It turns out that social networks are good for something other than sharing what's for dinner, notifying the planet what flavor of coffee's in your cup or looking up old pals from high school. Two clinical studies recently published in peer-reviewed journals tapped patients from different social networks to participate in what the studies are calling "citizen science." They challenged traditional assumptions about patient recruitment,...
sometimes a costly and time-consuming endeavor.
Boston Children's Hospital Informatics Program researchers Elissa Weitzman and Dr. Kenneth Mandl coauthored a study testing the willingness of TuDiabetes.org users to share glycemic control data (indicated by A1c stats) for the sake of public health surveillance. More than 1,600 patients -- 17% of the site's worldwide users -- signed up for the project, which was just the first study coming out of the ongoing surveillance project.
Users ported their A1c data to the study via TuAnalyze, an application the authors describe as "Facebook-like" in an article published in PLoS One.
"I think everybody recognizes that it's increasingly difficult and costly to obtain and maintain research cohorts, and the [traditional] model of research cohorts we've been working with is top-down, it's inflexible, [and] there are disparities between investigator and subject," Weitzman said in an interview with SearchHealthIT.com. "I don't want to say [it's] a dinosaur -- but I'm not sure it's going to be the way of the future. It's certainly not sustainable and scalable. We have an opportunity here to test a model that's very different."
A second study, released by PatientsLikeMe.com and published in the journal Nature Biotechnology, tackled a much broader concept -- a post-market drug study that refuted previous research suggesting that off-label use of lithium carbonate could effectively slow ALS, or amyotrophic lateral sclerosis. Almost 600 ALS patients previously signed up for the study, which tracked results of patients on the drug and collected almost 13 million data points, said PatientsLikeMe founder Jamie Heywood.
The Children's Hospital study simply examined patients' willingness to share data with researchers and each other. The PatientsLikeMe ALS study received mixed evaluations from experts. Yet the clear outcome of both was that patient social networks can connect researchers with willing participants. Ultimately, this might cut recruitment costs and trial times for research hospitals, who often expend much of both just getting a cohort who can give viable data, in the estimation of Nobel Prize-winning scientist Leland “Lee” Hartwell.
Weitzman said the default settings at the TuDiabetes site were established with high levels of privacy, with options for users to change them if they wanted. About 84% of users opted to share their de-identified A1c data with the other diabetes patients participating in the study. This also afforded participants more features, such as GIS Web mapping of their own data in relation to others who also opted to share.
"We found this very high level of what we called 'information altruism,'" Weitzman said. "People were willing, in a privacy-preserving model, to make individual decisions about how they were going to share their data."
It's very easy to do surveys or get data from online things, but we're well aware of how biased those can be from polling and other environments and how easy they are to manipulate the results.
Jamie Heywood, founder, PatientsLikeMe.com
About 30% of study participants went further, she continued, opting to publicly post current A1c test data to a public profile on the site. Not surprisingly, people who were doing better with their glycemic control were more willing to publicly display their A1c stats.
Weitzman and Mandl theorized that giving the users a choice of privacy settings helped engage the diabetics who participated -- and made it clear to participants they were sharing data for the sake of enriching public health.
That, and the fact that it was online, period, Mandl added. Since so few people answer land lines, or even keep them, conducting health surveys over the phone is difficult.
"If we are going to have the ability to do surveillance, we need to find people where they live -- which is online now," Mandl said. "These online health communities are engaging individuals by the hundreds of thousands. That's just in diabetes. And they're existing external to the health system."
Likewise, Heywood found PatientsLikeMe users willing to share details of their health issues with other patients in the online community -- even in cases when they ordinarily would not share such details offline. Moreover, a recent poll among PatientsLikeMe users also revealed other, more practical reasons they are uncomfortable sharing personal health information with employers and even physicians.
In this way, patients can achieve the compartmentalization and privacy they seek in their lives while living their offline social life "apart from their illness," as Heywood noted.
Patient social network takeaways for hospital CIOs
For CIOs at research hospitals, Mandl and Weitzman recommended that projects collecting and using patient data gathered through social networks start with traditional institutional review board (IRB) approval and also apply the usual HIPAA privacy and security protections, even though research projects don't always necessarily fall under HIPAA. In more patient-care-oriented projects, IRB approval might not be needed but HIPAA protections would be.
While meaningful use reporting requirements for diabetes patients will probably be satisfied using International Statistical Classification of Diseases (ICD) codes moving forward, Mandl sees patient social networks as an "important adjunct" to the public health surveillance objectives of the HITECH Act and meaningful use.
Heywood, too, sees some value for hospital IT folks who want to tap social networks to recruit patients for research studies. He also believes they can offer patient access that is hard to duplicate in the offline world.
However, Heywood views PatientsLikeMe more of an online Phase IV clinical research platform rather than simply a patient social network. He thinks it probably would not be simple for a hospital to duplicate his site's data collection successes overnight.
"It's very easy to do surveys or get data from online things, but we're well aware of how biased those can be from polling and other environments and how easy they are to manipulate the results," Heywood said.
"The 'social network thing' I always have mixed feelings about, even using it as a term. We didn't build PatientsLikeMe to be a social network. We happened to let it do that, but there was a point in time early on where we [discussed], 'Are we even going to have a forum, or are we just going to let them talk on other forums?'’ Heywood added. “It's easier to add a social network to a clinical research platform than to think about adding a clinical research platform to a social network."
Let us know what you think about the story; email Don Fluckinger, Features Writer.