Implementation of a Teleophthalmology Screening Network Using a Novel Handheld Fundus Camera on a Remote Island: Physician and Patient Perceptions from Real World Experience (Preprint)

BACKGROUND While teleophthalmology is gaining traction in recent years, it is taking center stage in the COVID-19 pandemics. However, most hospitals are yet ready due to a severe lack of real-world experience. Furthermore, there are limited number of studies evaluating the telemedicine application on remote islands. OBJECTIVE To evaluate the real-world clinical and referral accuracy, image quality, physician perceived diagnostic certainty, and patient satisfaction of a telemedicine eye screening using a novel handheld fundus camera in a rural and medically underserved population. METHODS This prospective study enrolled 176 eyes from a remote island. All participants underwent a comprehensive ophthalmic examination. Nonmydriatic retinal images obtained using a handheld fundus camera were reviewed by two retinal specialists to determine the image quality, diagnosis, and need for referrals. The agreement of diagnosis between image-based assessments was compared to that of binocular indirect ophthalmoscopic assessment. RESULTS Image quality of fundus photograph was considered ideal or acceptable in 97.7% and 95.5% eyes. There was considerable agreement in diagnosis between the indirect ophthalmoscopic assessment and image-based assessment by two reviewers (Cohen’s kappa=0.80 and 0.78, respectively). Likewise, substantial agreement in referrals was achieved. The sensitivity for referable retinopathy from the two reviewers was 78% [95% CI, 57%-91%] and 78% [95% CI, 57%-91%], whereas specificity was 99% [95% CI, 95%-99%] and 98% [95% CI, 93%-99%], respectively. For physician perceived certainty in diagnosis, 93.8% and 90.3% were considered either certain or reliable. Overall, 97.4% of participants were satisfied with their experience and greatly valued the telemedicine service. CONCLUSIONS The novel fundus camera-based telemedicine screening demonstrated high accuracy in detecting clinically significant retinopathy in real-world settings. It achieved high patient satisfaction and physician perceived certainty in diagnosis with reliable image quality, which may be scaled internationally to overcome the geographical barriers under global pandemic. CLINICALTRIAL N/A