Encouraging self-management has been viewed as one means of reducing health service utilisation and contributing to improved demand management. However, the processes and imputed relationship between self-management education skills and health service contact are poorly understood. This paper reports on data from an embedded qualitative study which ran alongside a randomised controlled trial in England designed to test the clinical and cost effectiveness of a self-care support policy which found no statistically significant reductions in health service utilisation. Drawing on concepts from the sociology of chronic illness, analyses suggest that the biographical and social context relevant to individuals' experience of living with a long-term condition, history of health service utilisation, and relationships with health professionals are relevant to understanding the impact of self-management education and related policies aimed at bringing about changes in service use. Our study suggests that future health policy assumptions about utilisation in the context of chronic disease management and self-care support polices may benefit by acknowledging the complex, contextual and recursive nature of health service utilisation operating in the life worlds of patients' experience of living with a long-term condition.