Skip to content

The New York Times recently published a feature article on a website called PatientsLikeMe. This is an online community like facebook or MySpace, but with a medical twist. The members have serious medical conditions, like Parkinson’s disease, multiple sclerosis, or HIV, and they use site to post quantitative information about their treatment and symptoms. The site then presents this information for all to see. For example, users can search the website for a drug and then view bar graphs illustrating reasons that members take the drug, the distribution of dosages, length of treatment, reasons for stopping treatment, and patient ratings of the treatment. Individual profiles also show line graphs plotting disease progression and showing major treatment events. The aim is to offer patients the information required to better tailor their own treatment.

It’s easy to think of both risks and benefits of this sort of website.

On the benefits side, PatientsLikeMe makes available a lot of potentially powerful information that could otherwise only have been brought to light through expensive trial. The website may thus enable patients to take greater control over their own conditions, and to acquire medical knowledge along the way. There is also the potential for doctors, researchers and pharmaceutical companies to tap into the information to better inform their own work.

Thomas Goetz, author of the Times piece, also hints at another interesting benefit. Much of the information presented on PatientsLikeMe is highly sensitive and personal. Under ordinary research conditions, it could not legally be publicised without the consent of the patient, which can be difficult and costly to obtain. In making privacy the default, privacy laws work well for those who highly value their privacy, or have little to gain from greater access to the sensitive information of others. But there are some who care less about privacy, and could gain a great deal from access to the sensitive medical information of others. For these people, privacy law works poorly, since the difficulties with obtaining consent (especially from large groups of people) limit the amount of medical information that is published, even from those who would be happy to provide it.

PatientsLikeMe mitigates this problem: those who have much to gain from access to the information of others, and who place little value on their own privacy, are likely to use the site. And once they use it, there is no significant practical barrier to presenting their own information. Moreover, nowhere along the way does any researcher need to approach patients seeking informed consent. Overall information supply is increased, especially to those who care most about it, while privacy rights are not threatened.

(As an aside, a similar situation issues exists in the area of organ donation, the topic of one of my earlier posts. In most Anglophone jurisdictions, active consent from the deceased person and/or family is required for organs to be transplanted from a deceased donor – the absence of dissent is not sufficient. The default position is thus non-donation. This is great for those who value post-death bodily integrity highly, and availability of transplant organs less highly. But since the costs of expressing consent are significant, donation rates tend to be low, which is bad news for those with the reverse preferences. But now consider a system in which only those who have signed-up to be donors would have access to organ transplants. Those who cared a lot about access to organs, and little about giving up their own organs after death, would now have a positive incentive to sign-up to be donors. This might increase organ supply while still protecting those who have a strong desire to retain their bodily integrity after death.)

On the risks side, it is easy to think of ways in which the information provided on PatientsLikeMe might be incomplete or skewed, and thus susceptible to misinterpretation. For example, though patient ratings of a drug may provide useful information about obvious side effects, they provide no information on treatment complications that only become apparent in the long term, or that are not associated with the drug by patients. Since information posted by patients is not the result of controlled experiments, it will also be susceptible to the placebo effect, and to interference from possible confounders. (For example, experimental treatments may appear to have a beneficial effect simply because those who use them tend to be highly motivated to seek the best treatment, and this motivation has other benefits.) These risks strike me as serious. But perhaps they could be diminished by having medical and scientific experts interpret the data presented on the website in the light of more robust evidence from clinical trials.


Goetz T. ‘Practicing PatientsThe

New York


23 March 2008



Share on