Despite growing evidence about treatments, many people living with diabetes have poor diabetes control even when healthcare is available. One difficult issue is how to apply medical evidence to individuals. This feasibility study explores change over time and the diversity of pathways to similar health outcomes, to understand how evidence can be tailored to the individual. Six people living with diabetes (two with type 1, and four with type 2) agreed to a series of interviews and diary-keeping. Reading the dataset for each individual reveals a person changing over time through interactions with people and their context. Identifying time as a theme is difficult, as it is ubiquitous. Outcome means little to those living with diabetes: they are living on through time. We developed attributes for each participant relevant to diabetes outcome, describing how they related to others and their environment, capturing emergent properties rather than detail. A similar health outcome could be achieved very differently. Comparison of patterns of attributes may be useful. However, the dynamic, relational nature of the attributes is easily lost from view. How people function in terms of time, change and interaction may be most important for tailoring interventions for improved health outcome.