Shrank WH, Choudhry NK, Swanton K, Jain S, Greene JA, Harlam B, Patel KP. Variations in Structure and Content of Online Social Networks for Patients With Diabetes. Arch Intern Med. 2011;171(17):1589-1591. doi:10.1001/archinternmed.2011.407
Author Affiliations: Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts (Drs Shrank, Choudhry, and Greene); Center for American Political Studies (Dr Shrank and Ms Swanton) and Department of the History of Science (Dr Greene), Harvard University, Cambridge, Massachusetts; Center for Medicare and Medicaid Innovation, Center for Medicare and Medicaid Services, Department of Health and Human Services, Baltimore, Maryland (Dr Jain); CVS Caremark Corporation, Woonsocket, Rhode Island (Dr Harlam); and Engelberg Center for Improving Value in Health Care, The Brookings Institute, Washington, DC (Dr Patel).
With the exponential growth of Facebook and other social networking sites, patients are increasingly seeking information and emotional support online, from other patients.1,2 Recent qualitative studies highlight the great potential that online social networks represent as a source of information and encouragement in chronic disease management.3- 5 However, recent research has also identified potential pitfalls in these unregulated sites.3,6 More than 25% of posts on Facebook sites for diabetic individuals were promotional in nature, largely directing patients to learn about products not approved by the Food and Drug Administration, and substantial efforts to gather data from Web site participants were noted.3 These findings underscore the need to guarantee the authenticity of participants and to ensure that online social networks are safe locations for patients to share information.
We know little about how commonly used sites validate the authenticity of participants' claimed identities or how they provide oversight over the content posted by participants. In addition, little is known about the structure of communication, the sources of funding, and the presence or use of advertising on these sites. Accordingly, we conducted a survey of online social networking sites to explore characteristics of these communities and to better inform physicians and patients about the choices available to them.
We identified the most commonly used online social networking Web sites focused on diabetes, a common but complex chronic condition. On November 17, 2010, using Google search methodology, we used the Boolean terms social network AND diabetes without language or geographic search modifiers and selected the top 300 results for further classification.7,8 We excluded any link to a secondary source (news, academic, or blog article), yielding 23 Web sites. Using a modified 3-part definition of social network sites developed by Boyd and Ellison9 as inclusion criteria, we identified 15 social networking sites as our final sample.
Each Web site was evaluated and data was abstracted by 2 reviewers (W.H.S. and K.S.). The frequency of use of each site was determined using statistics listed on the Web site or by communicating with site administrators. We evaluated membership requirements and authenticity checks for each site. We assessed health professional participation and oversight on each site by evaluating the source and nature of their role. We also assessed the presence, source, and content of advertising on each site by clicking on each tab on the site once and clicking through each advertisement once. We identified the source of funding for each site, when reported.
Membership ranged from 3074 to more than 300 000 patients; 3 sites had more than 150 000 members, 8 had more than 10 000 members, while 2 were confidential and 1 was unavailable (Table). Eighty percent of the sites linked directly to Facebook and 67% linked to Twitter. All but 2 of the sites had an explicit membership process, a requirement to post a comment. Yet the information required for membership was minimal and ranged from simply providing a user name and e-mail address to offering limited information about the presence and type of diabetes or relationship to a patient with diabetes. Only 1 site (http://www.tudiabetes.org) required an extensive profile to be sent to the site administrator for approval.
The level and type of health professional participation varied substantially across the sites. Site administrators reviewed the content of posts in 67% of sites we evaluated, and in 47%, administrators responded directly to questions from the members. Physicians were available to answer questions in 33% of the sites but systematically scanned postings and offered feedback in only 7% of postings. More than half of the sites used diabetes educators to answer member questions, and in 13% they scanned member discussions. Oversight practices also varied across the sites. In some (13%), there was no apparent policing of the sites. In others (67%), site administrators regulated postings to assure the validity of information purveyed, while others (7%) deleted any postings that appeared to be promoting any particular product. In 33% of sites, participants self-policed discussions and reported any activity thought to represent misconduct.
All the sites featured a discussion or question forum to facilitate communication. In some (67%), the topics were entirely member driven. In others, a moderator or site employee posted topics that any member could respond to. Staff members helped to categorize conversations in most (53%) sites, either by requiring members to post their discussions into prespecified categories or by categorizing posts once they were placed on the sites. One site required initial posts in each discussion thread to be phrased as a question. Forty percent of sites featured online “chatting”—opportunities that permit smaller groups to communicate directly. In addition, 93% of the sites featured blogging (online journals from individuals).
Industry advertising was permitted on all but 3 of the sites (Table). Half of the sites featured advertising from pharmaceutical manufacturers, 67% from diabetes device manufacturers, 13% from insurers, 67% featured products related to diet and exercise for diabetics, and 13% included advertising for not-for-profit services or foundations. Overall, 13% of the sites we evaluated either included no advertising or only not-for-profit advertising. Sources of funding for the sites varied dramatically, ranging from foundation sponsorship, industry sponsorship, advertisements, Web host sponsors, and voluntary donations. Three sites had no industry sponsorship.
Online social networks may play an increasing role in health promotion, as primary care physicians are asked to see increasing numbers of patients, limiting time for telephone consultations to answer questions related to chronic disease management, and as a Web-savvy population ages and develops more chronic diseases. Our evaluation of the commonly used online social networks focused on diabetes highlights their popularity and wide variability. Existing sites differ in their approach toward communication structure, authenticity and quality oversight, expert participation, and advertising or sources of funding. These metrics may be important to patients when selecting a community and may be of interest to health care providers who ultimately may advise patients about their particular needs. We hope additional research will further explore patient and health care provider perceptions about these metrics to build an evidence base to encourage social network development that will best promote patient health.
Correspondence: Dr Shrank, Brigham and Women's Hospital, 1620 Tremont St, Ste 3030, Boston, MA 02120 (firstname.lastname@example.org).
Author Contributions:Study concept and design: Shrank, Choudhry, Swanton, Jain, Harlam, and Patel. Acquisition of data: Shrank, Swanton, and Patel. Analysis and interpretation of data: Shrank, Choudhry, Swanton, Greene, Harlam, and Patel. Drafting of the manuscript: Shrank, Swanton, and Patel. Critical revision of the manuscript for important intellectual content: Choudhry, Jain, Greene, Harlam, and Patel. Statistical analysis: Shrank, Choudhry, Harlam, and Patel. Obtained funding: Shrank and Harlam. Administrative, technical, and material support: Shrank, Swanton, and Jain. Study supervision: Shrank and Harlam.
Financial Disclosure: None reported.
Funding/Support: This work is supported by a research grant from CVS Caremark and a career development award from the National Heart, Lung, and Blood Institute (K23HL090505-01) (Dr Shrank).