‘It is no substitute’: Ophthalmologist warns against using ‘Dr. Google’ to self-diagnose
More than ever, Americans are avoiding the expense of doctors' visits and turning instead to the internet, where online symptom checkers (OSCs) like WebMD attempt—oftentimes incorrectly—to diagnose patients based on a series of questions.
Writing in JAMA Ophthalmology, Rahul N. Kurana, MD, said as many as 6 out of 10 patients query “Dr. Google” about an immediate health issue before asking a licensed physician the same question. OSCs serve to self-diagnose and triage patients, and can factor significantly into an individual’s decision to either head to the emergency room or not seek care at all.
“To accomplish these goals, the assumption is that the OSC program is accurate,” Khurana, of Northern California Retina Vitreous Associates and the University of California, San Francisco, said in JAMA. “If that is not the case, the tools may lead to inappropriate use of health resources and poor outcomes.”
Khurana referenced a recent study that tested the accuracy of WebMD, a widely accessible and free OSC, for ophthalmic diagnoses. The research team used 42 validated ophthalmic clinical vignettes involving both emergent and nonemergent conditions to assess the tool, finding WebMD’s primary diagnosis was accurate in just 26% of cases. The correct diagnosis was included in the top three differential diagnoses in 38% of cases, while urgent cases like retinal detachment and acute angle-closure glaucoma were diagnosed incorrectly 79% of the time.
The finding that nearly four-fifths of urgent cases were diagnosed incorrectly is particularly concerning, Khurana said, because it could lead to an urgent case being triaged as nonurgent, limiting a patient’s access to care.
Another study that measured the accuracy of OSCs for diagnosing other medical conditions found the online tool reached a correct diagnosis in 34% of cases and was in the top three differential diagnoses in 51% of cases. The numbers were bleak in comparison to experienced physicians, who achieved a correct diagnosis 72% of the time and predicted a diagnosis in the top three differentials 84% of the time.
“With advancements in artificial intelligence, OSCs are likely to improve,” Khurana wrote. “Empowering patients with more information leads to greater engagement and better health outcomes. However, the accuracy must be validated before trusting these online resources.”
Khurana cautioned patients to exercise caution in using online resources to self-diagnose—especially since incorrect diagnoses can lead to unnecessary anxiety and inappropriate triage advice.
“At this time, the internet may be easy to access, but for a correct ophthalmic diagnosis, it is no substitute for an ophthalmologist,” he said.