Abstracts

Rationale: The distinction of epilepsy and psychogenic non-epileptic seizures (PNES) is a great clinical challenge. Misdiagnosis rates range between 10 and 30%, depending on the clinical setting. Factual items such as tongue-biting, incontinence or sleep seizures, have been shown not to differentiate between epilepsy and PNES. In view of this we and others have explored the “diagnostic” potential of conversational features which may otherwise be considered redundant or even irritating, such as an apparent difficulty or inability to describe individual seizure episodes, hesitations, restarts or who initiates a particular subject. Systematic analysis of such features in doctor-patient consultations in different in-patient, outpatient or research settings has revealed that the two patient groups are characterized by very different communication profiles (table 1)(1,2). Attention to these profiles can improve physicians’ diagnostic accuracy. However, the interactional, topical and linguistic features which make up the profile cannot be observed if physicians adopt a fact-oriented interviewing style dominated by interruptions and lists of closed questions. Here, we explore how these differentiating features may be elicited in routine clinical encounters.Methods: We used Conversation Analysis (CA) to analyse clinical encounters with 15 patients with video-EEG proven epilepsy or PNES. The encounters lasted less than 30min and followed a simple schedule (table 2). The physician did not mention seizures at the outset and made few inquiries. He listened to the patient without interrupting. He was encouraged to tolerate silence, or to use mmm, right or repetitions of something the patient has said to encourage elaboration. He was strongly discouraged from introducing information to which the patient has not already referred. We assessed to what extent the communicative differences between patients with epilepsy and PNES are observed in these encounters by asking a linguist unaware of the video-EEG result to predict each patient’s diagnosis on the basis of his/her communication profile. Results: The previously described differentiating features can be clearly observed in our brief clinical encounters. They allowed the linguist to accurately predict the patient’s diagnosis in all 15 cases.Conclusions: We describe an open interviewing style, which allows patients to develop their own communicative agenda. Previous studies have shown that this style is associated with higher patient satisfaction rates and compliance with treatment recommendations. This study shows that it also allows physicians to elicit interactional, topical and linguistic features, which can help them to differentiate more accurately between epileptic seizures and PNES.