Analyzing the benefits of PatientsLikeMe social network site

By John Grohol

Two research papers were published this month on the Health 2.0 Web site, PatientsLikeMe. PatientsLikeMe is arguably the only "real" health social network online today, because it allows patients to share actual data that matters with one another -- their personal health data.

(Other supposed health social networks seem more focused on the "social" than the "health," allowing for little integrated data sharing.)

The two research papers provide some interesting data points and insights into the disease process itself and how e-patients are using Web-enabled tools, such as PatientsLikeMe, to improve their own care.

The first study in the European Journal of Neurology looked at one of the diseases PatientsLikeMe covers, ALS. The researchers administered two surveys to basically look at people's perceptions and knowledge -- both patients and the patient's self-report about what their physicians told them -- about ALS.

From the paper's abstract:

RESULTS: A total of 247 ALS patients and 87 caregivers participated. Participants knew less about psychological symptoms than physical ones (72% correct responses versus 82%; paired t((333)) = -5.04, P < 0.001). Patients commonly reported being told by their doctor about physical symptoms such as problems walking (85%) or stiffness/cramps (74%) but not psychological issues like emotional ability (46%) or cognitive change (11%). The majority of patients (62%) and carers (71%) indicated a desire to be informed that cognitive change or dementia might occur.

CONCLUSION: ALS is a multi-system disorder. However, despite a desire for more information from patients and their carers, health care professionals continue to primarily address only the physical consequences of the disease.

The obvious problem with this sort of survey research is that you're relying on patient's self-report about what a third-party told them about the disease, rather than what they actually said. Historically, such third-party self-reporting can vary in reliability and accuracy.

Nevertheless, one of the points of this study was to show the type of research that could be more readily conducted with a 247-patient ALS community at your fingertips online.

The second study in the Journal of Medical Internet Research was more of an exploratory and narrative paper that gives people a sense of the ALS community using PatientsLikeMe and how a hopefully representative sample uses the community.

This study only looked at 123 comments out of 17,059 posted to the community. (One of the interesting footnote findings to me that out of those 17,059 comments, only 7852 -- about 46 percent -- were composed from scratch. The rest just used one of the default comments you could send another user (such as "Thank you").)

Another interesting footnote was that in the course of 15 months, only 63 percent of the registered users posted a comment from scratch to the site. What are the 37 percent doing on the site, one wonders, given that messaging is a pretty integral part of any popular social networking website?

Back to the study...

The thing that makes PatientsLikeMe different from a regular health support group site like NeuroTalk is the addition of patients easily adding their personal health data to their online health profile. Like many other PHRs online, one must manually enter in one's data and keep it updated regularly (for instance, when adding a new medication). In addition, you are encouraged to note changes in your condition -- getting better or worsening, and track anything and everything that might be related to your condition -- diet, treatments, changes in your life, work, supplements, exercise, you name it.

The results' section of the second study looked at three specific categories or types of comments: (1) targeted questions to others with relevant experience, (2) advice and recommendations, and (3) forming and solidifying relationships based on similarity.

Because this was a retrospective study, the only data relationships the researcher could draw were correlational and based upon the patients' self-report. The first category, targeted questions, illustrated these possible relationships. Unfortunately, such relationships are only as good as the data upon which they're based. The researchers gave no indication of the reliability or accuracy of the patients' data in their records, which would seem to be a good foundational base upon which to build studies such as this. I am not certain what the researchers base their assumption that such records are reliable and accurate.

The second category is interesting, but not unique to PatientsLikeMe. Virtually all online patient support communities are full of positive advice and recommendations that members share with one another.

The third category describes how many people in a health support community find others online with share activities and hobbies. Interesting, but there were no data presented that suggest such communications actually show people "solidify" their relationships based upon these shared hobbies.

This is an introductory study, and as such, it does provide some interesting insights into how patients use an online social network in health:

• Although small in number, the comments selected for this study represent an undetermined fraction of all uses of profile data. Nevertheless, they offer insight into the potential value of patients sharing health information.

• This study represents a first examination of the use of shared medical information, which is still a novel model for personal health data.

I'd like to see more information about the PatientsLikeMe data itself -- how many users enter in data regularly, how reliable their data entry is, and whether we can trust that data when it comes to analyzing it on a more global basis.

Psychologist John Grohol is a regular contributor to e-patients blog and also on his own blog Psychcentral.