July 4, 2012
Big Data development in Life Science
The PLC (http://weconsent.us) provides a solution by permitting research participants to contribute their data to a common consented environment enabling broad and multiple research uses. Importantly, patients can withdraw their data from the database at any time. The withdrawal does not operate retrospectively, so derivatives of the data, or even copies of it held on computer drives, are likely to remain available. Publications based on their data would also be unaffected by data withdrawal.
Big Data is becoming an increasingly more popular concept and new companies are launched almost every week. The first to go public - Splunk - was recently valued at $3 billion when it launched earlier this year on Nasdaq. Many players besides Splunk are approaching Big Data and some of these include Wavii, Metamarkets, Palantir and IBM. It seems like there is a race to collect, curate and analyze the vast amounts of unstructured data out there. On the non-profit side, it is reported in the latest issue of Fast Company that the Wikimedia Foundation will complete its first phase of development of Wikidata in August. Wikidata will extract data from Wikipedia to create a database focused on facts and figures with less subjectivity.
At the same time, more data is becoming available from various sources. When Facebook launched its $100-billion initial public offering, its userbase surpassed 980 million users worldwide. This means that nearly one in seven of the world’s population seems to be comfortable sharing personal data over the internet. What started out as a social network where we shared photos has transformed into a forum where we are comfortable in sharing geographical coordinates of our current locations and even position on whether we are organ donors or not. It is perhaps not far fetched to predict that we will be comfortable to share even more in the future. In spite of this trend, most of us still are more circumspect when it comes to sharing genetic, physiological and medical information online. An editorial in this month's issue of Nature Biotech brings up this issue and states that one key reason for poor uptake could be that there is still no simple and transparent way to track how personal data are being used, let alone a means to opt into, or out of, research using the data.
Portable legal consent (PLC)To tackle this problem a new type of consent process - Portable Legal Consent (PLC) - has been launched. The PLC aims to simplify informed consent and allow feedback of results to any participants. This will put donors in greater control of their own data, which will hopefully lead to more data being shared. The feedback mechanism is thought to provide an incentive for individuals to donate data since the patients in the 'traditional model' usually learn nothing of the research outcomes from their specimens or data, which is particularly true if the results are never published. This, however, demands that informed consent is overhauled.
NBT writes that in the United States: informed consent is based on a uniform 20-year-old, almost pre-internet set of regulations, colloquially known as the ‘Common Rule’. Under the Common Rule, the patient’s signature on the consent form following an ‘informing’ conversation creates a legal agreement that allows research (or medical procedures) to go ahead. The scope of research depends on the consent form; in some cases, biological specimens and associated data can be used only in the research described in the original consent. Alternatively, consent can be broader, extending to future research, within or without some limits. Most donations of tissue and data can be used only once, in the original research project. Any subsequent analysis, reanalysis or pooling with other data is breaking the law.
Creating the incentives to share personal data
A crucial element, besides trust, in building a successful platform where personal data is shared is of course to provide incentives to share. Google Health failed to attract users because of many reasons but one may have been that people saw no real benefit in uploading their data. One incentive for donating data or specimens to medical research could be learning about how the data has been useful for the research community. Another may be to get access to a network of patients with the same condition (PatientsLikeMe) or learning about their genetic profile (23andMe). With the latest trends in the Quantified Self movement there will likely be multiple other incentives to share data from consumer-diagnostic tests, biometric devices to measure glucose levels, heart rate and indications of stress. The latest issue of Entrepreneur writes that Quantified Self company FitBit says that the average user takes 43% more steps per day when they use their device. The biggest incentive, however, likely is the fact that FitBit users can track their progress online through infographics, pie charts and stated goals. For some products it may even be enough to just compile data into Facebook statuses to show friends and family how healthy you are (e.g. RunKeeper, WOD of the day, etc).
The question is not if we will share, it is when we will share our clinical data and where.