Why this doctor wants you to consult with 'Dr. Google' to help diagnose your symptoms

For someone experiencing a medical symptom, such as a sudden pain or fatigue, the internet can provide all kinds of answers. And while many doctors discourage their patients from consulting "Dr. Google," a new study proves it can be a reliable source. 

"When you use the internet to search, you actually get better at diagnosing yourself," lead researcher Dr. David Levine told As It Happens host Carol Off. 

Levine sees all kinds of cases as a general internist at Brigham and Women's Hospital in Boston. And as a researcher, his experience helps him reimagine how his patients can receive quality care. For him, that endeavour involves technology and decentralizing that care.

At the same time, as advanced as the internet has become, and as smart as search engines can be, he still hears his colleagues giving the same advice.

"We do hear that all the time, right? 'Don't go to the internet. Just come talk to me,'" he said.

Searching the internet for medical advice can lead a person to think about the worst-case scenarios. It can also lead to misplaced anxiety — known as cyberchondria — and miscommunications with actual doctors.

But there hasn't been enough research on this kind of internet search behaviour that proves it to be harmful, Levine said.

So he teamed up with researchers at Harvard to come up with some empirical measurements. They studied 5,000 people in the United States to see how the average person fares on an internet search for medical advice.

The findings were published this week in the journal JAMA Network Open.

"Probably, we should be allowing our patients to search for their symptoms," Levine said. "The worst-case scenario is that you end up at the doctor's office, just like you were going to."

Searches led to better diagnoses

The researchers asked 5,000 participants to diagnose a case, supposing that their loved one was ill. They gave the participants a list of symptoms that ranged from mild to severe, describing common illnesses including viruses, heart attacks and strokes.

Along with a diagnosis, the researchers asked them how quickly the patient would need treatment, as well as how worried they were about the case.

They asked the participants to do all of this without checking online.

"We were able to then say, 'Thanks for those answers. Now we want you to go search the internet,'" Levine said. 

In the end,  Levine said the results were  "very interesting."

"They did OK on diagnosis — they were about right half the time. They did pretty darn well on triage," he said. "Turns out, though, that when they were able to use the internet, they got about five per cent better at diagnosis."

The study also offered a glimpse into how people come up with their diagnoses and navigate the internet.

"About half of people use just the search engine alone. About just under half the people ended up going to a health specialty site, something like WebMD or Medline. And then a very, very small number of people ended up going to social media [and] news sites," Levine said. 

Since the study focused on human behaviour, there were more factors for the researchers to consider. Of the 5,000 American participants, most were white, at an average age of 45, with an even gender split. 

In their analysis, the researchers grouped the participants into different skill levels based on their demographics.

"Turns out women were much better than men at diagnosis ... folks who were older were better than folks who were younger ... [and it] also looks like folks that had more health-care experience or were sicker, in general, were also better at diagnosis," Levine said. 

"This is an exceptionally interesting part of this study that I think surprised us a lot."

Levine does not have an explanation as to why these patterns came up in the study, but he wants to understand them in order to help the average person succeed at a self-diagnosis.

"We are actually really interested in taking that … knowledge and trying to bottle it up into some really smart machine algorithms, not just for doctors and nurses … but for regular old folks who are searching the internet," he said.

"Maybe we can get a lot better at helping them get their diagnoses and triage right, so that it's not just a five per cent improvement ... and that we're actually getting close to being just as good as doctors without even needing to involve them."

Written by Mehek Mazhar. Interview with Dr. David Levine produced by Chloe Shantz-Hilkes.