How AI Can Transform ROP Care in Bangladesh

A new baby usually starts to see clearly within six months after birth. But what happens when an infant is born prematurely before the gestational age of 35 weeks or weighs less than 2,000 grams at birth? A premature baby risks developing Retinopathy of Prematurity (ROP), a preventable yet potentially blinding condition that disrupts the vision development process. As medical advancements are helping more and more premature babies survive, the need for making sure they are not left without sight becomes increasingly urgent.

Now is a critical time for Bangladesh in addressing ROP, as advancements in the neonatal health sector have improved survival rates for prematurely born babies, but limited screening, referral and treatment services mean that many babies are at risk of preventable childhood blindness. Bangladesh has made significant progress in addressing ROP through the formulation of national guidelines and the expansion of ROP services over the past decade. Still access remains uneven and many vulnerable infants are still missed from ROP screening and treatment. Studies suggest 20-40% of screened premature babies develop ROP, making it a growing public health concern. However, ROP-related blindness is largely preventable. Timely screening, early referral, and prompt treatment can save the sight of thousands of vulnerable children, making investment in nationwide ROP services both a health imperative and a commitment for the country’s future. Artificial intelligence (AI) proves to be a very helpful innovation for fighting ROP. The current way of diagnosing the disease involves a number of problems, such as a dependency on a small number of specialists and a highly subjective analysis of retinal images. Therefore, AI technology may assist in analysing images quickly and accurately in identifying those children who are at risk of developing the disease.

By helping trained medical personnel to capture retinal images for automated analysis, AI can reduce the burden on specialists, improve referral efficiency and ensure timely care of high-risk children. This feature may prove especially useful in situations when there is a shortage of resources and geographic barriers are preventing the provision of services.

Bangladesh is not the only country facing the growing challenges of ROP. Latin America and Southeast Asia face similar challenges because of the improvement in neonatal services which has saved many lives of premature children, but screening and treatment cannot keep up with this process. So the burden of preventable childhood blindness is growing in many countries.

Global experience suggests teleconsultation and AI can help address workforce shortages, expand eye screening coverage and enhance diagnostic accuracy when these are supported by national investment and strong health systems. The Orbis-Siloam Vision Partnership demonstrates the significant potential of AI-enabled ROP screening in resource-constrained settings. Since 2023, the initiative has leveraged AI-based teleconsultation platforms to facilitate the real-time analysis of retinal images, supporting earlier diagnosis and more timely clinical decision-making. This has helped strengthen the capacity of local healthcare providers to capture retinal images using cost-effective digital technologies. Experience also shows that AI-supported screening can substantially enhance access to timely ROP detection, referral and treatment services. It reveals an important lesson – implementation of technology-based solutions involving AI, teleconsultation and workforce capacity development provides an opportunity to develop and implement a reliable, efficient and sustainable strategy of widening accessibility to quality ROP services for many infants, thereby helping them to survive, maintain their visual functions and achieve their full developmental potential.
The current situation in Bangladesh is particularly challenging in relation to ROP. There is a need to ensure that the relatively high rates of neonatal survival will be followed by effective and quality screening processes. The use of AI solutions can serve as a viable strategy in this case.

This new technology presents an unprecedented chance for Bangladesh to revolutionise its ROP services. With AI-powered screening procedures, identifying patients at maximum risk is faster, leading to more efficient allocation of limited specialist resources to those patients who require immediate treatment in comparison to the current system of ROP screening and treatment. Utilised in conjunction with teleconsultation, AI allows extending screening procedures to rural and remote regions of the country rather than limiting it to big cities like Dhaka and facilitating remote assessment of the retinal images, thus, diagnosing ROP. Building on the ongoing collaboration between Bangladesh Medical University (BMU) and Orbis, Bangladesh can become a regional leader in AI-driven neonatal eye care, demonstrating how technology, training and health system strengthening can work together to prevent avoidable childhood blindness.

As premature and low-birth-weight infants are predominantly admitted to NICUs, training neonatologists and nurses in retinal image capture and integrating ROP AI at the district-level NICUs for ROP screening could establish a more responsive referral mechanism towards the apex ROP centres (hub). This approach would support the early detection and management of infants at risk of ROP, reduce workload at specialist end and unnecessary travel for the patients and caregivers. With its strong academic institutions, expanding digital infrastructure and ongoing investments in maternal and neonatal health, Bangladesh is well positioned to adapt these global innovations.

The application of AI in ROP screening might not only bridge the current gaps but also position Bangladesh to serve as an example for a combined technological approach to neonatal care and ROP screening in South Asia, showing how countries transitioning in neonatal survival can simultaneously safeguard vision outcomes for their most vulnerable populations.

For reaping the maximum benefit from AI in ROP screening, Bangladesh needs to adopt clear policies, ethical guidelines and regulations on how it should be used in the screening, referral and treatment process. A clear guideline in terms of data protection, cyber security, workforce education and infrastructure is crucial in ensuring confidentiality of patients’ data and successful implementation.

AI is there to enhance the clinical capacity. With the right supervision and qualified experts, AI will enable greater precision in diagnostics, efficiency in service delivery and access to sight-saving care. With the help of both ROP AI and an ethical approach, Bangladesh is likely to improve ROP services and prevent blindness in children.
It goes without saying, however, that the use of AI technologies in ROP treatment is essential for public health reasons. With more premature babies surviving after birth than before, it is imperative for Bangladesh to make sure that the lives saved also go hand in hand with the ability to see. 

In order to achieve that result, it is important to combine efforts of the government, medical facilities, development partners and the private sector to develop a comprehensive strategy aimed at transforming innovation into practice. In this case, AI may be one of the tools used to help fill the gap left by specialist capacity, but the key will be in responsible and safe implementation. Today, Bangladesh has a chance to become a regional role model in the field of AI-supported neonatal eye care. This innovative system can ensure the protection of the eyesight of the most vulnerable infants and prove the positive impact of innovative technologies in the healthcare sector.
_____________________________________________________________
Afroja Yesmin, Mohammad Yeadul Hassan, Dr Lutful Husain, Iqbal Hossain, and Dr Munir Ahmed works for Orbis International Bangladesh and Prof Dr Mohammad Abdul Mannan works at Bangladesh Medical University