Big data and social networks on health to improve medicine

The algorithms of datamining, a process in which macro data become knowledge for taking decisions, are currently used to analyse health information. From the use of big data, a greater understanding may be gained of everything surrounding human health from the preventive and clinical viewpoint. 

Although these might seem highly sophisticated information systems, there is still a long way to go. In this field, research is under way on how useful it is for medicine to use the data published on the social networks through crowdsourcing. In fact the voluntary participation of citizens connecting to a social network to share health questions produces valuable information that can be analysed and become a useful tool in prevention. 

There are many opportunities, but also challenges that have to be considered in creating virtual communities and social networks around health. Cloud computing technologies can improve these experiences in eHealth, which have to be projected along three premises:

1.- Building a sustainable social network system:

A base must be created that allows information to be created, stored and exchanged between different agents and data to be analysed. At this stage, the following must be borne in mind:

• Semantic interoperability. Given that the health institutions do not use a single common and standard vocabulary, the volume, variety and speed of big data mean that they would have to be integrated, shared, reused and made semantically available with the construction of a primitive ontology (descriptive, not programming computer language) that at the same time integrates other specific ontologies. 
• Account rendering, which may be imposed by law on the organisations that compile, produce or store medical data. The disperse geographic location of some providers in the cloud means that these laws might be breached. 
• Security and privacy. The housing of data in the cloud creates concern, for the service provider might access it, change the data accidentally or deliberately or even delete information. To reduce the security risks in the access to sensitive data, solutions have to be contemplated such as cryptography and public key infrastructure, the standardisation of APIs, and the security of the virtual machine. 
• Legislation. The fact that the data are in the cloud casts doubt on the belonging, custody or possession of the information in question. Agreements on terms of use are important to limit the rights of social network users depending on their role.
• Income and the financial model. The social networks designed for crowdsourcing are mostly free and give no advertising, banners or popups. The sale of anonymous data, clinical trials and market research surveys form the large part of the income of such platforms. 
• Reputation, credibility, quality control and transparency. The success of the social networks largely depends on their credibility on the market. 

2.- Promoting crowdsourcing on the social network:

If the social networks are used as a platform in which user participation is fostered, the four Ps of medicine can be achieved: preventive, personalised, predictive and participative. Many social networks have emerged, such as patientslikeme.com or 23andme.com, which offer an open platform to everyone who wants to share and compare their pathology with that of other people, or who wish to find out and control their own illness. 

To achieve participative supply in the social network, it is important to recruit, retain and assess the users, the semantic management of the queries and the quality of the contributions, as well as to continuously improve the user interfaces. When this is done, in addition to detecting preclinical symptoms, datamining algorithms can be applied to the forums of the social network to identify epidemiological patterns, to study the effectiveness and the side effects of a drug not contemplated in the trials, to monitor and participate in experiments in the real world and even to find new data on treatments, symptoms, progression and medical results.  

3.- Optimising access to these eHealth platforms:

The ineffectiveness of the present processes and the dependency on paper to store data are the principal barriers on access to eHealth data today. The challenge to increase participation in eHealth projects and on specific social networks is to ensure that medical monitoring is carried out with ubiquity, in other words, the step must be taken to treat the patient only at the hospital and to provide them with tools, such as mHealth to allow them to make a follow-up in any place and at any time. 
 

Challenges

The current lack of resources in health means that more and more health services are required of the cloud. The creation of platforms of social networks on health on the internet may therefore be a solution to this demand, but the challenges have to be faced. On the one hand the technical challenges, such as limited broadband, which can create bottlenecks in data transmission.

On the other hand, the most important technical challenge lies on the ICT departments of health institutions and bodies, which undergo a change of role from provider to consultant. Furthermore, the users must not be lost sight of, as they are more and more demanding with the measures of control and transparency of these eHealth services.  
 

Bibliographical reference

Deb, B; Srirama, S. Social Networks for eHealth Solutions on Cloud. Frontiers in Genetics, 3 September 2013 [access: 30 October 2013]. Available at: http://www.frontiersin.org/…