BackgroundThe NHS is committed to offering patients more choice. Yet even within the NHS, the meaning of patient choice ranges from legal ‘rights to choose’ to the ambition of establishing clinical practice as a ‘partnership’ between doctor and patient. In the absence of detailed guidance, we focused on preciselyhowto engage patients in decision-making.ObjectivesTo contribute to the evidence-base about whether or not, and how, patient choice is implemented to identify the most effective communication practices for facilitating patient choice.DesignWe used conversation analysis to examine practices whereby neurologists offer choice. The main data set consists of audio- and video-recorded consultations. Patients completed pre- and post-consultation questionnaires and neurologists completed the latter.Setting and participantsThe study was conducted in neurology outpatient clinics in Glasgow and Sheffield. Fourteen neurologists, 223 patients and 120 accompanying others took part.ResultsPatients and clinicians agreed that choice had featured in 53.6% of consultations and that choice was absent in 14.3%. After 32.1% of consultations,eitherpatientorneurologist thought choice was offered. The presence or absence of choice was not satisfactorily explained by quantitatively explored clinical or demographic variables. For our qualitative analysis, the corpus was divided into four subsets: (1) patient and clinician agree that choice was present; (2) patient and clinician agree that choice was absent; (3) patient ‘yes’, clinician ‘no’; and (4) patient ‘no’, clinician ‘yes’. Comparison of all subsets showed that ‘option-listing’ was the only practice for offering choice that was presentonly(with one exception, which, as we show, proves the rule) in those consultations for which participantsagreed there was a choice. We show how option-listing can be used to engage patients in decision-making, but also how very small changes in the machinery of option-listing [for instance the replacement or displacement of the final component of this practice, the patient view elicitor (PVE)] can significantly alter the slot for patient participation. In fact, a slightly modified form of option-listing can be used to curtail choice. Finally, we describe two forms of PVE that can be used to hand a single-option decision to the patient, but which, we show, can raise difficulties for patient choice.ConclusionsChoice features in the majority of recorded consultations. If doctors want to ensure a patient knows she or he has a choice, option-listing is likely to be best understood by patients as an invitation to choose. However, an important lesson from this study is that simply asking doctors to adopt a practice (like option-listing) will not automatically lead to a patient-centred approach. Our study shows that preciselyhowa practice is implemented is crucial.Future researchFuture research should investigate (1) links between the practices identified here and relevant outcome measures (like adherence); (2) whether being given a choice is better or worse for patients than receiving a doctor’s recommendation, taking account of clinical and demographic factors; and (3) how our approach could be fruitfully applied in other settings.FundingThe National Institute for Health Research Health Services and Delivery Research programme.