Exploring rare patient behaviour with sequential analysis: an illustration

To illustrate the application of sequence analysis to the study of rare patient behaviour in physician-patient dialogue. The rare behaviour in question here is patients' expression of emotional cues and concerns. We investigate which physician behaviours precede and follow such expressions.

Thirty-five cancer-patient consultations performed by four oncologists (two male and two female) were analysed. The consultations were coded with the Roter Interaction Analysis System (RIAS). Sequence analysis by means of Sequential Data Interchange Standard (SDIS) and the Generalized Sequential Querier (GSEQ) was applied to the coded data. Lag analysis (using RIAS categories) was applied to associate the given behaviour (patient 'concerns') with target behaviours (physician utterances).

For female physicians the significantly associated behaviour before the patient's expression of concern was reassurance, while male physicians also applied facilitation behaviour. After patients' expression of concern both reassurance and facilitating behaviour were shown by physicians of both genders.

Sequence analysis appears to be a clinically meaningful and statistically sound method for analysing are patient utterances and associated physician behaviour.