In recent years, social media has taken the world by storm. During the 12 years since it’s launch, Facebook – the most popular social media network – has amassed over 1.79 billion active users. Social media websites, such as Facebook, are able to collect huge amounts of personal information about their users. While much of the media coverage about this has focused on the ethical issues of using this data for profit through targeted adverts, others have turned their attention to potential beneficial uses of this data.
A group of researchers from the University of Cambridge recently published an article that assess possible ways in which this data could be used to help identify and target mental health conditions. We caught up with one of the authors, Dr Becky Inkster, who answered a few questions about the paper.
Hi Becky. Thanks for speaking to us today. Could you give a short summary of the main points discussed in the paper as well as the importance of the topic?
In our viewpoint article, we argue that we must start leveraging online social networking data to find innovative ways to improve the detection, treatment and prevention of mental health problems. We need to start asking questions, such as “Do the things that people put on social networking sites tell us something about their mental health, and can we use this information to help identify warning signs, or find new treatments?” In our article, we propose a novel online-offline framework, which combines Facebook data collection (of participants’ entire histories since joining the site) with pre-existing offline longitudinal mental health information already provided by these participants. Having access to prior offline clinical and behavioural measures across time will help to frame the meaning of novel, undefined Facebook measures. The goal is to detect online social factors that could indicate emerging psychopathology and/or predict later clinical outcomes. More specifically, we aim to identify more robust indicators of social isolation (e.g., social network structure, frequency of social interactions etc.), atypical or harmful language use (e.g., substance misuse or bullying), facial expressions associated with (or predictive of) mental health problems (e.g., flattened affect) and other indicators relevant to mental health (e.g. semantic coherence as a predictor of psychosis onset). Furthermore, we argue that online social networking interventions could potentially leverage pro-social connections and be beneficial in a therapeutic context. This idea is supported by evidence showing that Facebook users are influenced by the emotional content displayed by other users and a study involving homeless youth, in which those with more online pro-social connections had decreased alcohol intake and decreased depression-like symptoms. In our viewpoint article, we argue for the importance of developing, evaluating and implementing a way in which Facebook data could be translated from evidence-based findings into clinical social intervention settings. Its a very new and exciting area that needs much attention and evidence-based exploration.
What were your motivations behind writing this article?
I had been thinking about this online-offline framework psychiatry concept for several years actually, but it wasn’t until I read Dr Stillwell and Dr Kosinski’s groundbreaking work that I realised my ideas could all of a sudden move from being theoretically possible to actually do-able! Other fields and disciplines have already started using social media and social networking data. I just couldn’t understand why psychiatry and mental health research hadn’t yet. Obviously there are many ethical and confidentiality issues to address, but even still I just couldn’t understand why no one was trying to explore this. It seemed like the right place and the right time. Writing the Lancet Psychiatry article generated a great deal of support from the public, clinicians, patients, academics etc. So we will continue to move this work forward with great enthusiasm and collaborative spirit. I had also run some workshops with young service users to discuss these same ideas that were published in our article. Their responses by and large were very supportive of this research. In fact, they had valuable suggestions and ideas on how to carry out the work. They were very clear on their ethical views too, which ultimately is fundamental to the direction the research should go in.
People are sometimes worried about privacy issues when their data is used for research. What are some of the main ethical concerns that need to be considered when using social media as a resource for research?
This is going to be a central aspect to everything we do moving this work forward. As mentioned, one of the most critical aspects will be to design our research based on the needs and concerns of service users, especially young people. After all, the whole point of this is to supplement their treatment in ways that are innovative and relevant to their daily lives. Safeguarding and confidentiality will be key concerns as will be obtaining informed consent across different stages of illness etc. The research we’ve done so far and plan to continue doing will always be opt-in, fully informed consent with easy access to being removed from the study if they wish. This is a very new territory for ethics, for example, how would we navigate around the topic of introducing digital social networking therapies inside prison walls to strengthen pro-social connections with the outside world prior to their release back into the community? The same applies for in-unit mental health service users. This is just the tip of the ice burg. I am currently preparing a manuscript for publication that discussed a wide range of these and other topics involving the ethical implications of using social networking data in mental health research and clinical care.
How would you like to see this approach taken forward in the field of psychiatry research?
The first step is to collect Facebook histories from existing large-scale adolescent/young adult cohorts. We need to do this urgently in order to enrich the data already available (and while we still have consent to re-contact these individuals). The first goal is to run Proof-of-Principle and Comparison studies using the Cambridge-based ROOTS cohort and the Bristol-based ALSPAC cohort and so on. We use an automated process to collect participants’ Facebook histories in order to extract data such as ‘Likes’, status updates, comments, photographs, and the social structure of friendship networks. These will be analysed using data reduction techniques from machine learning and natural language processing, along with additional methods under development. We will then undertake clinical studies based on what we’ve learned from outcome measures to explore how they can be implemented in social intervention clinical settings, with the goal of making mental healthcare more accessible, especially to young people. I have an additional goal, which is to host an international conference on social media, psychiatry research and digital therapies. I am particularly passionate about developing therapies designed to provide improved access to innovative treatments in under-served, hard-to-reach populations, such as homeless youth, inner city gang members, young offenders and prisoners and those in poor personal circumstances. There is a lot of work to be done and so much potential to revolutionise mental healthcare.
Interview by Owen Parsons.