Biomed Analysis: A cautious welcome for online trials
Online clinical trials hold promise for research but scientists should think hard before using them in poorer countries, says Priya Shetty.
Facebook may be the home of many an embarrassing photo, and Twitter host to continually scrolling tedium, but online social networks can be used for so much more.
Scientists have caught on to this and are now using social networks to make their clinical trials virtual — using them to find participants for trials, or inviting patients to monitor and report their own symptoms after taking experimental drugs.
These technologies hold enormous promise for developing countries where resources are scarce and it can be difficult to recruit participants. But there is reason to sound a strong note of caution: research ethics for even conventional trials are often far from robust in these countries.
What's more, the scientists using social networks and those reviewing the ethics of the trial may lack training in research ethics; have to shoulder the burden of ethics review with little institutional support; or struggle to adapt international guidelines to local conditions.
Researchers should by all means be innovative in their approach to testing new health interventions, but they must understand the implications of taking trials online, and ensure that their methods are sound.
The pharmaceutical industry was early to grasp the usefulness of the Internet for recruiting participants to online trials. For instance, in 2009, the drug company Pfizer set up a YouTube video channel called Pfizer Clinical Trials, through which it recruits participants to trials on diseases such as lupus. 
Last year, the company went a step further, announcing that it would run the first 'virtual trial' on a drug called Detrol used to treat overactive bladders. The company will compare the results of the trial in 600 US patients with findings from a previous study on Detrol to validate this virtual approach to clinical research.
Pfizer will screen potential participants online, ship the medicines to their homes and then monitor the results remotely, using information that patients upload through computers or smartphones.
If it proves successful, this could transform research and development (R&D) — many pharma companies can no longer afford to bankroll the costs of clinical trials that increasingly run to millions of dollars. If it cut administration times, it could also mean that effective drugs get to the market quicker.
There are other ways that online communities are being used to test new treatments. Last year, a case-controlled observational trial run through an online community called PatientsLikeMe investigated whether people with amyotrophic lateral sclerosis (ALS) saw any benefit from lithium carbonate treatment. 
The researchers gathered self-reported data from people who had decided to take lithium despite a lack of concrete evidence for its effectiveness. Since ALS is a condition for which no treatment exists, many patients decide to take drugs without prescription.
The researchers then compared the symptoms of these lithium-taking people with those who hadn't taken lithium and, unfortunately for ALS patients, found a year of lithium treatment did not improve symptoms.
Patients with untreatable conditions or illnesses for which there is no viable therapy often experiment with new drugs, and these sorts of trials may be a good way to monitor what effect those treatments are having. By contrast, a conventional randomised controlled trial, in which one group is given the drug and the other a placebo, would not have received ethical approval to administer such highly experimental treatments to people.
The authors of the lithium study acknowledge that patients may be overly optimistic in their assessment of whether a treatment works (since they chose to take it in the first place). But they argue that without compensation or other benefits, patients have no incentive to provide false information.
Another major issue of concern in online trials is ensuring that the information provided is accurate. Although more than two-thirds of the participants in the lithium study provided information about their diagnosing physician, PatientsLikeMe did not confirm this information.
Social networks present identity issues too. Scientists who use social media to survey health attitudes, for example, have no way of being certain that their respondents are teenagers or geriatrics. Nor are there ethics regulations for a consent form, for example, for gathering information that people post on networks.
A review of the pros and cons of online clinical trials states that "the biggest threats to online trials are the security risks of electronic data collection, transmission, and storage", adding that it's unclear how well security measures work. 
This issue must be ironed out before online trials are conducted in developing countries, but signs suggest that it is possible. Careful IT management and the collaboration of mobile phone companies has ensured the security of data transmitted via mobile health technology, where patients or healthcare workers report symptoms via mobile phones, or treatment schedules are sent via SMS to patients.
Developing countries have many unmet R&D needs, and online trials could involve remote populations as well as lower the costs of trials. For all the potential of virtual trials, however, researchers must remember one thing: a trial is only as good as the ethics framework it operates in.
Journalist Priya Shetty specialises in developing world issues including health, climate change and human rights. She writes a blog, Science Safari, on these issues. She has worked as an editor at New Scientist, The Lancet and SciDev.Net.
 PfizerClinicalTeam, accessed 16/6/2012.
 Wicks, P. et al. Accelerated clinical discovery using self-reported patient data collected online and a patient-matching algorithm. Nature Biotechnology 29, 411 (2011)
 Paul, J. et al. The Internet and Clinical Trials: Background, Online Resources, Examples and Issues. Journal of Medical Internet Research doi: 10.2196/jmir.7.1.e5 (2005)