Time to look again at the status of social media and its impact on drug safety.
Unlike many other areas in life, the internet and social media have not yet played a major role in drug safety (DS) and pharmacovigilance (PV). Although the internet and email are now playing a bigger and bigger role in clinical medicine (electronic health records, email contact with your doctor or health care provider), telemedicine and other seemingly useful advances, the use of the internet has progressed more slowly with DS and PV.
What have we got so far?
Reporting and Collection of Adverse Events (AEs)
No question that the internet represents an excellent means of collecting drug and device AEs primarily from health care providers. One can report an AE to the FDA MedWatch and, in Canada to Health Canada MedEffect as well as in the UK to the “Yellow Card Scheme” and in Australia.
There are now sites and apps for patient and consumer reports on computers, and more and more on smartphones. These sites provide AE information and some also collect side effects. In the Apple App Store there are over 50 drug apps, many of which include AE data from FDA or other sources. Their primary purpose is to give patients or health care professionals (HCPs) information on drugs, devices, interactions and other pharmaceutical information. Some allow reporting of AEs. One interesting site is Medwatcher.org. This is a project run out of Boston Children’s Hospital and Harvard Medical School, created in collaboration with the FDA. You can report the basic AE information in a few fields and this information is sent to FDA within a few days. The reporter (usually a patient) may be called for further information. See the Q&A on this here.
The site Rxrisk is rather interesting. This site allows one to report an AE. There is a well thought out ten screen questionnaire which will take some time to fill out if done well; it covers demography, medical history, suspect drug, other drugs, impact on daily life and, on the last screen, asks if the report may be sent to FDA? A risk score is calculated using, presumably, an algorithm similar to the classic Venulet or Naranjo algorithms and a report can be prepared for the patient’s physician. This is nicely done but seems time consuming and requires a lot of data from the patient that he/she may not have or be willing to supply. Of course, the patient data is not medically verified and this should be done to increase the usefulness of the data.
Sermo is a widely used site used by physicians, but does not seem to have a mechanism to report AEs. If it does I could not find it.
Basically these are simply on-line flat forms that one fills in and submits. Although progress, it is not fully using information technology to get complete information yet. Each field could be made “smart” whereby filling in a particular field prompts situation specific and appropriate questions (e.g. if “female” is checked a question might pop up asking if you are or could be pregnant). Nonetheless this is a useful advance.
It’s likely that these sites will proliferate and become more sophisticated. FDA is regulating software and apps that touch on medicine and pharmaceutics. Some apps are becoming actual medical devices where an attachment can be plugged into the smartphone and transmit data (pulse, temperature etc.). This is a focus of FDA’s interest in order to insure that they are accurate and useful. Watch this area which will grow rapidly.
Clearly, social media sites have reached the mainstream. FDA works on several social media sites. FDA tweets – they have some 53,000 followers, but Justin Bieber has 51 million followers and President Obama has 43 million followers. On Facebook FDA has 115,000 likes but Justin has 65 million likes. On YouTube FDA’s most popular video is on Lasik surgery with 150,000 views but Justin Bieber (again) as well as Gangnam and Lady Gaga, amongst many others, have over 150 million views.
So what is the message here? It is that FDA and other agencies and health sources understand that social media sites are useful and certainly have captured the public’s imagination but have not yet figured out how to use or harness them to get and to transmit information vital to the public health.
As of May 2013, Facebook reported that it had about 1.11 billion people using the site each month. This would be about 1 out of every 6 or so human beings on earth. This is big data writ large! If this “database” could ever be tapped (“Anyone out there ever have a heart attack within a week after taking drug X?”) the potential to get actionable safety data is large.
Similarly, the pharmaceutical industry has not quite figured out yet how to use social media. Forbes and Ed Silverman commented on this in January 2014: “Tweet This: Pharma Still Fails To Embrace Social Media.”). They cite “anxiety over regulatory compliance, drug makers confront a loss of content control, privacy concerns, a stubborn lack of familiarity with social media and difficulties in trying to quantify to a return on investment.” However, the industry’s use of social media is primarily in marketing and making more information available. Collecting adverse events or educating on AEs is not high on their list. It will be interesting to see if the industry will actually do outreach to get safety data. I suspect this will ultimately be forced on the industry by health agencies. Currently FDA and other agencies do not require companies to use the social media or even periodically screen the internet for safety data. The EMA does require companies to look at its portal for safety information. This will surely change one day as companies will be required to do some sort of internet searching and screening.
Starting To Go The Next Step
Some folks are starting to understand better ways to use social media. One site that is worth looking at is patientslikeme.org. This is a site full of medical information; it is not primarily a drug site though there is drug and treatment information available. The site has information on various conditions. There are over a quarter of a million patients registered and some 17,000 forums/blogs/discussions with a fairly good filter to find the ones most appropriate for the patient searching. People share stories, anecdotes, complaints, data and are most interesting to read. They do not appear to be medically filtered so one is seeing “raw” data directly from the public. For example, for idiopathic pulmonary fibrosis, at the time of this writing, there are 1970 members of this site with the disease of which 54% are male with a histogram presented showing the age ranges of the patients. On the treatment page the members report their treatments, “perceived effectiveness” and whether they mild, moderate or severe or no AEs occurred. There is a link to clinical trials underway or planned which the user may try to join. AEs can be submitted to the FDA via the site. What is most interesting is the forum section.
There is some evidence building up that sites like this can identify trends or problems faster than the FDA, the industry and other classic sources of safety data and trending. A publication in Nature Biotechnology in 2011 entitled “Accelerated clinical discovery using self-reported patient data collected online and a patient-matching algorithm” by Wicks et al. claimed that data from patientslikeme showed that the use of lithium carbonate in ALS was not very effective. What we are seeing here is, in reality, a type of observational study. One suspects that observational studies using the internet and similar sites will become more common. Although these observational studies have their use, they are full of biases and are not a substitute, as the authors note, for double-blind, randomized, controlled trials.
This is leading us into the world of “computational pharmacovigilance” which, when “big data” arrives, may lead to a new way to do PV. At least one workshop has already been held on this in Italy in 2012.
We really are onto something here. The use of the internet, smartphones, tablets and other non-classic means of data collection and communication will open up the way for real-time, rapid analysis of drug use. We will be able to look at effectiveness, adverse events, and safety issues as well as lots of other useful information. This will be be picked up in a fraction of the time and, perhaps, at a fraction of the cost of more classic epidemiologic techniques. If only we could figure out how to do this and get the public interested in supplying data. Maybe Justin would like to work for FDA…