This study focused on patterns of communication between midwives and pregnant women and their implications for information, choice and control as now advocated in UK government policy. An earlier casenote audit evaluation of a new organisation of maternity care where midwives carry a personal caseload indicated no difference in quality standards of midwifery care from conventional care, yet women using the service gave a different view. In order to understand whether this difference might be an artefact of the research, responses to change, or a reflection of the limitations of using casenotes for research, an observation-based study was conducted. Forty interviews were observed in three UK settings: hospital clinic, GP clinic and women's homes. Interviews were tape-recorded and notes and drawings of interaction made. The transcripts were analysed using structured and qualitative approaches. The interactional patterns differed according to model of care i.e. conventional or caseload, and setting of care. Several key 'tasks' in the visits were noted, with risk screening and health education information being dominant in conventional care. A continuum of styles of communication was identified, with the prevalent styles also differing according to location and organisation of care. The hierarchical and formal styles discussed in earlier sociological work were the most common in conventional care, despite the focus of midwifery on being 'with-woman' and the recent policy emphasis on consumer choice. The caseload visits showed a less hierarchical and more conversational form and supported women's reports that this model of care offered them greater information, choice and control. The variation in patterns suggests that context is an important consideration in research of this type, with environment (both micro- and macro-level) and organisation of care influencing the ways in which the concepts of choice or consumerism operate in practice. Keywords: Maternity services; Midwives; Antenatal care; Consumer choice; UK