The most interesting comments are about PLM's experience with adverse events reported by its members. In short, PLM found that 7% of 500 randomly selected posts from the 364,000 posts contributed by patients within the PatientsLikeMe Forum during 2009 incorporated all four elements required for reporting an adverse event (an identifiable patient, a specific medication, an identifiable reporter and a reaction). This is 35 times the rate in the frequently cited Nielson study, which found only one such message in its random sample of 500 posts.
Recall that biopharma company UCB and PatientsLikeMe have partnered to create an online, open epilepsy community that captures real-world experiences of people living with epilepsy in the U.S. Part of this partnership includes a pharmacovigilance program to monitor the site for adverse events and report directly to the FDA adverse events associated with UCB products (see "Finally, a Drug Company Embraces Social Media, AEs Included!").
PLM said its study data "would extrapolate to approximately 20,000 events per year in our Forum alone. If we were to include data collected from all other sources (e.g. symptoms, treatment evaluations, side effects, annotations, private messages, research surveys) within PatientsLikeMe then the volume would be much higher. Given our current membership of approximately 55,000 patients representing only 15 diseases among hundreds that we plan to build, the contribution to drug safety and public health could be substantial."
PLM also reported that if patients were given the right tools that facilitate access to their personal medical data, the time required to report an AE to the FDA could be reduced by as much as 44% (from 36 minutes down to 20 minutes). In a pilot program in its 10,000 plus member Multiple Sclerosis community, 195 AERs were generated in one year. PLM found that "75% of these voluntary patient reports contained the four required elements for an adverse event and 24% reported a serious outcome."
Is this good or bad news for pharma companies hoping to cash in on the social media bonanza? Surely, many were bouyed by the Neilson data, which held out hope that AE's reported by patients on company-controlled or sponsored SM sites were not an insurmountable problem. The PLM data could bash those hopes.
But if drug safety is important for pharmaceutical companies, then they should welcome the opportunity that sites like PLM offer to improve the safety of their products. "Sites that integrate data collections systems and processes to capture the richness of patient experiences with drug products have an unprecedented opportunity to improve safety and outcomes for everyone who uses these products," said PLM in its comments to the FDA.
Hi John:
ReplyDeleteThis is indeed an interesting data point. I think the Neilsen study and this one are perhaps the two possible ends of the spectrum.
In our pharma work we are seeing AE rates closer to what Nielsen reported than what PLM is seeing. This may be due to the particular nature of the PLM community - as there is a more rigorous registration process for PLM than for many other communities.
Also, I would be surprised if the AE rate is not quite different across different conditions.
In any event, I think smart pharma companies are going to be using SM analysis to look for potential AE patterns - early signal detection. This enables them to act early if there are any issues.
Final point, smart pharma will not insist that all SM AE's meet the strict FDA guidelines. If you are looking for patterns then you want to know about any AE related to your drug - regardless of whether or not it meets all 4 criteria.
Thanks for sharing.
TO'B
John, good important post. I think incidence of AE's will definitely depend on nature of site. PLM, or something like iGuard.org, are ASKING for patients to report their experiences with drugs. People go there for that purpose. So I'm actually surprised the 7% number isn't higher.
ReplyDeleteFor promotional efforts on YouTube, FB, Twitter etc. where the purpose isn't med experience reporting my guess is the incidence is closer to the Nielsen number.
-Kevin
One more thing that makes PLM a bit unique is that they are there generally to talk about treatments and many do so with their real names and reference their own experience, which means that most conversations automatically contain three of the four criteria. This differs from the more general forums that Nielsen looked at where people more anonymous and talk more about disease conditions versus specific treatments. This means that two or three criteria are missing in most cases.
ReplyDeleteVery important and quality data from PLM though and something we should pay attention to. I think the most important point is that the system is so unwieldy that the average person would never go through the trouble to report an event. If MedWatch were simpler to use, it would get more reports and better quality reports that actually could be used to detect a safety issue in its early stages.
Jonathan Richman
Dose of Digital
The usefulness and relevance of this post is mirrored by the quality of the responses and responders.
ReplyDeleteHowever, and in general, does pharma really ever want to offer or ally themselves to marketing tools that actually encourage AEs and their reporting? At the Nielsen rates maybe but at these?
There are suggestions that this feedback will alert pharma earlier on and in general to issues with their products. So that they can what pray? Ask for a black box? Update sell sheets and reps to inform Doctors of potential risks? "Tweak" products in new clinical trials to try to remove some of the issues? At what cost, and at what delay to mkt & sales and in a down economy?
It's definitely the "right" way to go, but it goes heavily against the grain.
I assume by "grain" you mean "profit".
ReplyDeleteSales and profit, absolutely.
ReplyDeletePerhaps however pharma will evolve (not to do so, will anyway spell death sooner or later) and on its way to becoming the shining knight and beacon of hope (that it is in its better moments already) it will pass through its Microsoft moment - get thru clinical trials, launch, actively collect/encourage patient & physician feedback, raise awareness of issues quickly, study populations with AEs, discover which sub-pops are genetically or otherwise pre-disposed to reacting adversely,inform patients and doctors thus driving down AEs, tweak molecules and launch follow-on brands to 1) the database of patients who reported AEs in the first place 2) to all docs who now have version A and B (and c, d, e etc). Now that would be something.
I'm a bit late to the party, but here's my addition...
ReplyDeleteThe trend toward both listening and engaging on Social Media will be governed by the FDA's response to an influx of AEs. My experience is that you can add significant (specialty drug, educated ePatient market) to your current AE inflow by conducting listening engagements.
If the FDA takes this in stride and accounts for the outreach when reviewing a company's reports then everything will probably be ok. If they interpret the additional reports as having the same weight as spontaneous reports then I'd be worried we would see more requests for further studies or even products being removed from market. Comparing the two is like comparing a dragnet trawler to a sport fisher with a single line in the water. I just hope the FDA sees that.