Physicians grading of images from a digital hand-held nonmydriatic camera was effective in identifying referable DR, following pupil dilatation. This could be a feasible alternative modality to the existing opportunistic screening to improve the access and coverage.

Internal Medicine Specialties

Diabetes

Diabetic retinopathy

Diagnostic accuracy

Digital imaging

Screening

Diabetic retinopathy (DR) is a common complication of diabetes mellitus (DM), leading to sight loss if not detected and treated in time [1]. The International Diabetes Federation (IDF) estimated that cases of DM will increase to 629 million by 2045, with a significant burden (80%) in low and middle income countries (LMIC) [2]. Systematic DR screening (DRS) is a challenge in many of the LMICs due to limited resources [3]. The St Vincent declaration stated that all nations should make efforts to reduce DM related complications, including DR blindness [4]. It is unlikely that LMICs would be able to achieve this aim unless they adopt existing technologies.

The most common method of detecting DR in LMICs is direct ophthalmoscopy which has a low sensitivity and specificity even after rigorous training [5]. The mydriatic bio-microscopic examination by an ophthalmologist is not practical in these countries due to the low number of ophthalmologists and eye clinics which are already over-subscribed with more common blinding conditions such as cataract [6]. In these circumstances, DR is detected through opportunistic case detection only. Insufficient capacity and lack of screening infrastructure hampers efforts to implement DRS programs (DRSP) in these settings, and there is a lack of evidence of what works in LMICs [7–9].

Different models of DRS have been implemented in many parts of the world. In resource poor LMICs development of a DRS model is complex [10]. There are also poor recording systems to identify the people with DM (PwDM). Therefore, a comprehensive population based DRSP may not be feasible in LMICs in the near future. DRS also requires appropriate integration into routine care for sustainability [10]. One feasible model of systematic DRS in LMICs could be screening of PwDM when they attend for routine medical care. This can provide a participant list of PwDM, who can be offered screening at regular intervals. Integrated DRS at medical care clinics can also facilitate risk stratification and prioritisation of referrals to busy eye clinics. In these circumstances, a key consideration would be the availability of skilled human resources facilitating task shifting and sharing and efficient, cost effective and valid technology for DRS.

Digital retinal imaging is the most common DRS method used globally. Conventional desk-top digital cameras require significant physical space, skilled photographers and large image storage devices which incur high capital investment but are cost effective [11]. Hand held digital cameras are portable, require less space, minimum power consumption and less skills and training [12]. Nonmydriatic cameras which do not require pupil dilatation are more convenient for both patients and providers [13]. However, this may have an impact on image gradability and screening coverage [14]. To the authors’ knowledge there is no published data on this from Sri Lanka. This study aims to demonstrate the functional and technical feasibility of using a hand held nonmydriatic digital camera in a LMIC non-ophthalmic setting. We assess the DTA of DR detection by general physicians using this method compared to the local clinical reference standard of mydriatic indirect ophthalmoscopy and bio-microscopic examination by a retinologist.

Ethics approval was obtained from the ethics review committees of the London School of Hygiene & Tropical Medicine-United Kingdom and the National Eye Hospital-Sri Lanka. This study adhered to the tenets of the ‘Declaration of Helsinki’. A prospective screening intervention validation study was conducted between May 2017 and May 2018 at a tertiary level, public sector out-patient medical clinic in the Western province of Sri Lanka. The main outcome measure was detection of signs of DR (any DR or a referable level) by physician graders using captured digital images, according to a locally adopted guideline. The protocol of this validation study has been published in Journal of Medical Internet Research (JMIR-doi:10.2196/10900) and a summary is outlined below.

Summary of the methods

Nine general physicians from a tertiary level institution underwent a competency-based training programme delivered by two retinologists. The training included the following: capturing retinal fields using a hand held fundus camera (Zeiss-Visuscout100®-Germany), identification of signs of DR using images and DR grading according to an adopted classification system based on the United Kingdom - National Screening System [15] (Additional File 1-Table 1). Guidelines were used to standardize reporting of image quality, and ungradable images were classified based on the proportion of the retina visible for grading (Additional File 1-figure 1). Physicians were tested using a set of standard images of DR and the two who reached the required level of agreement with the retinologist (k=0.8-0.9) were selected as graders in the validation study.

The sample size (n=506) was chosen in order to estimate the sensitivity within a margin of error 10% (based on 95% confidence intervals), with an expected sensitivity of 70% and prevalence of moderate NPDR among PwDM of 20%. This included an additional 25% to take account of ungradable images (i.e., <50% of the retina visible). Interim analysis was undertaken to ascertain the level of ungradable images and, to take account of a higher than expected proportion of ungradable images, the sample size was increased to 700. A consecutive sample of diagnosed PwDM (>18years) without previous DRS at an eye clinic who were able to provide written informed consent were eligible to participate. Participants were identified at a tertiary level medical clinic, in the Western province of Sri Lanka.

In the index test imaging, two-field (1st field-macula cantered, 2nd field-disc centred) (Additional File 1-figure 2), 45-degree retinal images were captured in each eye before and after pupillary dilatation, using 2% phenylephrine, following adequate mydriasis (5-6 mm) by each physician grader. During grading, the non-mydriatic images were graded first. We calculated DTA at 3 levels for the non-ophthalmic settings: i.e., 1) any DR (detection of R1, R2, R3 and R4), 2) referable DR (R2 and above) and 3) detection of referable level and maculopathy combined with a visual acuity cut off (worse eye >6/18 Snellen visual acuity) (see Additional File 1-Table 1). The graders were masked to the history and clinical examination findings and pupil status of the images. The clinical reference test entailed a detailed, dilated fundus examination by an experienced retinologist using slit-lamp bio-microscopy with a 90D lens and indirect ophthalmoscopy using a 20D lens. The 7 field ‘Early Treatment diabetic Retinopathy Study’ references test was logistically not feasible in this resource poor setting. This examination took place as soon after imaging as possible. The retinologist was masked to the clinical status and physician graders’ findings.

For quality assurance, 15% of each non-mydriatic and mydriatic image sets were evaluated by the retinologist for technique, ability to image the required field and gradability. Fifteen percent of each hundred image sets were given back to the physician graders for double grading to assess the repeatability and intra-grader agreement in the 1st and 2nd attempts of grading images. A sample of the same image sets (n=212) were graded by the retinologist to calculate inter-grader agreement.

Analysis

Data were entered in to an MS Excel-16.0 worksheet and transferred to SPSS-Version-20.0 (Armonk-NY-IBM Corp-2011) and STATA/IC-Version-14.2 (Texas-77845-USA) for analysis. DTA variables of sensitivity, specificity and predictive values and agreement analyses (kappa statistics) were calculated with 95% confidence intervals, for each grader and each pupil status compared to the reference standard using individual eyes as the unit of analysis, considering each gradable eye as a separate case. Two approaches were used in the calculations to examine the impact of ungradable images on the outcomes. i.e., by excluding the ungradable images and by including ungradable as test positive in the analysis. As ungradable images indicate a requirement for referral to an eye clinic, we analysed ungradable images as screen positives to examine the sensitivity and specificity of detecting a need for referral. This also allows comparisons with previous studies, which have used both methods.

Subgroup analysis conducted for identification of presence/absence of DR (any DR), moderate NPDR and above with / without macular signs, to make recommendations for a referable criterion for the local context. We used different referable criteria in the analysis to understand the variation in DTA to assess the most suitable and accurate cut off level of DR without overloading the ophthalmologist’s clinic and also facilitating safe practice at a non-ophthalmic setting.

Participants’ characteristics

Of the 826 eligible PwDM identified from medical clinical records, response rate was 84.7% (700/826). Mean age of the participants was 60.8 years (SD ±10.08) and majority were women (66%, 462/700). Only a third of the participants were employed (27.9%, 195/700) and 79.1% (554/700) lived in the capital city of Colombo and hailed from low income families (88%, 616/700, monthly income < £150). Of these, 98.4% (689/700) had type 2 DM and 11 (1.6%) were diagnosed with DM at age < 30 years and were on insulin. The mean age at diagnosis of DM was 50.9 years (SD ±11.03) and mean duration of diabetes was 9.9 years (SD ±8.09). Mean fasting plasma glucose with in the last 3 months was 140.4 mg/dl (SD ±55.43). Additional co-morbidities included; hypertension (70%), hyperlipidaemia (57.3%), ischaemic heart disease (31.9%), nephropathy (9%) and neuropathy (35%). The Table 1 shows the characteristics of the PwDM in this study. The maximum time interval between index and reference test was 4 weeks.

Table 1. Participants’ characteristics

Variable	Categories	Results
Mean age	Mean	60.8 years (SD 10.1)
Sex	Male	34% (n=238)
	Female	66% (n=462)
Employment status	Employed	27.9%(n=195)
	Unemployed	41.0% (n=287)
	Retired	31.1%(n=218)
Monthly household income	Low (<£150)	88.0% (n=616)
	Middle (>£150 - <£300)	9.6% (n=67)
	High (>£300)	2.4% (n=17)
Ethnic group	Sinhalese	66.9% (n=468)
	Tamil	16.4% (n=115)
	Moor	14.1%(n=99)
	Other	2.6% (n=18)
Age at diagnosis of diabetes	Mean	50.9 years (SD 11.0)
Duration of diabetes	Mean	9.9 years (SD 8.1)
Current treatment of DM	Diet only	5.6% (n=39)
	Oral medication only	79.7% (n=558)
	Insulin only	5.6% (n=39)
	Oral medication and insulin	9.1% (n=64)
Fasting glucose level (within 3 months)	Mean	140.44 (SD 55.4) 95% CI (136.2-144.0)
HbA1c level (only n=42 available)	Mean	7.9 (SD 2.2) 95% CI (7.3-8.7)
Other comorbidities	Hypertension	70%
	Hypercholesterolaemia	57.3%
	Ischaemic heart disease	31.9%
	Nephropathy	9%
	Neuropathy	35%
	Leg / peripheral ulcers	5%
Age at diagnosis of hypertension	Mean	52.8% (SD 9.6%)
Family history	Diabetes	63.3%
	Hypertension	50.4%
	Hypercholesterolaemia	30%
	Ischaemic heart diseases	28.7%

Image gradability and number of images sets available for DTA analysis

Seven hundred PwDM were included in the study and 126 (15.2%, 126/826) were excluded (n=69-no consent and n=57-did not attend for imaging) (See Figure 1). Since both physician graders captured image sets of each participant, ideally there should be 1400 image sets (by eyes) for each grader for each pupil status. However, we ended up as shown in Additional File 2 - flow chart, due to technical errors in storage and failure to track PwDM (8-20 eyes, 0.6-1.4%) at the medical clinic. Overall ungradable proportion in nonmydriatic imaging was 31.0% (217/700) for at least one eye ungradable for either grader. In 12.0% (84/700) both eyes were ungradable for both graders. This decreased to 11.4% (80/700) and 1.1% (8/700) respectively, following pupil dilatation. We noted that 9 PwDM (18 eyes) did not attend for the reference test. In addition, reference test was not possible in 40 eyes (40/1400, 2.8%, in 21 participants: 37 advanced lens opacity, 1 posterior capsular opacity, 1 phthisical eye 1 and 1 eviscerated). After excluding eyes of those who did not attend and ungradable even at the reference test (total n=58) we left with 1342 image sets (by eyes) in DTA analysis. Overall there were 1041 DR positive eyes and 301 DR negative eyes as identified at the reference test. The technical failure rates at image set and grader level (by eyes) are described in the Table 2 and Additional Files 2 and 3.

Table 2. Gradability of the images as marked by each grader and agreement with the reference grader

		Nonmydriatic imaging		Mydriatic imaging
		Grader 1 N=1392	Grader 2 N=1391	Grader 1 N=1381	Grader 2 N=1380
Gradable
	100% ^a, n (%)	286 (20.5%)	352 (25.3%)	537 (38.9%)	605 (43.8%)
	75%, n (%)	308 (22.1%)	431 (31.0%)	395 (28.6%)	519 (37.6%)
	50%, n (%)	386 (27.7%)	276 (19.8%)	351 (25.4%)	186 (13.5%)
Ungradable
	<50%, n (%)	412 (29.6%)	332 (23.9%)	98 (7.1%)	70 (5.1%)
Inter-grader Agreement ^b, kappa (95% CI)		0.90 (0.85,0.95)	0.90 (0.86,0.95)	0.72 (0.56,0.89)	0.96 (0.89,1.03)

a. percentage of visibility in a given field, by eyes

b. Physician grader vs retinologist – grading a random sample of image sets (n=212, total n=424).

DTA after including ungradable images (primary analysis)

We aimed to demonstrate the DTA for referrals to eye clinic rather than the DTA of detecting DR in the primary analysis. When considering the ungradable images as screen positives, sensitivity of detection of any level of DR using nonmydriatic imaging was 82.7% (95% CI 78.4-86.5%) in grader 1 and 78.3% (95% CI 73.7-82.5%) in grader 2. However, since they were referring a higher proportion of ungradable, probably those who did not have the disease, specificity values dropped to 70.4% (95% CI 67.6-73.1%) in grader 1 and 76.2% (95% CI 73.6-78.7%) in grader 2 in nonmydriatic imaging. In mydriatic imaging when we included the ungradable images in the analysis sensitivity was 79.3% (74.7-84.8%) in grader 1 and 78.0% (95% CI 73.4-82.2%) in grader 2. The specificity value of grader 1 was 89.2% (95% CI 87.2-90.9%) and grader 2 was 91.5% (95% CI 89.7-93.1%). The sensitivity, specificity, NPV, PPV and kappa agreement at different levels of DR after including the ungradable images are described in Table 3.

Table 3. Diagnostic test accuracy of each grader by each pupil status (unit of analysis, by eyes, after including ungradable images)

Index Test			Sensitivity (95% CI) (%)	Specificity (95% CI) (%)	PPV (95% CI) (%)	NPV (95% CI) (%)	Kappa (95% CI) (%)
Any DR grading
	Non-mydiatric image
		Grader 1	82.7 (78.5, 86.5)	70.4 (67.6, 73.1)	47.4 (43.4, 51.5)	92.7 (90.7, 94.4)	0.42 (0.38, 0.47)
		Grader 2	78.3 (73.7, 82.5)	76.2 (73.6, 78.8)	51.6 (47.3, 55.9)	91.6 (89.6, 93.3)	0.46 (0.42, 0.51)
	Mydiatric image
		Grader 1	79.3 (74.7, 83.4)	89.2 (87.2, 90.9)	70.3 (65.6, 74.8)	93.0 (91.3, 94.5)	0.66 (0.61, 0.70)
		Grader 2	78.0 (73.4, 82.2)	91.5 (89.7, 93.1)	74.7 (70.0, 79.1)	92.8 (91.1, 94.3)	0.68 (0.64, 0.73)
Referable DR grading ^a
	Non-mydiatric image
		Grader 1	86.8 (79.5, 92.3)	71.7 (69.2, 74.2)	20.4 (16.9, 24.3)	98.5 (97.6, 99.1)	0.23 (0.19, 0.28)
		Grader 2	84.9 (77.3, 90.9)	77.3 (75.0, 79.6)	23.8 (19.7, 28.3)	98.4 (97.5, 99.0)	0.29 (0.23, 0.34)
	Mydiatric image
		Grader 1	88.7 (81.7, 93.8)	94.9 (93.6, 96.0)	59.1 (51.4, 66.6)	99.0 (98.4, 99.5)	0.68 (0.61, 0.75)
		Grader 2	92.5 (86.4, 96.5)	96.4 (95.3, 97.3)	68.0 (60.2, 75.3)	99.4 (98.8, 99.7)	0.76 (0.70, 0.82)
Maculopathy grading ^b
	Non-mydiatric image
		Grader 1	89.2 (83.5, 93.5)	70.1 (67.5, 72.6)	26.5 (22.7, 30.4)	98.2 (97.2, 98.9)	0.29 (0.25, 0.34)
		Grader 2	80.4 (73.5, 86.6)	77.0 (74.6, 79.3)	29.7 (25.3, 34.3)	97.0 (95.8, 98.0)	0.33 (0.28,0.38)
	Mydiatric image
		Grader 1	86.5 (80.4, 91.4)	91.5 (89.8, 92.9)	54.9 (48.5, 61.2)	98.3 (97.4, 98.9)	0.62 (0.56,0.68)
		Grader 2	82.4 (75.8, 87.9)	95.4 (94.1, 96.5)	68.2 (61.1, 74.7)	97.0 (96.9, 98.6)	0.71 (0.65,0.77)

a-Referable level DR – moderate non-proliferative DR and above

b-Maculopathy – presence of haemorrhage/s or exudates within 2-disc diameters of the centre of fovea

DTA after excluding ungradable images

The DTA estimates were calculated after excluding ungradable images (<50% of the field visible) in the next step as an accuracy measure of the modality. In the comparison physician’s grading using 2-field imaging against the clinical reference standard, in detection of any level of DR, there was no significant difference in DTA by pupil status, in each grader. Similar results were observed in detection of macular signs. Table 4 shows the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for each grader and for each pupil status for this analysis. A higher range of PPV values were observed in detecting a referable level DR (79.7%-92.8%) (moderate non-proliferative DR and above) compared to identification of macular signs (63.2%-73.5%) (presence of haemorrhage/s or exudate/s within 2-disc diameters of centre of fovea). However, such differences were not observed in NPV.

Table 4. Diagnostic test accuracy of each grader by each pupil status (unit of analysis, by eyes, after excluding ungradable images)

Index Test			Sensitivity (95% CI) (%)	Specificity (95% CI) (%)	PPV (95% CI) (%)	NPV (95% CI) (%)	Kappa (95% CI) (%)
Any DR grading
	Non-mydiatric image
		Grader 1	71.1 (64.9, 77.4)	95.6 (94.1, 97.0)	80.8 (75.0, 86.6)	92.7 (90.8, 94.5)	0.70 (0.64, 0.76)
		Grader 2	66.4 (60.0, 72.7)	95.4 (94.0, 96.8)	78.9 (72.9, 84.9)	91.7 (89.8,93.5)	0.66 (0.60,0.72)
	Mydiatric image
		Grader 1	76.2 (71.3, 81.0)	94.0 (92.6, 95.5)	79.1 (74.4, 83.9)	93.0 (91.4, 94.6)	0.71 (0.66, 0.76)
		Grader 2	75.2 (70.2, 80.1)	93.9 (92.5, 95.4)	78.3 (73.6, 83.1)	92.9 (91.3, 94.5)	0.70 (0.65, 0.75)
Referable DR grading^a
	Non-mydiatric image
		Grader 1	73.6 (61.7, 85.5)	99.7 (99.3, 100.0)	92.9 (85.1, 100.7)	98.5 (97.7, 99.3)	0.81 (0.72, 0.90)
		Grader 2	71.7 (59.6, 83.8)	99.0 (98.4, 99.6)	79.2 (67.7, 90.7)	98.5 (97.8, 99.3)	0.74 (0.64, 0.84)
	Mydiatric image
		Grader 1	81.8 (72.5, 91.1)	99.4 (99.0, 99.9)	88.5 (80.5, 96.5)	99.0 (98.5, 99.6)	0.84 (0.77, 0.91)
		Grader 2	89.4 (82.0, 96.8)	98.8 (98.2, 99.4)	79.7 (70.6, 88.9)	99.4 (99.0, 99.9)	0.83 (0.77, 0.90)
Maculopathy grading^b
	Non-mydiatric image
		Grader 1	78.1 (68.6, 87.6)	96.6 (95.5,97.8)	65.5 (55.5, 75.5)	98.1 (97.29, 99.1)	0.69 (0.60, 0.77)
		Grader 2	64.1 (53.5, 74.8)	98.1 (97.3, 99.0)	73.5 (63.0, 84.0)	97.1 (96.1, 98.2)	0.66 (0.57, 0.75)
	Mydiatric image
		Grader 1	81.0 (73.3, 88.7)	96.0 (94.9, 97.1)	63.3 (54.9, 71.6)	98.3 (97.6, 99.1)	0.68 (0.61, 0.75)
		Grader 2	75.3 (66.8,83.7)	97.8 (96.91, 98.6)	73.8 (65.3,82.3)	97.9 (97.1, 98.7)	0.72 (0.65, 0.79)

a-Referable level DR – moderate non-proliferative DR and above

b-Maculopathy – presence of haemorrhage/s or exudates within 2 disc diameters of the centre of fovea.

Sub-analyses of DTA

As a pragmatic approach for a resources poor non-ophthalmic setting, we reported the DTA of DRS using non-mydriatic imaging and dilatation of the pupils of only those who have ungradable images (two-step process). In this sub-analysis, the eye which was ungradable even following mydriasis were considered as screen positives. We derived a sensitivity of referable level of DR 81.1% (95% CI 72.9-87.9%) for grader 1 and 82.1% (95% CI 74.0-88.6%) for grader 2. The specificity values were 95.4% (95% CI 94.2-96.5%) for grader 1 and 97.1% (95% CI 96.1-97.9%) for grader 2 in this approach. We observed an improved level of PPV (59.7-70.2%) and NPV (98.4-98.5%) in this strategy. The details are described in Additional File 4.

We combined the DTA of the detection of referable level DR (moderate NPDR and above) with positive macular signs, using nonmydriatic imaging, where the sensitivity was 79.0% for grader 1 and 70.8% for the grader 2. These estimates improved to 84.5% and 85.8% respectively for grader 1 and 2 after dilatation. For the same referable level specificity values were 96.6% and 98.0% for grader 1 and 2 respectively and there was no significant change with the pupil dilatation. (nonmydriatic grader 1-97.3%, grader 2-98.4%).

We also incorporated visual acuity threshold for referrals (considering worse eye visual acuity 6/18 and above, retinopathy moderate and above and positive macular signs) and found a sensitivity of grader 1 was 98.3% (95% CI 94.9-99.7%) and grader 2, 97.4% (95% CI 93.5-99.3%). However, in the same referable level specificity values showed an overall reduction (grader-1 49.4%, 95% CI 45.3-53.5% and grader-2 51.6%, 95% CI 47.5-55.7%), probably due to high number of PwDM referred to the next level without ≥ moderate NPDR. These approaches will be useful in making recommendations for a referable level for a nonophthalmic setting.

Agreement analysis

The percentage of image gradability agreement, between index graders and retinologist (inter-grader agreement), in non-mydriatic imaging; grader 1 was 85.2% ([kappa] k=0.9, 95% CI 0.85-0.95) and grader 2, 78.5% (k=0.9, 95% CI 0.86-0.95). In mydriatic imaging, inter-grader gradability agreement of grader 1 was 76.2% (k=0.72, 95% CI 0.56-0.89) and grader 2, 72.7% (k=0.96, 95% CI 0.89-1.03).

We proposed grading the same images by a retinologist would provide a fair comparison for the physician graders in agreement analysis. However, here the concerns were limitations in the degree of view of a hand-held retinal camera and skills of capturing images by the physicians. In this analysis, in the grading of DR and macular signs, we found that inter-grader agreement was mostly >0.8 except for the grader 2 non-mydriatic images (any DR k=0.80-0.89, referable DR k=0.77-0.85, macular signs k=0.77-0.85). We observed a satisfactory level of agreement of the physician graders findings using a 2-field modality. The overall concordance of the results is described in Table 5.

Table 5. Agreement by comparing findings of sample of same images (captured by physicians) graded by retinologist (inter-grader agreement: physician grader 1 or 2 vs retinologist) (n=212, 424 image sets)

Index Test		Sensitivity (95% CI) (%)	Specificity (95% CI) (%)	Kappa value (95% CI) (k)
Any DR grading
Nonmydriatic image
	Grader 1	92.3 (86.4, 98.2)	96.8 (94.5, 99.1)	0.89 (0.83, 0.95)
	Grader 2	84.9 (77.3, 92.5)	94.9 (92.1, 97.7)	0.80 (0.73, 0.88)
Mydriatic image
	Grader 1	90.2 (85.0, 95.5)	96.7 (94.6, 98.8)	0.88 (0.82, 0.93)
	Grader 2	90.5 (85.5, 95.6)	95.0 (92.5, 97.6)	0.85 (0.80, 0.91)
Referable DR grading ^a
Nonmydriatic image
	Grader 1	80.0 (64.3, 95.7)	99.3 (98.3, 100.3)	0.84 (0.72, 0.96)
	Grader 2	79.2 (62.3, 95.4)	98.3 (96.9, 99.8)	0.77 (0.64, 0.91)
Mydriatic image
	Grader 1	77.1 (63.2, 91.1)	98.9 (97.8, 100.0)	0.80 (0.69, 0.91)
	Grader 2	97.0 (91.1, 102.9)	97.6 (96.4, 99.2)	0.85 (0.76, 0.94)
Maculopathy grading ^b
Nonmydriatic image
	Grader 1	94.3 (86.6, 102.0)	97.0 (94.9, 99.0)	0.85 (0.76, 0.94)
	Grader 2	75.0 (61.6, 88.5)	98.2 (96.7, 99.8)	0.77 (0.66, 0.88)
Mydriatic image
	Grader 1	88.1 (79.9, 96.4)	96.5 (94.5, 98.4)	0.82 (0.74, 0.90)
	Grader 2	76.6 (66.2, 86.9)	98.5 (97.3, 99.8)	0.80 (0.72, 0.89)

a-Referable level DR – moderate non-proliferative DR and above

b-Maculopathy – presence of haemorrhage/s or exudates within 2-disc diameters of the centre of fovea

Quality assurance

The index graders re-graded the coded image sets in a masked fashion independently without having access to the 1st attempt data. In this, first attempt vs second attempt weighted kappa agreement was calculated to assess the repeatability of DR grading at level of retinopathy. The kappa value of grader 1 was 0.69 (95% CI 0.60-0.78) and grader 2, 0.66 (95% CI 0.58-0.73). In comparison of grader 1 vs grader 2, first attempt kappa was 0.82 (95% CI 0.76-0.89) and in second attempt it was 0.74 (95% CI 0.66-0.83%) (see Additional File 5).

Reasons for ungradability of images, prevalence of DR and time and flow of the participants

We described the possible reasons for ungradability, using the highest recorded ungradability values, which was observed by grader 1. Of the 29.4% of ungradable images for grader 1-nonmydriatic imaging, 69.2% (285/412) had lens opacity. Among these 29.8% (85/285) eyes had significant level of lens opacity which required cataract assessment. Following reference test 37 eyes were identified as having lens opacities that required urgent cataract surgery. Overall, 75.6% of the participants had no DR, 16.7% mild DR (R1), 3.6% moderate NPDR (R2), 0.4% severe NPDR (R3) and only 1% had PDR (R4). Among the ungradable images in nonmydriatic imaging, 66.5% (274/412) had no retinopathy (R0), 19.9%-mild NPDR (R1), 1.7%-moderate NPDR (R2), 0.7%-severe NPDR (R3) and 1.5%-proliferative DR (PDR-R4) (see Additional File 3).

The mean time gap between index imaging and reference test was 3.6 days (SD ± 0.2) (95% CI 3.2-4.0, range 0-48 days). Six hundred and ninety-two PwDM completed the reference test examination and 98% (684/692) of them underwent retinologists examination < 4 weeks period.

We demonstrated that DRS by general physicians using a mydriatic two field technique was a feasible modality to detect a defined level of DR (sensitivity 88.7-92.5% and specificity 94.9-96.4%) in a non-ophthalmic setting, considering the level of DTA achieved. This may be suitable for LMIC settings where it would be difficult to implement full population based DRSP due to resource and information constraints. Compared to a locally accepted clinical reference standard, DRS using mydriatic 2-field strategy by general physicians showed a sensitivity of >80% and specificity of >95% after including ungradable images. This is comparable to accepted DRS standards in HICs [15]. We assumed that inclusion of ungradable images in the DTA analysis is a pragmatic approach for a non-ophthalmic setting, considering the requirement of refer those to the ophthalmologist’s clinic for further assessment and treatment. The proposed imaging strategy could act as a filter minimizing the number of referrals at ophthalmologist clinic, thereby reducing the strain on the system. The United Kingdom prospective diabetes study group (UKPDS) reported that 15.3% of those with signs of DR at baseline, required laser at 3 years [16]. Therefore, identification of even minor levels of DR will be beneficial to stratify the risk groups early.

Digital retinal imaging showed promising results in DRS [15]. The digital imaging systems have the advantage of instant availability of images for quality assessment and convenient storage and retrieval. Several studies have compared digital fundus photography with 7 fields used in early treatment diabetic retinopathy study (ETDRS)[17–19] or mydriatic ophthalmoscopy [20-21] in DRS and shown an acceptable level of DTA. The DTA studies from high income countries (HIC) used trained graders or ophthalmologists/retinologists in index test and table top static cameras with advanced technology such as wider angle and high resolution, which may be prohibitively expensive for LMICs. Though DTA is lower in this study, this strategy would be useful in a context where there is no systematic DRS. On the other hand, it may be arbitrary to compare the findings of this study with HICs. LMICs such as Sri Lanka require pragmatic solutions for control of visual loss due to DR with rising prevalence of DM.

The optimum number of retinal fields in a DRS strategy is a key factor that affects accuracy. The ETDRS 7-field strategy is considered to be the gold standard for DR detection but is not practical in a screening program [22]. Previous studies showed that single retinal field is inadequate to achieve required standards [23–27]. Studies have also demonstrated that 3-fields would not improve DTA of detection of any referable DR [28]. A non-mydriatic two field strategy in detection of sight threatening DR (STDR) in a HIC showed a sensitivity of 92% (95% CI 90-94%) and specificity of 96% (95% CI 95-98%) (proportion of ungradability - nonmydriatic 15.3-17.6%, mydriatic 1.4-2.1%) [17]. In our study, sensitivity was 71.7-73.5% and specificity 98.9-99.6% for detection of referable DR using nonmydriatic imaging. We could not achieve a higher level of sensitivity comparable with the studies done in HICs, due to poor image quality. The main causes of poor image quality are dark iris colour, poor pupil dilation status and lens opacity [13,29]. In HICs prevalence of cataract is less compared to LMICs like Sri Lanka [30–32]. We observed that sensitivity increased to 81.8%-89.3% when pupils were dilated. In addition, specificity was high irrespective of the pupil status, because physician graders were confident in grading in the absence of any signs. The study by Henricsson, M. et al. showed that dilatation and increasing number of fields to 5, the DTA improved to sensitivity of 93% and specificity of 91% [33]. It is apparent that one or more fields to the two central fields in DRS, has increased DTA minimally [18]. Therefore, a 2-field DRS strategy is justifiable for this context. In addition, we used retinologist’s clinical examination as the reference which Scanlon, PH. et al. showed that there was no significant difference in the assessment of DTA between using 7F ETDRS and slit lamp examination by ophthalmologists [21].

In some settings, several non-ophthalmological personnel had been employed in DRS. In our study we proposed DRS by trained general physician at medical clinic following assessment of barriers. The estimates from previous studies are comparable with the finding of our study [29]. In a study from the United Kingdom, DRS by general practitioners using 35mm colour images shown that detecting any level of DR was increased from 62.6% (95% CI 55.9-69.4%) with direct ophthalmoscopy to 79.2% (95% CI 73.6-84.9%) using retinal photographs (and specificity remained unchanged (direct ophthalmoscopy 75.0% (95% CI 69.5-80.5%) vs 73.5% (95% CI 68.0-79.1%)) [34]. They concluded that retinal photography by trained general practitioners in primary care setting could achieve an acceptable level of detection of STDR (87%) [34]. In our validation study physician graders showed a sensitivity range of 88.6-92.4% and specificity range of 94.8-96.3% in detection of referable level of DR using mydriatic imaging which is better than the reported studies. However, this may depend on the proportion of ungradable images. In our sub-analysis we included the technical failures as test positives, since physician graders refer these to the eye clinic. A review of 22 cross sectional photographic studies showed nonmydriatic retinal photography sensitivity range of 25-66% for general practitioners, 43-79% for optometrists and 27-73% for other non-ophthalmic health professionals and an overall specificity of >91% [35]. The sensitivity of detection of any level of DR increased to 87-100% for general practitioners and >91% for optometrists with pupil dilatation [35]. As a first line, this study has shown that physician graders are capable of DRS in a non-ophthalmic setting in Sri Lanka. However, we will have to study the effectiveness of this modality in a larger number to make specific recommendations.

In our study 75.6% of the participants had no DR, 16.7% mild DR (R1), 3.6% moderate NPDR (R2), 0.4% severe NPDR (R3) and only 1% had PDR (R4). A similar prevalence pattern was observed by Ding, J. et al. in a clinical sample in China, using similar exclusion criteria. In their study 77.6% (412/531) had no DR, 15.4% background DR (BDR) (27/531) only 9 had (1.7%) had PDR [36]. The technical failure rate in this study was also similar to our current study (2 field, 28.2%, 95% CI 24.5-32.3%). The mydriatic imaging technical failure rate was 5.1%-7.1% in our study, which is comparatively lower to a study in Israel by Mizrachi, et al [37] which recorded a prevalence of any DR as 18.3%. In their study, 14.4% were poor quality images and they reported that in 81-89 years of age group poor quality images proportion was very high (34.3-36.1%) [37]. Our study sample had low prevalence of DR, despite a clinical setting with longer duration of DM. A similar outcome was observed by Bhargava, et al. in Singapore [38] which showed a prevalence of DR 17.8% and clinically significant macular oedema of 1.69%. Similar level of prevalence of DR was seen in the study done by Ku, JY. et al. in Australia [24]. In their study only 1% had severe NPDR and 6% had PDR. The low prevalence of DR could be attributed to many factors. One reason for this would be excluding those who had undergone previous DRS and treatment. In Sri Lanka about 50% of the PwDM in clinics had DRS [39]. Forty percent (572/1398) of the PwDM had previous DRS or DR treatment in our study.

Nonmydriatic imaging has lower resolution and lower image quality leading to poorer detection of DR [13,40]. However, digital imaging has lower technical failure rates than imaging using colour slides [41]. The hand held nonmydriatic camera used in this study required a minimum of 3.5 mm pupil diameter and average pupil diameter in this study population was 2.01 mm at presentation. When pupils were dilated, proportion of ungradable images was reduced from 43.4% to 12.8%. Even at the reference test 37 eyes were ungradable due to lens opacity. The improvement of image quality in people with dark irises by pupil dilatation has been demonstrated in a previous study [13]. The referral of ungradable images to an ophthalmologist’s clinic is in the best interest of patient safety. Scanlon, P. et al. showed that in the > 80 years age group the technical failure rates reduced from 41.6% to 16.9% following mydriasis [14]. This study concluded that the odds of having one eye ungradable, increased by 2.6% (95% CI 1.6-3.7%) for each extra year of life since diagnosis of DM and major cause of ungradable images was having a central cataract (57%) [14]. Therefore, a nonmydriatic strategy with dilatation of pupil for ungradable images only would be more appropriate for this context. Another reason for low DTA in nonmydriatic imaging could be low resolution, which may have led to poor visibility of delicate signs such as microaneurysms, as suggested in the study by Henriccson, et al. [33].

Limitations

We excluded PwDM with previous eye screening or treatment, which reduced the proportion of people with DR, which may have introduced spectrum bias. However, the resulting sample included a wide range of pathologies, albeit with fewer people with more advanced disease. When considering any DR as referable level, there were 301 DR positive and 1041 DR negative eyes in the analysis. There were only 69 DR positive eyes when considering moderate NPDR and above as the referable level. However, PwDM who already visited the eye clinic would not usually participate in a screening programme, therefore the sample examined in this study reflects the PwDM who would be eligible for DRS. Another limitation was high proportion of ungradable images from nonmydriatic imaging compared to other studies. However, the populations in LMICs have a higher prevalence of untreated cataracts, which would prevent adequate retinal view and would require referral. In a potential DRSP, these participants will be referred to the next level of eye care and the patient would benefit from the imaging even if the DR status remains unknown.

The most common gold standard for a reference test would be the ETDRS 7-field image grading by an expert grader. However, it was not possible in this setting for a large sample due to resource and time limitations. In addition, any misclassifications in the clinical reference test could have been mitigated, with the second reference grader. In order to have a higher precision of the DTA, the sample size should be adjusted according to the reported low prevalence of higher graded of DR such as severe NPDR and PDR.

In this study we demonstrated that the DTA of the physician graders was closer to the standard practice of national level screening programs in other settings. We conclude that 2-field retinal imaging using a hand-held digital camera at a medical clinic, by physician graders, with dilatation of pupil of those who have ungradable images, provides a valid modality to identify referable level of DR. This strategy is an accurate screening method of detection of a referable level in a health care facility based PwDM who are at risk of developing STDR.

Ethics approval and consent to participate - The ethics approval was obtained from the London School of Hygiene and Tropical Medicine - United Kingdom and from the National Eye Hospital of Sri Lanka.

Consent for publication - Authors provide consent for publication in the BMC-Ophthalmology – Open source.

Availability of data and material - Any of the data and material produced in this study will be available under a reasonable request to the corresponding author, for non-commercial purposes.

Competing interests - No conflict of interest.

Funding - This study was funded by a PhD student project grant awarded to MMPNP, by Queen Elizabeth Diamond Jubilee Trust, coordinated through the Commonwealth Eye Health Consortium - United Kingdom.

Authors’ contributions - MMPNP conceived the idea and presented the concept and protocol at PhD upgrading. GVSM is the main supervisor and most senior author of this study. MMPNP conducted the study as a major component of a PhD research degree. MMPNP involved in concept development, study design, training of physicians, conducting the study in Sri Lanka, data acquisition, analysis and interpretation and manuscript preparation. GVSM and JY supervised the overall student project. DM is the adviser in statistics for the student. MMPNP conducted the data analysis under supervision of local statistician - Dr. Suwin Hewage (SH), Adviser at LSHTM - DM and statistician MK. All authors read and approved this manuscript for publication.

Acknowledgements

Prof.Clare Gilbert, Prof.Tunde Peto and Assistant Prof..David MacLeod for their role as advisers in conducting this study and the research student project overall.

-Local Collaborator – Association of Vitreo Retina Specialists of Sri Lanka – Dr.Charith Fonseka, Dr.Aruna Fernando and Dr.Shreeharanadan.

- Dr.Kapila Banduthilaka (KB) and Dr.Mangala Dhanapala at National Eye Hospital of Sri Lanka as trainers and managing those who required treatment following screening. KB for his role as the reference test grader.

-Local Research Team – Dr.Aruna Kulathunga for supervision at medical clinic, Dr.Heshani Dissanayaka and Dr.Lalani Pathirana for their contribution as physician graders.

-Dr.Suwin Hewage (SH) for his role in data management (including retinal image files) statistician in the local institution.

-Local Research Assistants – Dr.Renuka Balasubramaniyam, Dr.Abdul Quadir, Dr.Asanka Gunathunga, Dr.Anjali Umayanagana, Dr.Chathurangi Konara, Dr.Ruwanthi Samarasinghe

-All the local communities who participated in the study, staff of retinal clinic and medical clinic.

1 Klein BE. Overview of Epidemiologic Studies of Diabetic Retinopathy. Ophthalmic Epidemiol. 2007;14(04):179–83. PMID: 17896294. DOI:10.1080/09286580701396720

2. IDF. International Diabetes Federation Diabetes Atlas - Eighth Edition - 2017. http://www.diabetesatlas.org/resources/2017-atlas.html. Accessed 10 June 2018.

3. Lin S, Ramulu P, Lamoureux EL, Sabanayagam C. Addressing risk factors, screening, and preventative treatment for diabetic retinopathy in developing countries: A review. Clin Exp Ophthalmol. 2016;44(4):300–20. PMID: 26991970. DOI: 10.1111/ceo.12745

4. The Saint Vincent Declaration. Acta Ophthalmol Scand. 1997;75(S223):63. http://onlinelibrary.wiley.com/doi/10.1111/j.1600-0420.1997.tb00440.x/pdf. Accessed 15 June 2018.

5. Gerhs K, Chong L, Gusman G, Street D. Can we educate primary care physicians about diabetic retinopathy after graduation? Preliminary results of the diabetic retinopathy education study. Invest Ophthalmol Vis Sci. 1993;34:1182–5.

6. Bamashmus MA, Ophth FRC, Gunaid AA. Regular Visits to a Diabetes Clinic Were Associated with Lower Magnitude of Visual Disability and Diabetic Retinopathy- A Hospital-Based Historical Cohort Study in Yemen. Diabetes Technol Ther. 2009 Jan;11(1):45-50. PMID: 19132855. DOI: 10.1089/dia.2007.0299

7. Adriono G, Wang D, Octavianus C, Congdon N. Use of eye care services among diabetic patients in urban Indonesia. Arch Ophthalmol. 2011 Jul;129(7):930-5. PMID: 21746983. DOI: 10.1001/archophthalmol.2011.147

8. Burgess PI, Msukwa G, Beare NA V. Diabetic retinopathy in sub-Saharan Africa: meeting the challenges of an emerging epidemic. BMC Medicine. 2013;11(1):157. http://www.scopus.com/inward/record.url?eid=2-s2.0-84879800827&partnerID=tZOtx3y1. Accessed on 20 June 2018.

9. Muecke JS, Newland HS, Ryan P, Ramsay E, Aung M, Myint S, et al. Awareness of diabetic eye disease among general practitioners and diabetic patients in Yangon, Myanmar. Clin Exp Ophthalmol. 2008;36(3):265–73. PMID: 18412597 DOI: 10.1111/j.1442-9071.2008.01724.x

10. Khandekar R. Screening and Public Health Strategies for Diabetic Retinopathy in the Eastern Mediterranean Region. Middle East Afr J Ophthalmol. 2012;19(2):178–84. PMID: 22623855 PMCID: PMC3353664 DOI: 10.4103/0974-9233.95245

11. James M, Turner D, Broadbent D, Vora J, Harding S. Cost effectiveness analysis of screening for sight threatening diabetic eye disease. BMJ. 2000;320(7250):1627–31. http://www.bmj.com/cgi/doi/10.1136/bmj.320.7250.1627. Accessed on 20 June 2018.

12. Tran KT, Mendel TA, Holbrook KL, Yates PA. Construction of an inexpensive, hand-held fundus camera through modification of a consumer “point-and-shoot” camera. Investig Ophthalmol Vis Sci. 2012;53(12):7600–7607. DOI:10.1167/iovs.12-10449

13. Banaee T, Ansari-Astaneh MR, Pourreza H, Faal-Hosseini F, Vatanparast M, Shoeibi N, et al. Utility of 1% Tropicamide in Improving the Quality of Images for Tele-Screening of Diabetic Retinopathy in Patients with Dark Irides. Ophthalmic Epidemiol. 2017;24(4):217–21. https://doi.org/10.1080/09286586.2016.1274039. Accessed on 20 June 2018.

14. Scanlon P, Foy C, Malhotra R, Aldington S. The Influence of Age, Duration of Diabetes, Cataract, and Pupil Size on. Diabetes Care. 2005 Oct;28(10):2448-53. PMID: 16186278

15. Scanlon PH. The English National Screening Programme for diabetic retinopathy 2003–2016. Acta Diabetol. 2017;54(6):515–25. PMID: 28224275 PMCID: PMC5429356 DOI: 10.1007/s00592-017-0974-1

16. Kohner E, Stratton I, Aldington S, Holman R, Matthews D. Relationship between the severity of retinopathy and progression to photocoagulation in patients with Type 2 diabetes mellitus in the UKPDS (UKPDS 52). UK Prospect Diabetes Study Group. Diabet Med. 2001;18(3):178–84. PMID: 11318837

17. Baeza M, Orozco-Beltrán D, Gil-Guillen VF, Pedrera V, Ribera MC, Pertusa S, et al. Screening for sight threatening diabetic retinopathy using non-mydriatic retinal camera in a primary care setting: To dilate or not to dilate? Int J Clin Pract. 2009;63(3):433–8. PMID: 19222628 DOI: 10.1111/j.1742-1241.2008.01921.x

18. Perrier M, Boucher MC, Angioi K, Gresset JA, Olivier S. Comparison of two, three and four 45 degrees image fields obtained with the Topcon CRW6 nonmydriatic camera for screening for diabetic retinopathy. Can J Ophthalmol. 2003;38(7):569–74. http://www.ncbi.nlm.nih.gov/pubmed/14740798. Accessed on 20 May 2016.

19. Hansen AB, Hartvig N V., Jensen MS, Borch-Johnsen K, Lund-Andersen H, Larsen M. Diabetic retinopathy screening using digital non-mydriatic fundus photography and automated image analysis. Acta Ophthalmol Scand. 2004;82(6):666–72. PMID: 15606461 DOI: 10.1111/j.1600-0420.2004.00350.x

20. Aptel F, Denis P, Rouberol F, Thivolet C. Screening of diabetic retinopathy: Effect of field number and mydriasis on sensitivity and specificity of digital fundus photography. Diabetes Metab. 2008;34(3):290–3. PMID: 18406188 DOI: 10.1016/j.diabet.2007.12.007

21. Scanlon PH, Malhotra R, Greenwood RH, Aldington SJ, Foy C, Flatman M, et al. Comparison of two reference standards in validating two field mydriatic digital photography as a method of screening for diabetic retinopathy. Br J Ophthalmol. 2003;87(10):1258–63. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1920793&tool=pmcentrez&rendertype=abstract. Accessed on 20 May 2016.

22. Grading diabetic retinopathy from stereoscopic colour fundus photographs - An extension of the modified Airlie House classification: ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Group. Ophthalmology. 1991 May;98(5 Suppl):786-806. PMID: 2062513

23. Herbert HM, Jordan K, Flanagan DW. Is screening with digital imaging using one retinal view adequate? Eye (Lond). 2003;17(4):497–500. PMID: 12802350 DOI: 10.1038/sj.eye.6700409

24. Ku JY, Landers J, Henderson T, Craig JE. The reliability of single-field fundus photography in screening for diabetic retinopathy: The central Australian ocular health study. Med J Aust. 2013;198(2):93–5. DOI: 10.5694/mja12.10607

25. Farley TF. Accuracy of Primary Care Clinicians in Screening for Diabetic Retinopathy Using Single-Image Retinal Photography. Ann Fam Med. 2008 Sep-Oct;6(5):428-34. PMID: 18779547 PMCID: PMC2532778 DOI: 10.1370/afm.857

26. Suansilpong A, Rawdaree P. Accuracy of Single-Field Nonmydriatic Digital Fundus Image in Screening for Diabetic Retinopathy. J Med Assoc Thai. 2008;91(9):1397–403. PMID: 18843870

27. Kuo HK, Hsieh HH, Liu RT. Screening for diabetic retinopathy by one-field, non-mydriatic, 45° digital photography is inadequate. Ophthalmologica. 2005;219(5):292–6. PMID: 16123556 DOI: 10.1159/000086114

28. Murgatroyd H. Effect of mydriasis and different field strategies on digital image screening of diabetic eye disease. Br J Ophthalmol. 2004;88(7):920–4. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1772219&tool=pmcentrez&rendertype=abstract. Accessed 20 May 2016.

29. Gupta V, Bansal R, Gupta A, Bhansali A. Sensitivity and specificity of nonmydriatic digital imaging in screening diabetic retinopathy in Indian eyes. Indian J Ophthalmol. 2014 Aug;62(8):851-6. DOI: 10.4103/0301-4738.141039

30. Bourne RRA, Jonas JB, Bron AM, Cicinelli MV, Das A, Flaxman SR, et al. Vision Loss Expert Group of the Global Burden of Disease Study. Prevalence and causes of vision loss in high-income countries and in Eastern and Central Europe in 2015: magnitude , temporal trends and projections. Br J Ophthalmol. 2018 May;102(5):575-585. PMID: 29545417 PMCID: PMC5909755 DOI: 10.1136/bjophthalmol-2017-311258

31. Pascolini D, Mariotti SP. Global estimates of visual impairment: 2010. Br J Ophthalmol. 2012;96(5):614–8. PMID: 22133988 DOI: 10.1136/bjophthalmol-2011-300539

32. Murthy GVS, Gupta SK, John N, Vashist P. Current status of cataract blindness and Vision 2020: The right to sight initiative in India. Indian J Ophthalmol. 2008 Nov-Dec; 56(6): 489–494.

33. Henricsson M, Karlsson C, Ekholm L, Kaikkonen P, Sellman A, Steffert E, et al. Colour slides or digital photography in diabetes screening-a comparison. Acta Ophthalmol Scand. 2000;78(2):164–8.

34. Owens D, Gibbins R, Lewis P, Wall S, Allen J, Morton R. Screening for diabetic retinopathy by general practitioners: Ophthalmoscopy or retinal photography as 35 mm colour transparencies? Diabet Med.1998;15(2):170–5. PMID: 9507921 DOI: 10.1002/(SICI)1096-9136(199802)15:2<170::AID-DIA518>3.0.CO;2-H

35. Hutchinson A, McIntosh A, Peters J, et al. Review: Mydriatic retinal photography is the most effective test for detecting diabetic retinopathy. Evid Based Med. 2001;6(2):56. http://ebm.bmj.com/content/ebmed/6/2/56.full.pdf. Accessed 20 May 2018.

36. Ding J, Zou Y, Liu N, Jiang L, Ren X, Jia W, et al. Strategies of digital fundus photography for screening diabetic retinopathy in a diabetic population in urban China. Ophthalmic Epidemiol. 2012;19(6):414–9. PMID: 23171212 DOI: 10.3109/09286586.2012.716895

37. Mizrachi Y, Knyazer B, Guigui S, Rosen S, Lifshitz T, Belfair N, et al. Evaluation of diabetic retinopathy screening using a non-mydriatic retinal digital camera in primary care settings in south Israel. International Ophthalmology. 2014;34(4):831-837. PMID:24292883 DOI: 10.1007/s10792-013-9887-3.

38. Bhargava M, Cheung CYL, Sabanayagam C, Kawasaki R, Harper CA, Lamoureux EL, et al. Accuracy of diabetic retinopathy screening by trained non-physician graders using non-mydriatic fundus camera. Singapore Med J. 2012;53(11):715–9. PMID: 23192497

39. Katulanda P, Waniganayake YC, Ranasinghe P, Udai WM, Wijetunga A, Jayaweera M, et al. Retinopathy among young adults with Diabetes Mellitus from a tertiary care setting in Sri Lanka. BMC Endocr Disord. 2014;14(1):1–8. https://doi.org/10.1186/1472-6823-14-20

40. Taylor DJ, Goatman KA, Gregory A, Histed M, Martin CN, Warburton TJ, et al. Image-quality standardization for diabetic retinopathy screening. Expert Rev Ophthalmol. 2009;4(5):469–76.

41. Olson JA, Strachan FM, Hipwell JH, Goatman KA, McHardy KC, Forrester J V., et al. A comparative evaluation of digital imaging, retinal photography and optometrist examination in screening for diabetic retinopathy. Diabet Med. 2003;20(7):528–34. PMID: 12823232

Download PDF

Journal Publication

published 06 Apr, 2019

Read the published version in BMC Ophthalmology →

Editorial decision: Revise
06 Jan, 2019
Review #3 received at journal
03 Jan, 2019
Review #2 received at journal
25 Dec, 2018
Review #1 received at journal
24 Dec, 2018
Reviewers invited by journal
20 Dec, 2018
Reviewer #1 agreed at journal
20 Dec, 2018
Reviewer #2 agreed at journal
20 Dec, 2018
Reviewer #3 agreed at journal
20 Dec, 2018
Editor assigned by journal
10 Dec, 2018
Editor invited by journal
06 Dec, 2018

You are reading this older preprint version

Read the latest preprint version →

Diagnostic Test Accuracy of Diabetic Retinopathy Screening by Physician Graders Using a Hand Held Non-Mydriatic Retinal Camera at a Tertiary Level Medical Clinic in a Lower-Middle Income Country Setting

Status:

Journal Publication

Version 1

Abstract

Figures

Background

Methods

Results

Discussion

Conclusion

Declarations

References

Supplementary Files

Status:

Journal Publication

Version 1