It has to be said and a recent study makes it clear that you should not rely on “Symptom Checkers’ online in search for an accurate diagnosis. Given the percentage of accuracy within the cases presented to 23 different websites that handled online diagnosing, there are better ways to find out what health issues you might be facing. Such as a doctor.
The overall accuracy percent was somewhere around half, which is mostly a sign of clear hit-or-miss that most do not wish to chance upon when it comes to their well being. However, over 30% of Americans are reportedly resorting to the internet to provide them proper treatment and information on what diseases they could have.
In a study done by Harvard Medical School, twenty three ‘symptom checkers’ online have been put through the test and the results were worryingly inaccurate. Some were of more famous brand names such as Mayo Clinic, others lesser known.
Different types of symptoms were inserted into the fields of the alleged patients, based on true possibilities of real life diseases or conditions. Most websites provided a list of possible diagnoses instead of just one answer, taking the information and running it through an algorithm that will help them zero-in on the problem.
It works based on the likelihood of the disease matched-up with the input symptoms.
The percent of the right diagnosis presented on first place was of only 34% while most websites had an accuracy of 51% by reaching the right answer in the top three choices. It’s not the most hopeful statistic as there is a 50/50 chance to get properly diagnosed somewhere within the first three options. And most are not medically trained to decide which might be better.
When prodded even further, the chances of reaching the right conclusion in the first twenty presented options was of only 60%.
The study implied 45 different medical cases, among which 26 were common diagnoses while 19 were uncommon. Separated even further to better test the accuracy of online ‘symptom checkers’, 15 of the cases would have required emergency care, 15 were grave, but did not require it, while the remaining 15 could have been fixed with self care at home.
They found that that the proper advice was given in only 57%, while the majority exaggerated in the issue’s urgency and were more prone to advising for emergency medical care. That could perhaps be a redeeming quality in spite of its flawed nature. It is indeed better to be safe than sorry.
Advising people to instead urgently consult a doctor even when it’s not needed is better than the riskier alternative of dismissing it as nothing and allowing for your condition to worsen.
The conclusion remains that while Google, M.D. is not a reliable diagnosing tool, there is a way for these programs to be improved upon. With the technology we have available today, the percentage can be improved, but it needs some very important tweaks.
However, it’s not likely medical health consultants will become obsolete anytime soon.
Image source: medicaldaily.com