COVID-19’s online symptom-checking systems are almost always wrong – only a third of their diagnoses are correct. This is stated in a study by scientists from Edith Cowen University, published in the journal Medical Journal of Australia.
During the study, scientists analyzed 36 international mobile applications designed to answer user questions about the presence or absence of COVID-19 symptoms. The user describes the symptoms of the application, and it recommends that you consult a doctor if the symptoms are worrying.
Such services, according to the developers, should unload the doctors. However, in reality they are too often mistaken – only in a third of cases of the application do they give the right recommendations for going to the doctor, the analysis showed.
In particular, applications give accurate results in only 36% of cases, and only 49% of users receive the correct recommendations about the place and time of seeking medical help, the study said.
“Although the idea of using digital tools to test your symptoms seems tempting, in most cases, existing applications are unreliable at best and dangerous at worst”.
Michelle Hill, lead author of the research
Previously, scientists created two chatbots that they say help patients check their symptoms for coronavirus infection. They will be able to advise the majority of applicants and reduce the burden on medical workers.