Last month, Elissa Weitzman and others, researchers at the Children's Hospital Boston informatics program, published a paper in the Journal of the American Medical Informatics Association entitled "Social but safe? Quality and safety of diabetes-related online social networks," finding that only half of the ten communities they surveyed presented content consistent with diabetes science and clinical practice. The study found that the quality of clinical information, as well as privacy policies, varied significantly across these sites, and that some of the sites were wanting in terms of scientific accuracy, safeguards such as personal health information privacy protection, effective internal and external review processes, and appropriate advertising. (See InformationWeek for more.)
It seems to me that the researchers at Children's may be holding diabetes social networks to the wrong set of standards. All participants in these communities openly share information of the sort that is otherwise kept private, and many folks use these sites in order to learn about anecdotal evidence — often of just one case — where an individual patient has experienced positive results from a therapy that has not been proven through a double-blind study.
Albert Maruggi interviewed me about this study, the regulatory landscape for health care social media, and best practices for "ailment communities." You may listen to the interview and read Albert's accompanying piece at Social Media for Health Care.
Update 2/28/2011: See the latest from Susannah Fox at the Pew Internet and American Life Project on patient preferences for dealing with health professionals vs. family, freinds and fellow patients.