Improving teleophthalmology protocols in high-occupancy, high-demand, and remote settings

To the Editor:

Healthcare systems in Ecuador and other countries grapple with challenges such as high occupancy rates, surging demands, and limitations in precise diagnosis and monitoring, especially in remote areas. The Galapagos Islands (GIs) encounter similar hurdles, with a rising demand for diagnosis and a high prevalence of degenerative ophthalmological diseases that necessitate close monitoring (Fig. 1). Our goal was to integrate two pivotal articles from Eye to tackle these needs in the GIs. Nonetheless, limitations persist concerning the accuracy of the proposed devices and the availability of management options for this challenge. Meshkin et al. (2023) demonstrate a protocol that combines a questionnaire, clinical information, as well as photos captured by a smartphone and a Canon non-mydriatic fundus camera, within an emergency department for the purpose of triaging cases. However, there is a risk of misclassifying critical patients due to imprecise technology and other factors [1]. Balaskas et al. (2023) highlight the growing interest in remote vision monitoring apps and technologies to detect early signs of disease and improve patient care. Yet, limited evidence supports their effectiveness and usability, which may vary based on the devices and settings used [2]. Both studies emphasize the need for accurate diagnoses, user-friendly software, and simple implementation. While we agree with the authors, concise protocols are still needed to enhance the diagnosis and follow-up of ophthalmological diseases, particularly in remote settings like the GIs.

Fig. 1. The Galapagos Islands are in urgent need of improved teleophthalmology approaches.

Similar to other geographically isolated areas worldwide, the Galapagos Islands face similar healthcare challenges, which makes it an ideal location to test new teleophthalmology protocols. Utilizing available public data from Sistemas Medicos de la Universidad San Francisco de Quito (SIME, USFQ), we identified three prevalent impairing conditions that significantly impact the population: Age related Macular Degeneration (AMD), Diabetic Retinopathy (DR), and Hypertensive Retinopathy. Additional imaging findings suggest the possibility of other diseases, but confirming them requires further diagnosis, either through the development of a new device or in-person physical examination. These findings emphasize the importance of future research and healthcare strategies for remote regions. Created with BioRender.com.

In this letter, we propose a protocol that relies on an initial patient assessment to identify individuals at risk and gather relevant factors influencing their condition. Subsequently, rapid screening is conducted by capturing eye photos using smartphones and tablets, facilitated by AI and a remote specialist (RS). For further confirmation, an advanced precision camera is employed to assess both the external part of the eye and the fundus. Once the data is compiled, it is analysed by AI and validated by an RS (Fig. 2). We believe that combining these multiple parameters will significantly improve the efficiency of patient assessment, resulting in accurate diagnoses. This proposal represents a valuable strategy to consider.

Fig. 2. Optimized protocol for high occupancy, demand, and remote settings.

This figure illustrates a protocol designed for application in high occupancy, demand, and remote areas where access to medical care is overwhelmed and limited. The protocol emphasizes the importance of creating a comprehensive medical history to gather patient information, enabling healthcare staff to assess the risk level. A rapid screening technological device is then utilized to capture anterior segment eye images. Following the rapid screening, an initial evaluation is provided by an AI system and a remote specialist. If necessary, a confirmatory assessment is conducted for patients with suspected high-risk diseases or those with an undetermined diagnosis, followed by a subsequent evaluation by both the AI system and a remote specialist. This protocol aims to minimize the risk of misclassifying critical patients due to imprecise technology and other factors. Furthermore, it incorporates a double-check system involving both a specialist and AI to generate evidence supporting the effectiveness and usability of automated diagnosis. Created with BioRender.com.

Author contributions

The manuscript was written, reviewed, and data curated by SC, MBT, GC, CR, CG, JO, and AC. AC, CG, and JO contributed to revising of the manuscript in terms of information and applications. AC supervised, mentored, administered the work, and conceived this letter. The final form of the letter was critically reviewed and commented on by SC, MBT, GC, CR, CG, JO, and AC, leading to its approval.

Funding

Sistemas Médicos de la Universidad San Francisco de Quito, SIME - USFQ. Escuela de Medicina, Colegio de Ciencias de la Salud COCSA, USFQ, Quito, Ecuador. These funding sources had no role in the design of this study and will not have any role during its execution, analyses, interpretation of the data, or decision to submit results.

Competing interests

The authors declare no competing interests.