Type 1 diabetes is a disease defined by the inexorable autoimmune destruction of pancreatic beta cells, leading to endogenous insulin deficiency. C-peptide, a 31 amino acid protein that joins the alpha and beta chains of insulin in the proinsulin molecule, is well established as a marker of endogenous insulin secretion. Circulating levels within people with type 1 diabetes demonstrate persistence of insulin secretion, in some cases, for many years after diagnosis. Additionally, histological analyses of donor pancreata have provided evidence for persistent immunoreactive insulin-positive beta cells. These findings have challenged the dogma that all beta cells are destroyed at, or soon after, onset of type 1 diabetes. Although it is clear there is some relationship between residual C-peptide and preserved beta cell mass, residual C-peptide alone cannot distinguish between loss of beta cell mass and reduced functionality. As such, C-peptide level remains a contested surrogate for the aetiopathological definition of type 1 diabetes, which is used across disease classification and as the end point in many intervention trials designed to preserve beta cell function. A fundamental difference between type 1 diabetes and type 2 diabetes is that the former is characterised by rapid progression to endogenous insulin deficiency due to autoimmune beta-cell destruction. Since histological classification is impossible in living humans with type 1 diabetes, C-peptide-defined type 1 and type 2 diabetes have been used as the endpoint in the development and validation of classification models which combine clinical features and biomarkers to improve classification of disease at diagnosis. In Chapter 2, I aimed to I validate a classification model that was previously developed on a C-peptide outcome in a clinical cohort, against a histological definition of type 1 diabetes. This classification model combined age, body mass index (BMI), autoantibody status and type 1 diabetes genetic risk score (T1D GRS), with its predictive performance tested on samples defined histologically as having type 1 diabetes and non-type 1 diabetes from the Network for Pancreatic Organ Donors with Diabetes (nPOD) biobank. Strong predictive performance of the model in this setting demonstrated that the C-peptide outcome, used in its development, is representative of histologically defined disease, confirming that C-peptide is a robust, appropriate surrogate outcome that can be used in large clinical studies where histological definition is impossible. In the 1970s it was crystallised that type 1 diabetes is a disease mediated by the autoimmune destruction of insulin producing beta cells. Since then, the centrepiece of many disease modifying intervention trials has been to augment the survival of functional beta cells, assessed via measures of preserved C-peptide secretion. However, there are clear differences in disease progression between children and adults with recent suggestions that, even within children, differences are driven by underlying endotypes. In Chapter 3, across disease duration, I compared the trends of decline of C-peptide in a cohort of living children from the UK Genetic Resource Investigating Diabetes, to the trends of decline of pancreatic beta cells in organ donors from the combined nPOD and Exeter Archival Diabetes Biobank (EADB), through stratifying by newly described age at diagnosis associated endotypes. I demonstrated that C-peptide loss and beta cell loss, in all age at diagnoses studied, mirror one another across duration of disease. I demonstrated that proportionally fewer children diagnosed <7 retained C-peptide after one year of diagnosis, with the levels of retained C-peptide being lower at diagnosis that those diagnosed at older ages. I showed these trends of loss are almost identical in pancreas donors, with proportionally fewer children retaining islets containing insulin positive beta cells after one year of diagnosis, with fewer insulin positive beta cells at diagnosis as compared to donors diagnosed at older ages. The results in this chapter are indicative of the differences in disease progression in children. The rapid depletion of C-peptide and beta cells in children diagnosed < 7 years is suggestive that early intervention close to or before diagnosis may be most time critical, and should additionally be considered in planning and interpretation of intervention trials. Preserving C-peptide is unequivocally beneficial to a person diagnosed with type 1 diabetes, associating with reduced frequency and severity of self-reported hypoglycaemia and fewer long term microvascular complications, as evidenced originally from DCCT. In Chapter 4, using insights from continuous glucose monitoring (CGM) technology, I demonstrated that higher levels of endogenous insulin secretion around the time of diagnosis impact glycaemic variability, but are not associated with hypoglycaemia. The work in this chapter adds to findings from previous studies of longer duration diabetes to offer a more complete picture of the impact that variation in C-peptide levels have on glucose control in people with type 1 diabetes. Increased use of flash and continuous glucose monitoring has enabled more detailed, daily insights into glycaemic control within type 1 diabetes, the relationships of such with C-peptide have been explored within this thesis. This technology however offers a wealth of opportunity for exploring the lived experience type 1 diabetes, in relation to daily glucose control. In Chapter 5 I developed upon the skills I had refined in CGM data analysis, exploring the impact that free-lived high and moderate intensity exercise have on glycaemic control in type 1 diabetes, as compared to an individual's non-exercise "normal". I compare monitored glucose traces from 10 adults with type 1 diabetes that each completed three, 14-day intervention periods of: home-based high intensity interval exercise, home-based moderate intensity continuous exercise and a free-living non-exercise control period. A key part of this analysis was the careful comparison of the glucose traces in each exercise intervention period to the glucose traces within the non-exercise of control period, in order to understand how much exercise perturbed an individual from their "normal" . In this analysis I found that the exercise modes assessed increase glycaemic variability and hypoglycaemia in the 4 hours after exercise, had a modest effect on glycaemic variability overnight, but increased glycaemic variability and hypoglycaemia the day after exercise. The findings in this chapter suggest that developing focused clinical guidance around time periods post-exercise, and accounting for "everyday life", may improve the management of blood glucose in type 1 diabetes and ultimately reduce barriers to exercise. In the majority of endocrine conditions, the hormone in question is measured as part of routine usual care. In diabetes this is not yet the case. In this thesis I provide evidence that C-peptide is a robust surrogate marker of functional beta cells in clinical settings and demonstrate how an estimate of a patients C-peptide reserve could benefit clinical management. In addition to C-peptide level, I explore how exercise is another influential factor on glucose control in type 1 diabetes. Throughout this thesis I aim to place findings within the context of the lived experience of type 1 diabetes. After all, it is the people living with type 1 diabetes that are the reason we continue our research.