CHILDSTAR: CHIldren Living With Diabetes See and Thrive with AI Review.

Background: Artificial intelligence (AI) appears capable of detecting diabetic retinopathy (DR) with a high degree of accuracy in adults; however, there are few studies in children and young adults. Methods: Children and young adults (3-26 years) with type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM) were screened at the Dhaka BIRDEM-2 hospital, Bangladesh. All gradable fundus images were uploaded to Cybersight AI for interpretation. Two main outcomes were considered at a patient level: 1) Any DR, defined as mild non-proliferative diabetic retinopathy (NPDR or more severe; and 2) Referable DR, defined as moderate NPDR or more severe. Diagnostic test performance comparing Orbis International's Cybersight AI with the reference standard, a fully qualified optometrist certified in DR grading, was assessed using the Matthews correlation coefficient (MCC), area under the receiver operating characteristic curve (AUC-ROC), area under the precision-recall curve (AUC-PR), sensitivity, specificity, positive and negative predictive values. Results: Among 1274 participants (53.1% female, mean age 16.7 years), 19.4% (n = 247) had any DR according to AI. For referable DR, 2.35% (n = 30) were detected by AI. The sensitivity and specificity of AI for any DR were 75.5% (CI 69.7-81.3%) and 91.8% (CI 90.2-93.5%) respectively, and for referable DR, these values were 84.2% (CI 67.8-100%) and 98.9% (CI 98.3%-99.5%). The MCC, AUC-ROC and the AUC-PR for referable DR were 63.4, 91.2 and 76.2% respectively. AI was most successful in accurately classifying younger children with shorter duration of diabetes. Conclusions: Cybersight AI accurately detected any DR and referable DR among children and young adults, despite its algorithms having been trained on adults. The observed high specificity is particularly important to avoid over-referral in low-resource settings. AI may be an effective tool to reduce demands on scarce physician resources for the care of children with diabetes in low-resource settings.

Autonomous artificial intelligence increases real-world specialist clinic productivity in a cluster-randomized trial.

Autonomous artificial intelligence (AI) promises to increase healthcare productivity, but real-world evidence is lacking. We developed a clinic productivity model to generate testable hypotheses and study design for a preregistered cluster-randomized clinical trial, in which we tested the hypothesis that a previously validated US FDA-authorized AI for diabetic eye exams increases clinic productivity (number of completed care encounters per hour per specialist physician) among patients with diabetes. Here we report that 105 clinic days are cluster randomized to either intervention (using AI diagnosis; 51 days; 494 patients) or control (not using AI diagnosis; 54 days; 499 patients). The prespecified primary endpoint is met: AI leads to 40% higher productivity (1.59 encounters/hour, 95% confidence interval [CI]: 1.37-1.80) than control (1.14 encounters/hour, 95% CI: 1.02-1.25), p < 0.00; the secondary endpoint (productivity in all patients) is also met. Autonomous AI increases healthcare system productivity, which could potentially increase access and reduce health disparities. ClinicalTrials.gov NCT05182580.

Visual Impairment and Risk of Self-Reported Road Traffic Crashes Among Bus Drivers in Bangladesh.

PURPOSE: This study assesses the prevalence and the causes of visual impairment among bus drivers undergoing screening in Bangladesh and associations with self-reported crashes. METHODS: Eye health screenings including refraction and questionnaires were conducted at 10 bus terminals in 7 districts of Bangladesh from June through August 2019. Presenting near and distance visual impairment and self-reported road traffic crashes were recorded. RESULTS: Among 700 participants, nearly 1 in 5 (n = 126, 18.0%) had presented visual acuity (VA) in the better-seeing eye ≤6/9, not meeting the vision standard of Bangladesh for bus drivers. A majority of drivers (n = 492, 70.3%) had near or distance refractive error, and most who failed to reach the driving standard (88.1%, n = 111) could be improved with readily available treatment, either glasses or cataract surgery. A history of 1 or more road traffic crashes while driving a bus or minibus was reported by 62 (8.97%) participants. In multivariable models, factors significantly associated with near or distance visual impairment included older age [odds ratio (OR) per year 1.15, 95% confidence interval (CI): 1.12-1.18, P < 0.0001] and having no driver's license (OR 1.80, 95% CI: 1.04-3.13, P = 0.037). Self-reported history of a motor vehicle crash was associated with near or distance visual impairment (OR 2.45, 95% CI: 1.09-5.49, P = 0.030), even when adjusting for other factors such as age and weekly miles are driven. CONCLUSIONS: Ensuring that bus drivers are screened to meet the required visual standards for driving while referring those who do not for treatment, can contribute to safer roads in this high-risk setting.

Vision Impairment and Productivity Among Female Garment Workers in Bangladesh: A Cohort Study.

PURPOSE: To assess the prevalence of near and correctable distance visual impairment among screened participants in the garment industry and to explore associations with income, age, and urban versus rural residence. METHODS: Vision screenings were conducted at 4 garment factories, 2 urban and 2 rural locations during September and October 2019. Distance vision impairment was the presence of uncorrected vision of <6/12 in either eye, correctable to ≥6/7.5 with distance refraction. Near vision impairment was defined as 1 or more of the following: 1) either eye with presenting near vision 6/12 in the same eye; 2) having been prescribed near add spectacle power in examination records; and/or 3) clinical diagnosis of presbyopia at the time of screening. Demographic information and monthly income were self-reported by questionnaire completion. RESULTS: Among 915 participating workers (100% female, 18 to 70 years), 29.2% (n = 267) and 26.8% (n = 245) had correctable distance and near vision impairment respectively. Prevalence of near vision impairment was significantly higher among rural residents (34.2%, n = 160), compared to urban (19.0%, n = 85, P < 0.0001) with the largest differences in the 35 to 39 (68.2% vs 44.2%, P = 0.0019) and 40+ (85.9 vs 48.9%, P < 0.0001) year age ranges. Prevalence of near vision impairment was already high among urban (20.4%, n = 20) and rural (23.0%, n = 17) workers aged 30 to 34 years. In simple linear regression models, participants with near vision impairment earned $13.3 [standard error (SE) 2.44, P < 0.0001] less per month than those without, while urban residents earned $40.6 (SE 1.74, P < 0.0001) more than rural dwellers. In the final multivariate linear model, both near vision impairment ($6.51 lower monthly earnings, SE 1.84, P = 0.0004) and urban residence ($43.2 higher monthly earnings, SE 2.39, P < 0.0001) remained significantly associated with income. CONCLUSIONS: This study found high rates of near vision impairment among female garment workers, particularly rural dwellers, and at a younger age than expected. The high prevalence and association between near vision impairment and lower income suggest that focusing on industries with a high proportion of female workers, such as readymade garments, may be effective in addressing gender disparities in vision impairment and its economic impact.

Risk factors associated with retinopathy in young people with type 1 diabetes in Bangladesh.

Objectives: Diabetic retinopathy (DR) is the most common microvascular complications seen in children and adolescents with type 1 diabetes. The aim of this study was to evaluate the prevalence of retinopathy and its association with other risk factors in young people with type 1 diabetes. Methods: This study was a cross-sectional study, which was done as part of the ongoing complication assessment in the paediatric diabetes clinic in BIRDEM (Bangladesh Institute of Research and Rehabilitation of Diabetes Endocrine and Metabolic Disorders), a tertiary care hospital. Children, adolescents and young adults with type 1 diabetes who were having diabetes duration >2 years were included in this study. Retinopathy was detected using fundal photography, and grading was done by National Screening Committee of UK by trained ophthalmologists. Results: Diabetic retinopathy was observed in 44 (6.6%) patients. Majority (95.4%) of them had early diabetic retinopathy in the form of mild NPDR (nonproliferative diabetic retinopathy) (R1). Patients with retinopathy had higher HbA1c 9.6[8.4-12.3] vs 9.1 [7.9-10.8] (P = .013), longer duration of diabetes 7.6 [5.5-10.7] vs 6.0 [4.5-8.2] years (P = .001) and were older 21.5 [18.0-23.0] vs 18 [16.0-21.0] years (P = .0001) compared with those without retinopathy. On multivariate regression analysis, higher age and median HbA1c were significantly associated with DR. Conclusions: Higher HbA1c was the only modifiable risk factor for development of DR in our study population. Early detection of DR with improvement of glycaemic control may reduce the risk of progression of severe stages of the disease.

Rapid assessment of avoidable blindness and cataract surgery coverage among forcibly displaced Myanmar Nationals (Rohingya refugees) in Cox's Bazar, Bangladesh.

AIM: To determine the prevalence and causes of blindness, vision impairment and cataract surgery coverage among Rohingya refugees aged ≥ 50 years residing in camps in Cox's Bazar, Bangladesh. METHODS: We used the Rapid Assessment of Avoidable Blindness (RAAB) methodology to select 76 clusters of 50 participants aged ≥ 50 years with probability proportionate to size. Demographic and cataract surgery data were collected using questionnaires, visual acuity was assessed per World Health Organization criteria and examinations were conducted by torch, and with direct ophthalmoscopy in eyes with pinhole-corrected vision <6/12. RAAB software was used for data entry and analysis. RESULTS: We examined 3,629 of 3800 selected persons (95.5%). Age and sex adjusted prevalence of blindness (<3/60), severe visual impairment (SVI; >3/60 to ≤6/60), moderate visual impairment (MVI; >6/60 to ≤6/18), and early visual impairment (EVI; >6/18 to ≤6/12) were 2.14%, 2.35%, 9.68% and 14.7% respectively. Cataract was responsible for 75.0% of blindness and 75.8% of SVI, while refractive error caused 47.9% and 90.9% of MVI and EVI respectively. Most vision loss (95.9%) was avoidable. Cataract surgical coverage among the blind was 81.2%. Refractive error was detected in 17.1% (n = 622) of participants and 95.2% (n = 592) of these did not have spectacles. In the full Rohingya cohort of 76,692, approximately 10,000 surgeries are needed to correct all eyes impaired (<6/18) by cataract, 12,000 need distance glasses and 73,000 require presbyopic correction. CONCLUSION: The prevalence of blindness was lower than expected for a displaced population, in part due to few Rohingya being ≥60 years and the camp's good access to cataract surgery. We suggest the United Nations High Commissioner for Refugees include eye care among recommended health services for all refugees with long-term displacement.

Burden of eye disease and demand for care in the Bangladesh Rohingya displaced population and host community: A cohort study.

BACKGROUND: There is a growing awareness that addressing chronic as well as acute health conditions may contribute importantly to the well-being of displaced populations, but eye care service has generally not been prioritized in crisis situations. We describe a replicable model of eye care provision as delivered by Orbis International and local partners to the Rohingya and host population in Cox's Bazar, Bangladesh, and characterize the burden of vision impairment and demand for sight-restoring services in this setting. METHODS AND FINDINGS: Orbis International and local secondary facility Cox's Bazar Baitush Sharaf Hospital (CBBSH) provide eye care support to the Rohingya population and the host community of all ages in Cox's Bazar, Bangladesh, with fixed vision screening locations established in Camps 4 and 11 of the Kutupalong refugee settlement. Structured outreach targets these camps and four surrounding local subdistricts, with referrals made as needed for refraction (glasses measurement) and cataract surgery to CBBSH. Between February 2018 and March 2019, 48,105 displaced Rohingya (70.3%, among whom 71.6% were children and 46.5% women) and 20,357 local residents (29.7%, 88.5% children, 54.4% women) underwent vision screening. Displaced Rohingya sought services from a total of 12 surrounding camps, within which coverage was 17.3%, including 43.3% (27,027/62,424) of children aged 5-11 years and 60.0% (5,315/8,857) of adults ≥ 60 years old. The prevalence of blindness (presenting acuity < 3/60) among Rohingya patients exceeded that among local residents by 3- to 6-fold in each 10-year age group between 18 and 59 years (P < 0.001 comparing vision between the two groups in this age range), and the prevalence of cataract requiring surgery was also higher in Rohingya patients (18-29 years: 4.67% versus 1.80%, P = 0.0019; 30-39: 7.61% versus 2.39%, P < 0.001; and 40-49 years: 7.91% versus 3.77%, P = 0.0014). A limitation of the study is lack of data on population prevalence of eye disease. CONCLUSIONS: The burden of untreated eye disease is very high among the Rohingya, particularly those in their peak working years who could contribute most to the resiliency of their community. Demand for eye care service is also great among children and adults in this population with many competing healthcare priorities. Research is needed, building on strong evidence of benefit in settled populations, to explore the specific impact of vision care on the well-being of displaced populations.