• A new study published in The Medical Journal of Australia suggests that online symptom checkers only provide accurate diagnoses 1/3 of the time.
  • A team of researchers from Edith Cowan University (ECU) evaluated the accuracy of 36 different symptom checkers, inputting symptoms of 48 different “patients.”
  • They pulled clinical information for their patients from training material for health care providers as well as other reliable medical resources.
  • In addition, a “correct” diagnosis was given for each patient’s symptoms by a group of medical professionals.
  • The research team considered two main measures: correct diagnosis and correct triage advice provided by the online symptom checkers.
  • The team found that the symptoms checkers produced the correct diagnosis as the top result 1/3 of the time and within the top three results about ½ of the time.
  • Additionally, the triage advice produced by symptom checkers almost always recommended more urgent care than necessary.

Quick Summary

According to a new study “The quality of diagnosis and triage advice provided by free online symptom checkers and apps in Australia,” published in The Medical Journal of Australia finds that online symptom checkers give us an accurate diagnosis only 1/3 of the time. In other words, googling or otherwise researching our health symptoms probably isn’t the best idea.

Investigation

Many of us are guilty of googling our symptoms, looking them up on WebMD, or submitting them to a symptom checker. This can lead to health anxiety or hypochondria, which is why we’re often advised against doing so—but how helpful (or harmful) can researching our symptoms actually be? That’s what Edith Cowan University (ECU) sought to find out.

Their study evaluated the accuracy of 36 different symptom checkers, inputting symptoms of 48 different “patients,” so to speak. The symptom checkers then provided diagnostic advice, triage advice, or both. The researchers pulled clinical information for their patients and their corresponding illnesses from training material for health care workers as well as other reliable medical resources. Additionally, the “correct” diagnoses for each patient’s symptoms was given by a group of medical professionals.

Findings

The research team evaluated two main measures: correct diagnosis advice as well as correct triage advice given by the symptom checkers. For the former measure, they used the following definitions:

  • An accurate diagnosis is defined as: including the correct diagnosis as the top result or among the top results (e.g., in the top 3 or top 10 potential diagnoses)
  • An inaccurate diagnosis is defined as the correct diagnosis not being included in the top 10 results.

They concluded that the diagnostic advice given by symptom checkers varied in quality; however, they only produced the correct diagnosis as the top result 1/3 of the time and within the top three results about ½ of the time. Additionally, the triage advice produced by symptom checkers almost always recommended more urgent care than necessary. For example, for a medical issue that is widely considered to be non-urgent, these symptom checkers would deliver a “urgent” or “emergency” result instead.

Implications

More people are scouring the internet than ever before to better understand their health symptoms and to seek health care advice. Unfortunately, symptom checkers (and other online resources) often provide inaccurate, incomplete, or otherwise unsuitable information for these individuals.

Limitations

  • This research focused solely on user habits of symptom checkers in Australia.
  • Researchers investigated the most-used symptom checkers, but could have missed some widely used by Australians.
  • The “patients” were fairly simple, as they did not have co-morbid conditions, and therefore did not possess the level of complexity of real patients.

Source

Hill, M. G., Sim, M. & Mills, B. (2020, May 11). The quality of diagnosis and triage advice provided by free online symptom checkers and apps in Australia. The Medical Journal of Australia. Retrieved from https://onlinelibrary.wiley.com/doi/full/10.5694/mja2.50600

(2020, May 17). New research finds ‘Dr. Google’ is almost always wrong. NeuroscienceNews. Retrieved from https://neurosciencenews.com/dr-google-wrong-16408/