Google hasn’t quite replaced the doctor yet. A recent study out of Harvard Medical School finds that the web isn’t a very good way to figure out what ails you. The study showed that even sites touted as sophisticated symptom-checkers were extremely unreliable.
Many popular symptom-checkers, such as WebMD, respond to symptoms by suggesting many — sometimes dozens — of possible ailments, with the most likely condition at the top.
However, the study found that the sites only yielded the correct diagnosis as the top result in 34 percent of cases. In 42 percent of cases, the correct diagnosis was not even listed in the first 20 results. (The researchers asked respondents to enter different sets of symptoms linked to various illnesses into 23 symptom-checking websites to gauge site accuracy.)
The websites weren’t much better at assessing triage, particularly for less serious medical conditions. When the correct diagnosis referred to an “emergent” case, the sites assigned the appropriate triage 80 percent of the time, compared to only 55 percent of non-emergent cases. In cases that required only self-care, the sites evaluated the triage correctly only 33 percent of the time.
The study thus concluded that symptom-checkers are often encouraging people to seek emergency care when it is not necessary. Four of the sites — iTriage, Symcat, Symptomate, and Isabel — always recommended people to seek professional care rather than self-care.
There are ways such sites could improve their performance, including by incorporating local epidemiological data into their algorithms, the researchers suggested. Sites could also be far more effective if they included clinical data from the individual’s electronic health records.
Although the researchers urged caution when using symptom-checking websites, they stopped short of suggesting that individuals should give up on online tools entirely.
“Symptom checkers may, however, be of value if the alternative is not seeking any advice or simply using an internet search engine,” they wrote.