Crowdsourcing health research: A new chance for patients and tech people

Olivier Mirat
7 min readMar 16, 2018


Being stuck with a disease or an injury is hard. In the modern age of the Internet and of technology, there might be a way to turn those painful experiences into something positive: crowdsourcing health research. MyAIguide, for instance, is an open platform to crowdsource health research and to help participants with their health issues, mostly through the creation of AI assistants. For the programmers, designers, marketing experts, health professionals and investors, MyAIguide’s open platform also represents an exciting opportunity to make a positive contribution to the world. Let’s take a look at what crowdsourcing health research has to offer, and why one might or might not want to take part in it.

What’s the traditional way of doing clinical research?

In traditional clinical research, scientists design an experiment to try to answer some specific question about a disease, and then recruit patients who must visit a hospital, clinic, research center or doctors’ office to participate. The advantage of this kind of research is that the scientists in charge of the project usually know what they’re doing, they are part of an establish research organization, they have to follow specific ethical protocols to ensure patients safety, and they have access to advanced technologies [1]. Participating in this kind of research is usually not so easy. You need to find a study that takes place at the right time, that you qualify for, and that is close to where you live. There’s also usually a limit on the number of participants. If a drug is being tested on you, you may or may not be able to continue using that drug after the trial ends depending on the situation [2]. You do however get paid for your contribution.

The Internet, Technology, and Disruption:

For better or worse, we live in an age were everything is being disrupted by technology. Traditional taxi drivers are facing the competition of GPS-enabled smartphone apps, you can buy your books, music and diapers online, find someone to walk your dog on the internet, and a lot more.

Perhaps because of its high amount of regulations, the health sector has been more resistant to disruption. But this is now rapidly changing in both the healthcare sector and clinical research [3].

Crowdsourcing health research:

With crowdsourced health research, the idea is to let anyone participate to a study over the Internet, usually without the need of visiting a specialized facility like a hospital. This usually consist in having participants track health information about themselves through a mobile app or a website. In some cases, participants also get their blood drawn and analyzed, their genome or microbiome analyzed, or wear various fitness trackers. A big advantage of this kind of research is that many people can participate without much restrictions, which can result in collecting huge amounts of data that traditional research usually can’t gather. An other interesting benefit is that the feedback loop to the patient can be shortened: in some cases, the patient can benefit from his own contribution much faster. This video from the Apple Research Kit [4], provides a good example of this: participants provide data about autism and in return get a self-diagnosis tool for their child.

Similarly, MyAIGuide is currently collecting data about repetitive strain injuries and chronic pain, and already provides personalized health recommendation to people who participate in the study.

So if you want to turn your injury or your painful condition into something positive but you can’t find a traditional clinical study to participate in, a crowdsourcing study might be the best option. However, be warn, this kind of activity has its pitfalls and potential risks.

The pitfalls and dangers of health research crowdsourcing:

Needles don’t live on the internet:

The biggest and most obvious pitfall of crowdsourcing health research is that needles don’t live on the internet! In other words, the technologies that participants have at their disposal in the comfort of their homes is more limited than what traditional research labs have. I don’t know about you, but I personally don’t have a MRI machine in my garage. Technology usually does get more affordable and more accessible with time, so the quality and quantity of technologies accessible by common mortals should only get better with time.

Tricky ethical considerations and safety threats:

In Western countries, traditional clinical research is heavily regulated, mostly to protect patients. This is however unfortunately not necessarily always the case in developing countries (see this TED video). For crowdsourced studies, protection of patients can sometimes be an issue.

The Apple Research Kit experiments are probably slightly less protective of patients compared to traditional trials: there can be some issues with data protection, minors enrolling in studies or with making sure that patients really understand what they are getting into [5]. But the ResearchKit really doesn’t deserve much blame, and every app using it now needs to go through an Institutional Review Board to verify that the study is ethically acceptable, just like traditional clinical trials have to [6].

In some cases, crowdsourced health research can also be organized by for-profit corporations. 23AndMe, for instance, is a very well-known direct to consumers genotyping company: if you pay them, they will send you a tube to spit in, you send them back the tube and they give you information about your DNA. By using this crowdsourced approach, they’ve managed to compile the largest DNA database ever created, and they’ve used it for very worthy research purposes [7]. However, in order to create this huge database, they initially advertised themselves, for the most part, as a company that could tell you what disease you were at risk of developing. This might have been a misleading way of presenting things and the FDA initially forced them to shut down their health section, and then let them reopen it under much stricter conditions [8].

Don’t think you’ll always get that much out of it in the short-term:

An average 23andMe user, for example, usually won’t learn anything really important from the service, the results are usually mostly just amusing, as shown in this review. Unless you want to check if you might conceive a child with an hereditary disease [9], or if you are looking for relatives [10].

Another great example comes from PatientsLikeMe, a health social network: in 2008, a small research team published a paper suggesting that a drug called Lithium could slow the progression of ALS. Hundreds of members of the social network then started testing the drugs on themselves, and the study eventually got published in Nature [11]. Unfortunately, the result was that the drug was ineffective. Since then, the data collected by PatientsLikeMe has lead to many more publications [12], none of them seem life changing for patients in the short-term, but science usually progresses in small steps, so this is pretty much normal.

Looking at the crowdsourced studies proposed by the Apple Research Kit [4], to the exception of the autism smartphone diagnosis tool previously mentioned, it doesn’t look like the other projects could help you in the short term. They might in the longer term however, and they will most probably help others in the future.

So what can you do now?

The author of this article is leading MyAIguide, a crowdsourced research project currently focused on repetitive strain injuries and chronic pain. The idea is to look at how different treatments influence pain symptoms over time. Among other treatments, the effect of rest is being studied extensively. It is indeed currently unknown how to best balance rest and exercise when suffering from pain issues: gaining a better understanding of it could have a huge positive impact [13]. MyAIguide is also committed to help its participants: everyone who enrolls in the project will receive free personalized health recommendations after they start tracking health data about themselves. So if you have repetitive strain injuries or other forms of chronic pain, you should definitely consider enrolling in MyAIguide.

On the other hand, if you are a programmer, designer, writer, marketing expert, health professional, journalist, investor or if you think you could help us grow MyAIguide in any way: feel free to get in touch with the author of this article (at:

MyAIguide is guided by very high ethical standards, and is build from the bottom-up by passionate people. Even though MyAIguide is only looking at chronic pain issues at the time of this writing, the aim is to study a lot more diseases in the near future.


[1] : The official information provided by the FDA regarding traditional clinical trials:

[2] : After a clinical trial ends, you sometimes get to continue taking the drug, and sometimes you don’t:

[3] : This article shows how in some ways, the digital health industry is following the same trends the internet did in the 90s:

[4] : The Apple Research Kit is currently probably one of the most well-known crowdsourcing framework :

[5] : Some ethical considerations over the Apple Research Kit: . Overall, the Apple Research Kit really doesn’t deserves to be blamed though.

[6] : Legal guidelines apps using HealthKit need to follow:

[7] : One of the many publications based on 23andMe’s database: . More publications can be found here:

[8] : 23andMe recently rebooted after initially advertising itself as a company that could tell you what disease you were at risk of developing:

[9] : 23andMe can be useful if you want to have children and want to check if they might get a rare heritable disease:

[10] : 23andMe can be used to find relatives you might not know of:

[11] : a report by PatientLikeMe regarding what happened with the lithium study:

[12] : a pdf contaning all the publications based on PatientsLikeMe is downloadable here:

[13] : For injuries and painful conditions, finding the right amount of rest and exercise is often considered an art form: