The last 10 years have witnessed a dramatic increase in the number of people going online to seek out information and advice, to research health issues and connect with those sharing their experiences. NHS Choices, for example, has seen a steady increase in usage every year since 2011, with over 31 million unique visitors to the site in January 2015 alone.
This report focuses on one particular aspect of this new environment: the use of online web forums. These are similar to ‘bulletin boards’ – spaces on which people can gather to discuss their interests, seek advice from those with shared experience, and often simply exchange the chatty pleasantries which constitute everyday life. Typically, these communities will spring up around a specific topic, and a wide range of forums exist which allow users to discuss mental health issues, usually in anonymity. While some forums are private, visible only to those who create an account, many are freely and publicly available.
This report investigates the growing number of mental health related discussions taking place on public online forums, and to explore the potential for the use of computational techniques to provide robust, actionable insight from these conversations to a wide audience – from healthcare professionals and policymakers to those affected by mental health issues themselves – as well as addressing the technical and ethical challenges posed by the collection and analysis of online forum data.
Collecting and analysing one million publicly available forum posts, this study details investigations into three specific areas: an examination of posts from users asking for urgent help online; public discussion around Cognitive Behavioural Therapy; and comments by mental health patients suffering from specific physical conditions. The report also includes a number of posts, paraphrased for the sake of user privacy, which illustrate the types of discourse which takes place on these forums. The methods used to conduct this analysis are described in a detailed methodological annex.
Online forum data represents a potentially valuable store of information, which remains broadly untapped by the healthcare profession. But the process of gaining any useful insight from these data for health professionals, commissioners, researchers, and patients is difficult. In addition to technical and methodological challenges, there are also major ethical and professional considerations of using this type of data, which need to be carefully addressed. These include protecting highly vulnerable participants from harm, and acknowledging that while forums may be in the ‘public’ domain, the expectation of privacy amongst users may be high. With a small number of exceptions, such as those noted above, little research has yet been conducted into the potential use of these technologies to make sense of the mass of unstructured public health data.
The paper explores potential uses and developments of this approach in mental and physical health, technical issues including the further development of software and tools and the ethical issues of approaches using open, but often highly personal, unstructured health data for analysis.