The big picture

There are some obvious statistics we can look at if we want to know how effective diagnosing ourselves on the internet can be.

For example, let’s look at another, even more recent, Pew study that surveyed people who used the internet to diagnose themselves.

Of the people surveyed, 41% of these online self-diagnosers said that a medical professional confirmed their diagnosis and 2% more said that a medical professional partially confirmed it.

That sounds pretty good!  41% is a lot of people correctly diagnosing themselves.

But as more numbers are looked at, we see a much more complicated picture.

For example, another 18% of respondents said that the medical professional told them their self-diagnosis was incorrect.

Now, this wouldn’t seem so bad.  After all, the person doing the research went to a medical professional and was able to get the correct diagnosis in the end, it would seem.

But what we don’t realize is that there are, in fact, other costs to doing our own research online besides being wrong.

The options

The internet is full of options.  So where do people like you and me go to diagnose ourselves?  And of those options, what are their relative strengths and weaknesses?  Let’s take a look.

Google, Google, Google

Do you use Google to start your personal medical mystery search?  It would seem you’re not alone.

For the vast majority (82%) of Americans who search for a diagnosis online, the first place they go when researching is a search engine like Google or Bing.  Of course, this means we’re putting an enormous amount of trust in these technologies when we’re trying to diagnose ourselves.

We’ve all heard stories of friends or others who have diagnosed themselves through Google, and there have even been news stories that have circulated about such cases, like Tim Stauffer of the San Diego Padres diagnosed himself with appendicitis.

But are these stories representative of the effectiveness of search engines for self-diagnosis?  After all, search engines were not built for diagnosis, they were built to help us find information on the web.  So, how well do they really work?

The problems with search engines:

1. One of the main issues with them is that since their goal is to allow access to all the web, legitimate sites with updated information are often mixed in with illegitimate sites, spam sites, and more.A recent study delved into the numbers of sites that, when accessed through a search engine search, gave accurate information.  The results: only 39% gave “accurate information”. That is a devastatingly low number (and coincidentally is incredibly close to the percentage of patients who accurately diagnosed themselves).
   Clearly, the “openness” of search engines is one of the biggest drawbacks, although we tend to see it as a strength.
2. Another issue with search engines (and diagnosing ourselves online in general) has nothing to do with the end result but the process: anxiety.  The act of self-diagnosis requires that we sort through all the possible diagnoses we might have, which almost always includes quite serious ones.The anxiety created through self-diagnosis has become so widespread that a term has even been coined for it: cyberchondria.  How many people have suffered from cyberchondria?According to the most well-known study about cyberchondria, self-diagnosing on the internet increased the anxiety of 40% of the self-diagnosers.  And this number applies to both people who successfuly diagnosed and didn’t.  In other words, even if we are successful in our self-diagnosis we may have also needlessly increased our inner anxiety.
3. One more issue that ends up affecting all people who search, whether they succeed in their search or not, is the sheer amount of time spent looking.  For example, of the people who have used our site, the average time they have spent trying to diagnosing themselves on the internet was 337 hours.  That’s more than 14 full 24 hour days. We specialize in hard-to-diagnose cases, so our numbers may be higher than normal, but the fact remains that doing all the work ourselves is an enormous drain of energy and time, even if we end up succeeding.

Ask your friends: on Facebook

The next huge way people have begun looking for answers to their medical diagnoses has been in the next logical place you would expect: social media.  About 20% of Americans use social media as a source of health information, according to a recent study, with 94% turning to Facebook.

Social media sites, especially ones as connected as Facebook, offer a lot of potential for diagnosis, and like search engines, there are quite a few anecdotes about people who have been able to diagnose themselves or their family through on it.

But, again, social media sites are not specifically made for diagnoses.  Although there seems to be potential, can we truly rely on the power of social media to find the answers to our medical mysteries?

The problems with social media:

1. All the problems for search engines exist on social media.  The difference between social media and search engines is that instead of your filter being Google’s algorithm, the filter are your friends or whoever you happen to be connecting with.  While this may be effective in specific cases, the same issues you face with search engines you face with friends and others: the chance of getting directed to unreliable sites or getting unreliable advice is high, and anxiety can go through the roof.
2. Another big issue for asking on public sites like Facebook are the privacy issues.  This may be part of the explanation for why people have turned more to search than social.  Of people polled, the majority worry about privacy issues when it comes to sharing medical problems on the net.  This is for good reason: sharing our potential diagnoses with the world may affect our employability, social perception, and more.
3. Accountability: unlike a physician, your friend or acquaintance will not be held accountable for the advice he or she gives you.  This is a big problem, especially when coupled with their lack of medical expertise.  In other words, the “friend” filter is less effective because they are not worried about being held accountable for their advice.

Of course, there are other tools people use for self-diagnosis, but in essence they come down to a choice between an algorithm not built for medical diagnosis and a social media site or forum where there is no accountability for suggested diagnoses, and no way to filter the good responses from the bad.

The next generation of online diagnosis

The question, then, is whether we should be using the internet to diagnose ourselves at all, or whether it’s simply too dangerous.

Since just about every method of online self-diagnosis seems to increase anxiety, have a 60% chance of being incorrect, and reduce our privacy, there seems to be a good case for not using the internet at all for diagnosis, and to simply rely on our eventual visit to a physician.

But there’s a good case for using the internet as a tool as well:

1. There’s no way to stop it – At the end of the day, it seems that most of us are too enticed by the empowerment of self-diagnosis, even if are aware of its limitations.  The fact that it is so widespread makes it clear that using the internet as a tool for diagnosis is not going away anytime soon.  It would be best, perhaps, to find and develop better tools, rather than try to convince people that they’re wrong for using it.
2. The potential – The fact that so many people have successfully self-diagnosed, without training or degrees, and that there have been so many cases of miraculous diagnoses provided either through a search engine or social media site, is a testament to the potential of the internet as a diagnostic tool.  A 41% success rate is great when we take into account that people are using sites that are not even built for diagnosis.  What if a site, or sites, were able to harness the potential of these two platforms?

CrowdMed: tapping the potential

CrowdMed was essentially created with these issues in mind.  When it was developed, the potential of social media and algorithmic tools to help a person reach a diagnosis were clear to us, and we realized that if a tool was built specifically to aid a person on their quest for a diagnosis, we could take the best of these tools and bring out their potential.

What CrowdMed does: CrowdMed is essentially a combination of a private social network site and a search engine.

1. Social network – As the name of our site implies, we use the “crowd” to help a person find a diagnosis.  In our case, the vast majority of the crowd aiding a person in their quest for a diagnosis (what we call “medical detectives”) are medical experts.  Like the rest of the world, we see the potential in having a lot of people, especially medically trained professionals, giving input on a case.  But how do we mitigate the negative effects of social media?
2. The algorithm – As we described, we also use a similar method to our site as Google uses.  However, whereas Google uses its algorithm to rank web pages on the internet, we use our algorithm to rank our medical detectives, as well as the diagnoses they provide.  The goal with this algorithmic technology is to address the issues that social media presents: that of a lack of accountability and no filter.As our medical detectives contribute to our site, our system rewards them for their success (or doesn’t, because of lack thereof).  The medical detectives that succeed the most move up in our site, have more say, recognition, and input on arriving at each person’s diagnosis.  Google uses a system called “PageRank” that is similar to this.
   We then rank the suggested diagnoses, based on popularity, rank of the medical detectives who submitted their suggestions, and more.

It is in these ways that we are able to address the potential as well as the challenges of social media and search engines.  Whereas search engines rank the results of your searches by factors that do not necessarily correlate to an accurate diagnosis, our system ranks specifically around the factors that will help our patients find the answers they are looking for, while also taking advantage of the varied input of social media platforms.

The Results:

Of course, all the theory in the world can’t mean anything without results to back it up.

1. Accuracy – Despite the fact that we deal with some of the hardest medical cases in the world, the amount of our users who claim we helped them get closer to a diagnosis is now at a steady 60%.  That means that we do 33% better than a population that looks up more common ailments than our system normally deals with.
2. Time – The combination of social and algorithmic help on our site means that the user does not have to do the very jobs that take so long: the search and the analysis.  Their main job? Stay in touch with their medical detectives.
3. More results – You can see more of our results in our last two data blog posts: Proof The Internet Can Help Diagnose You and Are Doctors The Only Ones Qualified To Diagnose Us?

 

Are you in need of a diagnosis? Click here to submit your case.

Want to be part of our medical detective community? Click here to join.