Artificial intelligence-supported diabetic retinopathy screening in Tanzania: rationale and design of a randomised controlled trial

INTRODUCTION

Globally, diabetic retinopathy (DR) is a major cause of blindness. Sub-Saharan Africa is projected to see the largest proportionate increase in the number of people living with diabetes over the next two decades. Screening for DR is recommended to prevent sight loss; however, in many low and middle-income countries, because of a lack of specialist eye care staff, current screening services for DR are not optimal. The use of artificial intelligence (AI) for DR screening, which automates the grading of retinal photographs and provides a point-of-screening result, offers an innovative potential solution to improve DR screening in Tanzania.

METHODS AND ANALYSIS

We will test the hypothesis that AI-supported DR screening increases the proportion of persons with true referable DR who attend the central ophthalmology clinic following referral after screening in a single-masked, parallel group, individually randomised controlled trial. Participants (2364) will be randomised (1:1 ratio) to either AI-supported or the standard of care DR screening pathway. Participants allocated to the AI-supported screening pathway will receive their result followed by point-of-screening counselling immediately after retinal image capture. Participants in the standard of care arm will receive their result and counselling by phone once the retinal images have been graded in the usual way (typically after 2-4 weeks). The primary outcome is the proportion of persons with true referable DR attending the central ophthalmology clinic within 8 weeks of screening. Secondary outcomes, by trial arm, include the proportion of persons attending the central ophthalmology clinic out of all those referred, sensitivity and specificity, number of false positive referrals, acceptability and fidelity of AI-supported screening.

ETHICS AND DISSEMINATION

The London School of Hygiene & Tropical Medicine, Kilimanjaro Christian Medical Centre and Tanzanian National Institute of Medical Research ethics committees have approved the trial. The results will be submitted to peer-reviewed journals for publication.

TRIAL REGISTRATION NUMBER

ISRCTN18317152.

Artificial intelligence for diabetic retinopathy in low-income and middle-income countries: a scoping review

Diabetic retinopathy (DR) is a leading cause of blindness globally. There is growing evidence to support the use of artificial intelligence (AI) in diabetic eye care, particularly for screening populations at risk of sight loss from DR in low-income and middle-income countries (LMICs) where resources are most stretched. However, implementation into clinical practice remains limited. We conducted a scoping review to identify what AI tools have been used for DR in LMICs and to report their performance and relevant characteristics. 81 articles were included. The reported sensitivities and specificities were generally high providing evidence to support use in clinical practice. However, the majority of studies focused on sensitivity and specificity only and there was limited information on cost, regulatory approvals and whether the use of AI improved health outcomes. Further research that goes beyond reporting sensitivities and specificities is needed prior to wider implementation.

Prevalence, Risk Factors, and Complications of Diabetes in the Kilimanjaro Region: A Population-Based Study from Tanzania

BACKGROUND

In sub-Saharan Africa, diabetes is a growing burden, yet little is known about its prevalence, risk factors, and complications. To address these gaps and help inform public health efforts aimed at prevention and treatment, we conducted a community-based study assessing diabetes epidemiology.

METHODS AND FINDINGS

We conducted a stratified, cluster-designed, serial cross-sectional household study from 2014-2015 in the Kilimanjaro Region, Tanzania. We used a three-stage cluster probability sampling method to randomly select individuals. To estimate prevalence, we screened individuals for glucose impairment, including diabetes, using hemoglobin A1C. We also screened for hypertension and obesity, and to assess for potential complications, individuals with diabetes were assessed for retinopathy, neuropathy, and nephropathy. We enrolled 481 adults from 346 urban and rural households. The prevalence of glucose impairment was 21.7% (95% CI 15.2-29.8), which included diabetes (5.7%; 95% CI 3.37-9.47) and glucose impairment with increased risk for diabetes (16.0%; 95% CI 10.2-24.0). Overweight or obesity status had an independent prevalence risk ratio for glucose impairment (2.16; 95% CI 1.39-3.36). Diabetes awareness was low (35.6%), and few individuals with diabetes were receiving biomedical treatment (33.3%). Diabetes-associated complications were common (50.2%; 95% CI 33.7-66.7), including renal (12.0%; 95% CI 4.7-27.3), ophthalmic (49.6%; 95% CI 28.6-70.7), and neurological (28.8%; 95% CI 8.0-65.1) abnormalities.

CONCLUSIONS

In a northern region of Tanzania, diabetes is an under-recognized health condition, despite the fact that many people either have diabetes or are at increased risk for developing diabetes. Most individuals were undiagnosed or untreated, and the prevalence of diabetes-associated complications was high. Public health efforts in this region will need to focus on reducing modifiable risk factors, which appear to include obesity, as well as early detection that includes increasing awareness. These findings highlight a growing urgency of diabetes prevention in this region as well as the need for treatment, including management of complications.

Reasons for poor follow-up of diabetic retinopathy patients after screening in Tanzania: a cross-sectional study

Background

Diabetes is an emerging public health problem in sub-Saharan Africa. Diabetic retinopathy is the commonest microvascular complication of diabetes and is a leading cause of blindness, mainly in adults of working age. Follow-up is crucial to the effective management of diabetic retinopathy, however, follow-up rates are often poor in sub-Saharan Africa. The aim of this study was to assess the proportion of patients not presenting for follow-up and the reasons for poor follow-up of diabetic patients after screening for retinopathy in Kilimanjaro Region of Tanzania.

Methods

All diabetic patients referred to a tertiary ophthalmology hospital after screening for retinopathy in 2012 were eligible for inclusion in the study. A randomly selected group of patients from the community-based diabetic retinopathy screening register were identified; among this group, follow-up was assessed. Interviews were conducted within this group to inform on the reasons for poor follow-up.

Results

Among the 203 patients interviewed in the study 50 patients (24.6 %) attended the recommended referral appointment and 153 (75.4 %) did not. Financial reasons were self-reported by 35.3 % of those who did not attend the follow-up appointment as the reason for non-attendance. Multiple logistic regression analysis showed that the patient report of the clarity of the referral process (p = 0.014) and the patient report of whether a healthcare worker told the patient that diabetic retinopathy could be treated (p = 0.005) were independently associated with attendance at a follow-up appointment. Income per month was not associated with attendance at a follow-up appointment on multivariate analysis.

Conclusions

Financial factors are commonly cited as the reason for non-compliance with follow-up recommendations. However, the reasons for poor compliance are likely to be more complicated. This study highlights the importance of health system factors. Improving the clarity of the referral process and frequent reminders to patients that diabetic retinopathy can be treated are practical strategies that should be incorporated into screening programmes to increase attendance at subsequent follow-up appointments. The results from this study are applicable to other screening programmes as well as those for diabetic retinopathy.

Electronic supplementary material

The online version of this article (doi:10.1186/s12886-016-0288-z) contains supplementary material, which is available to authorized users.

Diabetic retinopathy in Tanzania: prevalence and risk factors at entry into a regional screening programme

OBJECTIVE

The number of adults with diabetes in sub-Saharan Africa (SSA) is expected to almost double by 2035. This study investigated the prevalence of diabetic retinopathy (DR) and its risk factors at entry into a community-based screening programme.

METHODS

All persons with diabetes screened for retinopathy at entry into a screening programme in Kilimanjaro Region, Tanzania between November 2010 and December 2014 were included. Fundus photographs were taken with a Topcon retinal camera following pupil dilation. Data were collected on BP, random blood sugar, duration of diabetes, BMI and visual acuity on entry.

RESULTS

A total of 3187 persons were screened for DR. The prevalence of any DR was 27.9% (95%CI 26.4-29.5%) with background diabetic retinopathy (BDR), pre-proliferative diabetic retinopathy (PPDR) and proliferative diabetic retinopathy (PDR) having a prevalence of 19.1% (95% CI 17.7-20.4%), 6.0% (95%CI 5.2-6.8%) and 2.9% (95%CI 2.3-3.5%), respectively. Maculopathy was present in 16.1% (95%CI 14.8-17.4%) of participants. Multivariable logistic regression analysis for the presence of any DR found independent associations with duration of diabetes (P < 0.0001), systolic BP (P < 0.0001), random blood sugar (P < 0.0001) and attending a government hospital diabetic clinic (P = 0.0339).

CONCLUSIONS

This study is the first to present data from a DR screening programme in SSA. The results will provide policymakers with data to aid planning of DR screening and treatment services in the African region. The study highlights the importance of managing comorbidities within DR screening programmes.