Investigating sequences during clinical communication: the case of the cancer clinic
Gerals Humphris (University of St Andrews, UK)
Many clinicians and patients will support the view that what is said in the clinic when a diagnosis of a malignancy is made is an important conversation.
This statement is not controversial. When it comes to make recommendations to clinicians on how to go about presenting
potentially life-changing information is when there is some debate. Not surprisingly a great deal of effort
has been expended over the decades to study how these conversations can be best managed.
Interviews with clinicians and patients have derived some important clues.
Our work in Medicine has concentrated on observation. I will present a recent study (FORECAST) that has been conducted
with colleagues at the Edinburgh Cancer Centre focusing on patients with breast cancer receiving
the final major treatment of their primary malignancy, namely radiotherapy. We recruited nearly 100 women at
the start of their treatment and followed them over a period of about 4 weeks, collecting audio-tapes of review meetings
with their staff and then inviting these patients to rate their distress about 2 months following the end of treatment.
The attention of our group has been to code the emotional content of the patients' speech, determine
the response of the staff and see if this has any impact on resulting distress in the longer term.
We have used a variety of statistical techniques including latent growth curve models and multi-level modelling to help us
interpret our findings. Difficulties that have presented themselves has been
the effort to identify the emotional speech in lengthy conversations and pointed us to examine the potential to look for
proxies such as parameters in the sound waves of the speech. Furthermore our
ability to detect patterns in speech are fairly crude. Some progress has been made through some multi-disciplinary work
which has already proved fruitful (e.g. Hidden Markov Modelling) however we are exploring the possibilities of revealing
new structures in these review meetings that staff have with their patients in a new study (FORECAST2).