Online symptom checkers far from accurate

Online symptom checkers and apps only provided the correct diagnosis about a third of the time, but could still be worthwhile compared to Dr. Google, a new study suggests.

Services often send people for care when staying home would be better

Online symptom checker services often send people for care when staying home would be better. (David Goldman/Associated Press)

Online symptom checkers and apps only provided the correct diagnosis about a third of the time, but could still be worthwhile compared to Dr. Google, a new study suggests.

Unlike internet search engines that can lead users to confusing and unsubstantiated information, symptom checkers hosted by medical schools such as Harvard, hospitals such as the Mayo Clinic and government agencies such as the U.K.'s National Health Service are more sophisticated.

People are asked to list their symptoms and then a computerized algorithm suggests whether to seek medical care immediately, like for a heart attack or stroke, visit a doctor in the next few days or just rest at home.

Researchers at Harvard Medical School put 23 different symptom checkers through the paces based on 45 clinical vignettes used to teach and test medical students.

Symptom checkers are part of a larger trend among both doctors and patients to use the internet for services ranging from seeking advice to e-visits and health apps.

Overall, the symptom checkers gave appropriate triage advice, like whether to get to a doctor quickly, in 57 per cent of cases, senior author Dr. Ateev Mehrotra, an associate professor of health-care policy and medicine at Harvard Medical School and his colleagues reported in this week's issue of the BMJ, formerly known as the British Medical Journal.

"Their performance is not perfect by any means and should not be treated as gospel," Mehrotra said in an email. "But they could provide better information than simply Googling the internet with your symptoms."

For cases where people should go to emergency, the symptom checkers correctly recommended that in 80 per cent of cases. In comparison, previous studies suggested search engines correctly pointed to emergency medical treatment 64 per cent of the time.

Mehrotra's sense is triage is the most important aspect of the tools because its often a key question for people. For a patient with fever, headache, stiff neck and confusion, knowing whether it's from meningitis or encephalitis for example isn't nearly as important as being told to get to an emergency department right away.

Like the telephone triage lines they may eventually displace, symptom checkers tended to send people for care when staying at home might be reasonable, the researchers found.

Checkers with the most accurate diagnoses (Isabel, iTriage, Mayo Clinic, and Symcat) were not on the list of the programs that did the best job of recommending the appropriate level of care for a given case (, Steps2Care and Symptify). 

The symptom checkers gave the correct diagnosis first 34 per cent of the time and listed the correct diagnosis within the top 20 in 58 per cent of the evaluations.

Doctors are thought to have a diagnostic accuracy of 85 to 90 per cent and studies comparing the accuracy of telephone triage recommendations to those of in-person physicians ranged from 61 per cent to 69 per cent, the research team said. There wasn't a comparison of doctors and symptom checkers.

While symptom checkers are not likely to go away, they could encourage patients to worry unnecessarily. A concerning diagnosis could fuel cyberchondria — escalated anxiety from self-diagnosing online. 

Symptom checkers are still in their infancy. The researchers suggested ways to improve their performance, such as including local data on incidence of infectious diseases in the community or data from electronic medical records.

The study was funded by the U.S. National Institute of Health. 


To encourage thoughtful and respectful conversations, first and last names will appear with each submission to CBC/Radio-Canada's online communities (except in children and youth-oriented communities). Pseudonyms will no longer be permitted.

By submitting a comment, you accept that CBC has the right to reproduce and publish that comment in whole or in part, in any manner CBC chooses. Please note that CBC does not endorse the opinions expressed in comments. Comments on this story are moderated according to our Submission Guidelines. Comments are welcome while open. We reserve the right to close comments at any time.

Become a CBC Member

Join the conversation  Create account

Already have an account?