The ethics and editorial challenges of internet-based research

Facebook is a well-known online social networking site that, by the end of 2013, was reported to have 1.23 billion users worldwide [[1]]. Despite its main purpose as a social networking site, data posted on Facebook was recently used in a large scale research project, causing public outcry [[2]]. The research investigated whether emotional states can be transferred to others through social contagion. By reducing the amount of negative or positive content visible in a user’s News Feed, the researchers were able to investigate whether this had an effect on those users’ own posting behaviours [[3]].

The outcry surrounding the publication of this research has brought the ethical issues of internet-based research to the forefront. While Facebook’s study appears unique in size and controversy, internet-based research involving human subjects is not new. It is an issue that we, as journal editors, have been dealing with for some time.

We have encountered many studies that involve the use of the internet in some way. We have seen analyses of data posted on online fora, analyses of videos posted on YouTube and studies looking at the effect of website content on health-related behaviours. All of these studies have raised questions, including whether the protocol used required approval from an ethics committee, whether informed consent was obtained, and the extent to which participants knew that their data might be used for research.

Studies may now involve participants across the world. This allows for larger sample sizes, as demonstrated by the recent Facebook study, which involved 689,003 people [[3]].

“Internet-based research” is a broad term and includes many different study designs of varying levels of invasiveness and risk. Internet-based research can include both “the Internet as a tool for research” [[4]] and “the Internet as a locale or venue for research” [[4]]. It encompasses: the use of information already available on the internet (with no direct interaction with human subjects); using the internet as an interventional tool; using the internet to recruit research subjects; research about the internet and its effects; and research about internet users [[4]].

The possibilities offered by the internet bring with them many challenges, including verification of participant identity, follow-up and support of participants, ethics approval, informed consent and support of vulnerable groups. In this editorial we focus on what we see as the two main issues for editors – ethics and informed consent.

Journals require that research involving human subjects, their data or material must have been performed in accordance with the Declaration of Helsinki and approved by an appropriate ethics committee. This also applies to internet-based research. An appropriate ethics committee is usually accepted to be an ethics committee local to the participants, and therefore able to weigh up the risks and benefits to local participants. Where research takes place in a single or few locations, this is straightforward, but how does this work for internet-based research? There is no ‘local’ ethics committee and researcher may not even know the locations of participants. Does an ethics committee in one country have jurisdiction to grant approval of a study that involves participants in another country? Should ethics approval be sought in every country from which participants may take part? Perhaps even more challenging is when researchers are based in a country where ethics approval is not required for the particular study design, but participants may be based in a country where it is. These questions raise ethical and logistical challenges which must be balanced against each other.

A fundamental principle underpinning research involving human subjects is the need for informed consent. The need for informed consent from research participants is stated in the Declaration of Helsinki [[5]] and reflected in our editorial policies [[6]].

The Facebook study states that, as all users agree to their Data Use Policy prior to creating an account, they have given informed consent to participate in this research [[3]]. Facebook’s Data Use Policy states that they may use the information about their users “for internal operations, including troubleshooting, data analysis, testing, research and service improvement” [[7]]. It is our interpretation that this does not include research for external publication. It is also not possible to guarantee that those posting were aware of this statement, and therefore that their data could be used in scientific research.

Internet-based research also raises issues surrounding data protection and participant privacy. This is particularly a problem when data that was posted on the internet for other purposes is then used for research. The assumption is that, if information is posted on the internet and in the public domain, it can be considered available for use for any purpose [[4]]; but how far are participants aware of this assumption when they post their data? This becomes particularly problematic if research subjects are children or other vulnerable groups who may post information about themselves without understanding the extent to which their information could be disseminated or used. While individual countries have regulations, such as the Data Protection Act in the UK [[8]], that control the use of private data, there needs to be a global consensus on the use of publically accessible data when the extent of the owners’ (the study participants’) awareness of its use is unclear, especially when the participants are from a country that is different from the researchers’ or are a vulnerable group.

In addition to this guidance from individual institutions, the Association of Internet Researchers has produced a report containing a series of considerations for researchers and those with oversight of internet-based research [[14]]. In the USA, the Secretary's Advisory Committee on Human Research Protections has produced a document outlining considerations and recommendations relating to conducting and reviewing internet-based research, which includes a list of twenty regulatory considerations [[4]].

For manuscripts reporting standard, non-internet research, our editorial polices aim to ensure the work we publish has been conducted ethically and within the required national regulations and frameworks. Given the lack of definitive and comprehensive guidance for internet-based research, we often struggle to know whether research conducted using the internet is ethical and conducted in compliance with the authors’ local regulations and laws such as the Data Protection Act in the UK.

Of the 86 committees we contacted, we have to date received responses from three committees. They told us that they review research proposals that use the internet in the same way as other forms of research and that, so far, they have little experience of reviewing such research. This indicates that there is currently no formal guidance for the conduct of internet-based research in the UK.

It is inevitable that the internet will increasingly become a ‘venue’ and source of data for human research. Guidance on the ethical conduct of such research is needed, not just for those who conduct and review the research, but for journal editors too. It is important that guidelines exist to ensure that internet-based research meets the same ethical and publication standards set for other research involving human subjects. A consensus on ‘best practice’ for internet-based research will be difficult to achieve given the very broad range of research circumstances in which it is used. We currently deal with these issues as they arise on a case by case basis and would welcome comments and views from fellow editors on what they would consider to be best practice.

SLH and JP contributed to drafting this editorial.