Perhaps I am a purist, but as a member of the healthcare analytics community I have been following with much concern the growth of social media where patients enter conditions, treatments, and outcomes, for retrieval as statistical summaries by others.
A recent NY Times article May 28th called out two social media websites that encourage consumers to describe their medical conditions and what treatments they have been using, namely Cure Together and Patients Like Me.
Having perused and joined these websites, there are certainly positive aspects. One is encouraging patients to track their progress toward achieving health goals, such as weight loss. Another is a venue for desperate patients with serious conditions to quickly link to others who are suffering similarly, and learning about potential treatments.
However, the major point of concern is how data on self-reported patient conditions, treatments and outcomes is being aggregated, summarized, packaged, and even sold as psuedo-outcomes results, or pseudo market research data.
Has healthcare social media research gone from the messaging and linguistic interpretation, as described in
Medical Marketing and Media in May 2009, to more quantitative clinical outcomes analyses?
For those viewing such statistics, Caveat Emptor. There are many sources of biases and unknowns here.
What clinical trials, outcomes research studies, and formal market research surveys accomplish, among other things, is provide clinicians, patients, and healthcare companies with statistically reliable data on treatments and outcomes by applying the scientific method, so that results can be properly interpreted. There are screening criteria for inclusion, test vs. control methodologies, and careful interpretation of results. Pharmaceutical data providers like IMS also hire statisticians to ensure their prescription and claims data are accurately projected.
By contrast, self-reported treatments and outcomes on consumer social media have none of that. Anyone can enter any condition, or any treatment, presuming the right terms are used and laypeople are self-reporting their outcomes using their own interpretations. It is the self-reporting bias taken to the Nth degree.
The websites sited above issue data tables regularly, package them, even create pseudo E-books for sale, and are aiming to re-sell as "research" for health organizations. These may be curiosities, or signposts, but are unlikely to be interpreted as any serious efficacy results.
It will be fascinating to watch the ramifications of this from the medical community, or even patient health advocacy groups.
