Studying how people talk to help diagnosis

There are different forms of applied conversation analysis. One of these is known as diagnostic conversation analysis (DCA). DCA an emerging specialty within healthcare research that involves analysing recordings of consultations within specific patient populations to identify distinct word choices and sentence structures in patient descriptions of symptoms. This helps with identification of differences in how diagnostic sub-groups present their problems and can be used to support differential diagnosis. 

We know that diagnosis can be a complex process and it often involves an iterative approach, particularly for conditions for which there is not a straightforward diagnostic test. When doctors do clinical reasoning, they draw on a variety of sources to formulate a diagnosis, including observations of a patient’s behaviour and talk, not just what they say, but how they say it.

While using this diagnostic intuition to interpret patient talk is a supported process within clinical reasoning, there is not a lot of evidence to support its application in clinical practice. The risk of unconscious bias increases when relying on intuitive processes, especially when patient descriptions of illness experience are the primary source for diagnosis. The lack of evidence makes it difficult to teach, even when it is accurate.

DCA offers a systematic approach to identifying observable condition-specific talk-based indicators. It involves identifying talk-based indicators that are specific to a condition and establishing if these can be used to assist with differential diagnosis. DCA has been used to differentiate between conditions such as epilepsy and non-epileptic seizures and dementia and functional memory disorders.

In one of these studies, researchers used conversation analysis to identify conversational profiles of people with dementia versus functional memory disorders. What they're able to do with these profiles is develop tools to help with differential diagnosis between functional memory disorders and dementia. They then tested these conversational profiles to see if they could really help with differential diagnosis between dementia and functional memory disorders. They found that conversational cues can be used to aid the screening and referral process.

By analysing conversation patterns, linguistic features, and interactional behaviours, these studies have demonstrated the potential of DCA to enhance diagnostic accuracy and improve patient care. 

Using DCA provides a systematic approach to supporting more accurate application of talk-based assessment within clinical reasoning, reducing the reliance on intuitive judgments and increasing the reliability of diagnosis. Further research will open the door to integrating such tools into clinical practice.