Scholarlink

Background <p>Diabetic retinopathy is one of the leading causes of preventable blindness among people with diabetes. Despite the availability of effective screening methods, access to timely retinal examination remains limited in many rural and resource-constrained settings because of shortages of ophthalmologists, high equipment costs, and logistical barriers. Artificial intelligence (AI)-enabled portable fundus imaging has emerged as a promising approach to improve early detection and expand screening services at the primary healthcare level.</p> Objective <p>This study examines the clinical effectiveness and practical implementation of an indigenous, low-cost AI-enabled portable fundus camera system for detecting diabetic retinopathy in rural Indian primary healthcare centers where specialized ophthalmological services remain largely inaccessible.</p> Methods <p>We conducted a 6-month prospective validation study across eight primary health centers and screened 612 diabetic patients. The portable device’s diagnostic accuracy was compared against gold-standard dilated fundus examination by retinal specialists. Implementation barriers, cost-effectiveness, and healthcare worker acceptance were systematically evaluated.</p> Results <p>The AI-enabled device demonstrated 91.3% sensitivity and 88.7% specificity for detecting referable diabetic retinopathy, with particularly strong performance in identifying proliferative cases (94.2% sensitivity). The system reduced diagnostic time from weeks to less than 15&#xa0;min per patient, with a per-screening cost of ₹185 compared to ₹2,400 for traditional referral pathways. Healthcare workers achieved operational proficiency within three training sessions.</p> Conclusion <p>This indigenous technology offers a scalable, cost-effective solution for diabetic retinopathy screening in resource-limited settings. The device addresses critical healthcare access gaps while maintaining diagnostic accuracy comparable to specialist examination. Regulatory approval and national program adoption would significantly impact preventable blindness reduction in India’s diabetic population.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI-driven clinical decision support for diabetic retinopathy screening using portable fundus imaging in low-resource health systems

  • 期刊论文

  • 发表日期：

    2026年7月30日

查看全文
  • Venkata Nagaraj Kakaraparthi,
  • Vamsi Krishna Gannamaneni,
  • Paul Silvian Samuel,
  • Ravi Shankar Reddy,
  • Lalitha Kakaraparthi

摘要

Background

Diabetic retinopathy is one of the leading causes of preventable blindness among people with diabetes. Despite the availability of effective screening methods, access to timely retinal examination remains limited in many rural and resource-constrained settings because of shortages of ophthalmologists, high equipment costs, and logistical barriers. Artificial intelligence (AI)-enabled portable fundus imaging has emerged as a promising approach to improve early detection and expand screening services at the primary healthcare level.

Objective

This study examines the clinical effectiveness and practical implementation of an indigenous, low-cost AI-enabled portable fundus camera system for detecting diabetic retinopathy in rural Indian primary healthcare centers where specialized ophthalmological services remain largely inaccessible.

Methods

We conducted a 6-month prospective validation study across eight primary health centers and screened 612 diabetic patients. The portable device’s diagnostic accuracy was compared against gold-standard dilated fundus examination by retinal specialists. Implementation barriers, cost-effectiveness, and healthcare worker acceptance were systematically evaluated.

Results

The AI-enabled device demonstrated 91.3% sensitivity and 88.7% specificity for detecting referable diabetic retinopathy, with particularly strong performance in identifying proliferative cases (94.2% sensitivity). The system reduced diagnostic time from weeks to less than 15 min per patient, with a per-screening cost of ₹185 compared to ₹2,400 for traditional referral pathways. Healthcare workers achieved operational proficiency within three training sessions.

Conclusion

This indigenous technology offers a scalable, cost-effective solution for diabetic retinopathy screening in resource-limited settings. The device addresses critical healthcare access gaps while maintaining diagnostic accuracy comparable to specialist examination. Regulatory approval and national program adoption would significantly impact preventable blindness reduction in India’s diabetic population.