Family physicians can learn to identify patients with obvious retinopathy with the right equipment, but they still need to refer patients with diabetes to eye care specialists. That, according to findings presented last last week at the annual meeting of the American Association of Family Physicians (AAFP).
James M. Gill, MD, MPH, from Christiana Care Health Services, in Wilmington, Delaware, United States, and colleagues selected 11 family physicians from the Delaware Academy of Family Physicians (AFP) to participate in a 4-hour training workshop. The workshop was given by an education specialist and ophthalmologist and designed to help the physicians screen for diabetic retinopathy by using a new non-mydriatic ophthalmoscope that allows a view of the retina that is three to five times wider than that of a standard direct ophthalmoscope.
Following the workshop, the physicians examined 29 patients with diabetes who were selected to represent a spectrum of retinal abnormalities, from no disease to severe disease. They assessed two retinal fields in each of the patients’ eyes as "normal", "abnormal" or "cannot evaluate", based on the standardized criteria they learned in the workshop. The physicians’ findings were compared to the ophthalmologist’s evaluation, using the same ophthalmoscope and criteria, as well as the results of comprehensive indirect ophthalmoscopy and retinal diagrams in the ophthalmologist’s office.
Overall, the physicians rated 41 percent of retinal field as "normal", 14 percent as "abnormal" and 46 percent as "cannot evaluate". In contrast, the ophthalmologist, using the same tool and criteria, rated 43 percent as "normal", 21 percent as "abnormal’ and 36 percent as "cannot evaluate". With the indirect criteria, the ophthalmologist rated 46 percent as "normal", 54 percent as "abnormal" and none as "cannot evaluate".
Compared to the ophthalmologist using the same criteria, the physicians’ sensitivity with respect to detecting retinal abnormality was 36 percent, and their specificity was 55 percent. When both the physicians’ "abnormal" and "cannot evaluate" scores were collapsed together to form a "refer" category, as opposed to the normal or "do not refer" category, sensitivity increased to 70 percent. Compared to the indirect criteria, and again using the "refer" and "do not refer" categories, physicians’ sensitivities were 75 percent, and their specificities were 59 percent.
According to Dr. Gill, physicians’ accuracy rates "were not high enough to bypass the current standard of sending all [diabetic] patients for an eye exam. However, the methods [the physicians learned in the workshop] are much more accurate than what we have today, and actually they’re very efficient."
"Probably the key issue is that, in patients for whom retinopathy is identified by their family physician, there’s probably an increased likelihood they will go to an eye specialist because you’re telling them they have a problem," he concluded.