Some things I have learnt while using corpus methods to study health communication

On a dark winter afternoon in December 2011 (before CASS existed), an email from the Economic and Social Research Council informed me that they would fund the project ‘Metaphor in End of Life Care’, which I had put forward with several colleagues from Lancaster. The project involved a combination of manual analysis and corpus methods to study the metaphors used in a 1.5-million-word corpus consisting of interviews and online forum posts by people with advanced cancer, family carers and healthcare professionals. Even in the elation following that announcement, I could not remotely foresee how important that project would become in my (professional) life, and how much interest it would attract from healthcare professionals, charities, the media and the general public. That interest, I now realise, is due to two main factors. First, the topic of communication about cancer and end-of-life is important and intrinsically interesting. In particular, most people have strong feelings about the metaphors used for cancer, such as the ubiquitous but much hated ‘battle against cancer’ that is ‘lost’ by people who die of the disease. The second factor in the success of the project is even simpler: corpus methods actually make sense to people.

At the beginning of the project, I expected that it would be difficult to get audiences outside (corpus) linguistics to understand and appreciate our methods. I was wrong. I have consistently found that people generally, and healthcare professionals in particular, quickly become interested in corpus methods and appreciative of what can be learnt from them. After all, healthcare professionals are used to large-scale quantitative studies and statistical analyses, and therefore do not find corpus-based research at all alien. The dissemination of our findings has therefore been much better received than I would ever have imagined.

A paper we published in BMJ Supportive and Palliative Care has been the most read in that journal from February 2015 to January 2016. In that paper we provide evidence from our patient data of the potential negative, disempowering effects of Violence metaphors (e.g. when a patient says: ‘I feel such a failure for not winning this battle’). However, we also show that those same metaphors are motivating and empowering for some people, who proudly embrace the identity of ‘fighters’ (e.g. when another patient says: ‘cancer and the fighting of it is something to be proud of’). Our paper shows a similar pattern for Journey metaphors, which are used in preference to Violence metaphors in policy documents and guidelines on cancer and end-of-life in the UK’s National Health System (e.g. ‘my cancer journey’). In our patient data, Journey metaphors are used in positive and empowering ways by some people, and in negative and disempowering ways by others. From a healthcare professional’s point of view, this means that there is no easy one-size-fits-all approach to communication about cancer. Rather, it is essential to be sensitive to the language used by individual patients, and to adapt one’s own responses accordingly. The fact that this message has attracted so much interest is due in large part to the nature and amount of evidence that we have been able to provide by using corpus linguistic methods. In retrospect, I shouldn’t have been surprised.

Welcome to our newest senior research associate – Gavin Brookes!

CASS just keeps getting fuller! Gavin Brookes is the newest senior research associate to join the centre, and will be working on our “Beyond the checkbox – understanding what patients say in feedback on NHS services” project. Here’s a little about Gavin, in his own words:

received_10153870011246093I am very excited to begin my role as Senior Research Associate working with Professor Paul Baker on the CASS project, “Beyond the checkbox – understanding what patients say in feedback on NHS services”. The purpose of this research is to help the National Health Service better understand patient feedback with a view to improving frontline healthcare service provision (you can find more info. here: This project is corpus linguistics at its most applied. Its aims are timely and have clear and significant practical consequences and I am thrilled to be a part of it!

I am endlessly fascinated by the relationship between discourse and social life and have adopted corpus linguistic, (critical) discourse analytical and multimodal approaches to investigate this relationship in my research to-date. My enthusiasm for this project will come as little surprise when I tell you that I am particularly interested in how discourse shapes and represents our experiences and understandings of health and illness. My ESRC-funded doctoral research, undertaken in the School of English Studies at The University of Nottingham, examines the discursive construction of a contested condition known as diabulimia in a specialised corpus of online health messages.

Outside academia I spend my time walking, travelling, reading fantasy and science fiction novels, partaking in pub quizzes, and following my beloved (if perpetually under-achieving) Mansfield Town FC. I am delighted to be here and can’t wait to learn more about, and get involved in, the research that is being undertaken within the Department.

Congratulations to our newest CASS PhD student!

We are excited to be welcoming Craig Evans to the centre in October, as the recipient of a PhD studentship which was awarded to CASS for winning the Queen’s Anniversary Prize. Here is a little about Craig, and the project he will be working on, in his own words:

Craig Evans photoI am delighted to have been offered the opportunity to study for a PhD at Lancaster University from October. The PhD is part of a studentship funded by the Queen’s Anniversary Prize, and will involve working with a 40-million-word corpus of NHS patient feedback forms. A primary aim of the research will be to identify ways of improving patients’ experience of the NHS using corpus-based discourse analytic approaches.

I developed an interest in corpus linguistics and discourse analysis during my undergraduate degree in English Language and Linguistics, which I studied at the University of the West of England. There, I used corpus methods in a number of projects. A particular highlight was a study of the media representation of state care for children, where I investigated keyword differences between corpora using tabloid and broadsheet articles. This formed part of my undergraduate dissertation on the topic of care leaver identity, which helped to cement my interest in how social reality is constructed in discourse, especially in relation to care practices.

I am currently studying for an MA in Discourse Studies at Lancaster. When I’m not studying, I like to go walking in the countryside, mostly in the Forest of Bowland which is near to where I live. Other things about me: I enjoy watching films, in particular documentaries and psychological thrillers. I like the novels of Graham Greene, although my favourite novels are Sartre’s The Age of Reason and Mann’s The Magic Mountain. I spent my early teens living in Muscat, Oman. I like to visit different European cities when I get the chance. I like comedy, especially satire. And finally, I love music: I have varied tastes, but must admit that I listen to more 80s electronic pop than I probably should.

Beyond the checkbox – understanding what patients say in feedback on NHS services

In 2016 I will be working on a new project in CASS, which has received funding from the ESRC (£61,532 FEC). The purpose of this project is to help the National Health Service better understand the results of patient feedback so that they can improve their services. The NHS gathers a great deal of user feedback on its services from patients. Much of this is in “free text” format and represents a rich dataset, although the amount of text generated in the thousands of feedback forms patients fill in each year makes it unfeasible to undertake a close qualitative analysis of all of it. Categorisation-based approaches like sentiment analysis have been tried on the dataset but have not found to be revealing. In this project we will be working with the NHS to first identify a set of research questions they would like to be answered from the data, and then we will use corpus-based discourse analysis to draw out the main themes and issues arising from the data. We will focus on four key NHS services – dentists, GP practices, hospitals and pharmacies. From these services alone we have around 423,418 comments to analyse, totalling 105,380,697 words. Some of the issues we are likely to be focussing on include: what matters most for patients, the key drivers for positive and negative feedback, indicators in comments that might trigger an alert or urgent review and differences across providers/services or by socio-demographic group.