Pediatric Telemedicine-based Retinal Screening in Ghana

Introduction Retinopathy of prematurity (ROP) is increasingly observed in neonatal intensive care units (NICUs) across Sub-Saharan Africa (SSA), driven by improved access to supplemental oxygen and rising survival rates of premature infants.[1-3] Fortunately, treatment with intravitreal bevacizumab is widely accessible and affordable throughout the country, costing less than $30 per eye, including physician time, drug, and instrument costs. Despite treatment availability, there are only 12 retina specialists or pediatric ophthalmologists in Ghana, primarily concentrated in the country’s capital, Accra. Scarce providers have resulted in limited bandwidth for consistent ROP screening. Screening for ROP and other pediatric retinal eye conditions is predominantly performed with binocular indirect ophthalmoscopy (BIO) which involves a skilled and knowledgeable physician to examine and grade each baby that meets screening criteria.[4] This approach has long been considered the gold standard for screening worldwide, although it has been notably difficult to sustain in resource constrained environments.[5,6] BIO is not scalable due to time and distance constraints between each screening center. As such, only one center in the country performed ROP screening regularly before 2024.[3] Telemedicine-based ROP (TeleROP) screening programs are a promising and proven solution to these constraints in Low- and Middle-Income Countries (LMICs). In comparison to BIO, which necessitates substantial physician skill as well as time for commute and examination, TeleROP offloads the time involved in imaging and face-to-face contact with the patient to non-physicians who are trained to be expert fundus photographers.[7,8] Images are then remotely graded by a physician or some combination with an expert technician grader, increasing the quaternary expert on-patient assessment time. Africa Eye Imaging Centre, a non-governmental organization, was established in Ghana in 2023 to leverage such telemedicine-enabled frameworks to identify and address the most significant causes of retinal blindness. In this report, we share our experiences and challenges in deploying our telemedicine-based ROP screening program. It is our hope that the lessons from this program will enable and inspire similar solutions in LMICs. Leveraging Local and Quaternary Experts with Telemedicine-based Screening We piloted our TeleROP screening program at Komfo Anokye Teaching Hospital (KATH) in Kumasi, Ghana, in partnership with Stanford University (under KATH IRB protocol AMP1/201/24). Screening is performed by trained community health workers (CHWs) using a wide-angle digital fundus imaging (WDFI) camera (Forus 3nethra NEO, Bengaluru, India and the Natus RetCam Envision, Wisconsin, USA). Each eye is imaged using a standardized 6-image protocol (one anterior and five posterior images: optic nerve-centered, superior, inferior, nasal, and temporal). Images are uploaded to a secure picture archiving and communication system (PACS), which manages patient scheduling, image storage, and grading. Within 24 hours, images are graded remotely by a Ghanaian ophthalmologist, with concurrent validation by a Stanford-based quaternary TeleROP (qTeleROP) expert. The TeleROP Severity Score (TeleROP-SS) is utilized to simplify communication of disease severity to caregivers and healthcare providers.[9] Patients identified with incidental findings or treatment-warranted ROP are escalated for repeat screenings or in-person examination by an ophthalmologist as indicated. The key to success in any ROP screening program is both consistent screening and diagnostic accuracy. Telemedicine enables simultaneous grading by both local ophthalmologists and a remote qTeleROP expert, ensuring high-quality assessments. Additionally, simultaneous grading of the same photos creates a unique opportunity for knowledge transfer for long-term capacity building and sustainability of local ophthalmologists who may lack prior experience in ROP grading. Importantly, this telemedicine-enabled capacity-building model can be generalized beyond ROP to other pediatric eye conditions that similarly require highly specialized expertise. For example, telemedicine consultations with remote quaternary experts in retinoblastoma may enable earlier recognition, timely diagnosis, and improved opportunities for globe salvage.[10] Lessons Learned and Programmatic Adaptations 1. Image Grading and Training ○ Initial Strategy: Rely on local ophthalmologists for grading. ○ Challenge: Limited local expertise in ROP grading led to diagnostic uncertainty and variability. ○ Adaptation: Implemented simultaneous grading by local ophthalmologists and a remote qTeleROP expert (with over 20 years of TeleROP experience), providing real-time feedback and structured training modules to enhance local grading capacity and consistency. 2. Photographer Staffing ○ Initial Strategy: Train NICU nurses to perform retinal imaging. ○ Challenge: NICU nurses were already burdened with clinical responsibilities and lacked sufficient time to reliably perform imaging. ○ Adaptation: Dedicated, full-time community health workers were trained to become experts as retinal imaging technicians. They are full-time employees for the ROP-focused non-governmental organization. They have now taken tens of thousands of photographs across thousands of examinations and demonstrated excellence in quality and efficiency. A dedicated ROP NICU nurse is on stand-by for all examinations and joins for any systemically high-risk neonates. Infants are on continuous monitoring for their heart rates and oxygen saturations whilst undergoing the screening process. 3. Screening Criteria ○ Initial Strategy: Adopt standard U.S. screening criteria (≤1500 g birth weight or ≤30 weeks gestational age, with the first exam at 4 weeks of life, or after 31 weeks post menstrual age). ○ Challenge: Early experience revealed treatment-warranted ROP in infants outside traditional U.S. screening criteria, highlighting the need for broader inclusion criteria in the Ghanaian context. Additionally, numerous reports from SSA showed evidence of aggressive ROP (AROP) occurring as early as 2 weeks of life.[11,12] ○ Adaptation: Screening criteria were expanded to ≤2000 g birth weight or ≤32 weeks gestational age, reducing likelihood of missing at-risk infants. The first screening examination was selected to occur at two weeks of life or at discharge, whichever is earlier, to capture at-risk infants before they leave the NICU. More data is needed to validate the current protocol before narrowing of screening criteria can be safely performed. Inpatient and Outpatient Screening with Centralized Referral Centers As the TeleROP program matured and demonstrated initial success at KATH with inpatient screenings, we recognized the need to extend screening services beyond our primary site. Many NICUs across Ghana lack access to ophthalmologists trained in ROP screening, creating significant gaps in care. To address this access gap, we established two additional regional centers in the Greater Accra and Northern Regions. Additionally, we began to expand our referral network to accept outpatient referrals from surrounding, smaller neonatal units. We utilize a bespoke picture archiving and communication system (PACS-ROP) to schedule and coordinate exam grading and follow-up. The PACS-ROP serves as a centralized hub for image storage, grading, and patient management. Structured data collection within PACS-ROP (including zone, stage, plus disease, TeleROP-SS,[9] and incidental findings) facilitates standardized reporting of examination findings to parents and referring NICUs with longitudinal tracking of disease progression. Infants meeting screening criteria from regional NICUs are now routinely referred to our centralized imaging hubs for outpatient screening. This expansion has significantly increased patient volume, improved regional access to timely ROP screening, and facilitated earlier detection and treatment of vision-threatening disease. Additionally, this regional referral model has strengthened collaboration among pediatricians, neonatologists, and ophthalmologists across multiple institutions, fostering greater awareness and education about ROP screening and management. Lessons Learned and Programmatic Adaptations 1. Scheduling, Counselling and Compliance ○ Initial Strategy: Request NICU attendings and nurses to provide parental counseling. ○ Challenge: NICU attendings were responsible for a large patient census (e.g., over 150 beds), and nurses were similarly overextended. As a result, dedicating sufficient time to provide effective counseling on ROP was often not feasible. ○ Adaptation: Dedicated community health workers (CHWs) were trained to take ownership of both inpatient and outpatient counseling. Every qualifying parent receives in-person counseling prior to hospital discharge, emphasizing the importance of ROP screening and follow-up. During hospitalization, screening compliance is exemplary, assuming the infant is medically stable to photograph. At discharge, parents are given an appointment card with the scheduled date for the next screening. Similarly, during outpatient visits, a new card is provided based on point-of-care preliminary grading conducted by the photographer. This immediate, tangible feedback has proven to be a key driver of adherence. In cases where the preliminary grade differs from the final review by an attending ophthalmologist, CHWs proactively contact the family with updated instructions and next steps. Additionally, flexible outpatient scheduling, aligning with regimented weight-checks, and regular communication with caregivers, improved adherence rates. 2. Referral Coordination and Communication ○ Initial Strategy: Rely on informal communication channels and existing relationships for referrals from regional NICUs. ○ Challenge: Informal referral processes led to inconsistent patient follow-up, incomplete documentation, and missed screenings. ○ Adaptation: Formalized referral protocols and standardized documentation processes, including clear eligibility criteria, structured referral forms, and regular communication between referring NICUs and our imaging hub. This ensured consistent patient tracking, improved adherence to screening schedules, and enhanced overall program efficiency. We are also in the process of drafting national screening guidelines in conjunction with the Ghana Health Service to establish ROP referrals as standard of care. Treatment Patients identified with treatment-warranted ROP (Type 1[13] or Aggressive ROP) are promptly referred to local retina specialists for intervention. Treatment options include intravitreal bevacizumab (IVB) injections and laser photocoagulation. Regular communication between local ophthalmologists and qTeleROP experts ensures timely decision-making and intervention. Patients requiring surgical intervention (e.g., advanced retinal detachments) are referred to tertiary teaching hospitals, although financial and logistical barriers remain significant challenges. Lessons Learned and Programmatic Adaptations 1. Treatment Accessibility and Cost ○ Initial Strategy: Rely on only retina attendings and pediatric attendings for all ROP treatments. ○ Challenge: Limited number of ophthalmologists trained in pediatric retinal interventions restricts treatment availability. Additionally, high treatment costs were prohibitive for most families. ○ Adaptation: Implemented structured training programs for ophthalmology retina fellows and pediatric ophthalmology attendings under the supervision of the local vitreoretinal surgeon. For all patients, the senior retinal surgeon treats one eye, and a fellow or attending from another sub-specialty treats the other eye. This expands treater redundancy and expertise in ROP care. This training parallels the existing program for adults, where we have, as noted above, conducted multiple indirect laser teaching sessions on adult patients with proliferative diabetic retinopathy to expand physicians qualified to perform laser. ROP laser skills are taught after building competency for adult laser treatment. 2. Surgical Intervention (Vitrectomy) ○ Initial Strategy: Refer advanced cases requiring surgical intervention to tertiary hospitals. ○ Challenge: Vitrectomy remains prohibitively expensive and logistically challenging for most families, often costing the equivalent of an entire year’s income. ○ Adaptation: Currently, no low-cost solution exists. Families receive detailed counseling regarding expected costs and prognosis, and the decision to pursue surgery is left to the caregivers. This underscores the urgent need for affordable surgical retina solutions in LMICs. Program Outcomes and Future Directions The TeleROP program demonstrated significant cost-effectiveness, with projected costs of 450 US dollars per infant treated, compared to 4,605 US dollars for traditional binocular indirect ophthalmoscopy (BIO)-based screening programs in similar LMIC settings.[14] The success of this pilot initiative has informed plans for scaling the TeleROP model throughout Ghana and potentially other SSA countries. Moving forward, we aim to further decentralize screening by establishing additional imaging hubs in strategic regional locations. This will reduce travel burdens for families, improve screening adherence, and enable earlier detection and treatment of ROP across Ghana. Continued collaboration with regional NICUs, ongoing training of local healthcare providers, and investment in telemedicine infrastructure will be essential to achieving these goals and ensuring sustainable, equitable access to pediatric retinal care across West Africa. Limitations and Challenges Despite early successes, significant challenges remain. High initial capital investment for imaging equipment remains a barrier for widespread adoption. Additionally, secure and intuitive telemedicine platforms are essential for timely image upload, grading, and patient management. Further operational improvements and region-specific data collection are necessary to optimize screening criteria, reduce screening burden, and minimize loss to follow-up. Finally, country-specific screening and treatment guidelines will be established to make ROP screening and treatment the standard of care. Conclusion Telemedicine-based ROP screening (TeleROP) represents a promising, scalable solution for addressing pediatric retinal blindness in LMICs, particularly in SSA. By delegating screening to trained non-physician technicians and leveraging remote grading by local ophthalmologists and international quaternary experts, telemedicine programs can significantly reduce blindness due to pediatric conditions like ROP. Continued investment in infrastructure, training, and operational optimization will be critical for sustainable expansion and long-term success. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.