Nonexpert Clinic-Based Screenings Effective for Disease Detection

Researchers say that graders outside of ophthalmic clinics demonstrated a high level of effectiveness in detecting eye disease on fundus photographs.

Elevated rates of ocular disease and an aging patient population, many of whom live in rural areas without easy access to eye care, have left ophthalmologists spread thin. Due to this growing demand, some clinicians are turning to outside sources to perform tasks such as diagnostic photo grading. Although non-expert interpretations of fundus photography and other diagnostic imaging tests may not seem ideal, a study published in the American Journal of Ophthalmology is demonstrating that, for screening purposes, this option can be effective, especially in the cases of diabetic and thyroid eye diseases.

The investigations reviewed an eye disease screening program at a general medical clinic, diabetes clinic, and thyroid clinic at Chiang Mai University Hospital in northern Thailand. The participants were culled from diabetes, thyroid, and general clinics.

Individuals reviewed included patients at least 50 years old. Any participants with a positive finding on fundus photography, visual acuity testing, or tonometry were referred to an ophthalmology clinic. Additionally, a 10% random sample with negative results in both eyes were referred for evaluation by an ophthalmologist.

Of the screenings, 229 participants of the total enrolled (n=889) had at least 1 positive test, and 189 of them presented for an ophthalmic examination. Of 76 patients with normal screening results who were referred for an examination, 50 attended. The study found that fundus photography screening had the highest yield for diabetic retinopathy (sensitivity 67%), visual acuity screening for cataract (89%) and intraocular pressure (IOP) screening for glaucoma and glaucoma suspects (sensitivity 25%). At least one of the diseases of interest was detected in 25% of participants from the diabetic clinic, 34% from the thyroid clinic, and 21% from the general clinic. 

The researchers explain that the present study employed nonexpert staff to take and interpret fundus photographs, which sets it apart from previous research. The training of these personnel to recognize abnormalities, they note, allows the bypassing of the logistics and cost associated with a reading center.

“The photogrades assigned by the nonexpert staff in this study had relatively high predictive value for the diseases of interest, suggesting this screening program model may be possible even in areas with relatively low ophthalmologist support,” the study explains. “

Nonexpert point-of-screening image interpretation is able to detect several eye diseases at once, thus providing an advantage over currently available automated systems.”

Study limitations included possible bias due to the higher ratio of attendees who failed a portion of their screening (83%), compared with those who presented a normal screening (66%). Additionally, the study did not include anterior segment imaging to assess for cataract, due to the high probability that it would be detected in a measure of acuity.


Ausayakhun S, Snyder B, Ausayakhun S, et al. Clinic based eye disease screening employing non-expert fundus photo-graders at the point of screening: diagnostic validity and yield. Am J Ophthalmol. Published online April 2, 2021. doi:10.1016/j.ajo.2021.03.029.