Persistent hyperglycemia predicts the development of microvascular complications (e.g. retinopathy and nephropathy) in patients with diabetes. However, the relationship between glucose control and organ damage is complex. Reducing hemoglobin A1c (A1c) prevents or delays microvascular complications but cardiovascular disease (CVD) and mortality are inconsistently affected. Thus, there is a need to develop new quality measures beyond A1c alone to better identify and treat patients at risk for complications and mortality. A1c variability, as measured by fluctuations in A1c over time, is a strong candidate. Several studies show a significant relationship between increased A1c variability and microvascular disease and CVD. While A1c variability carries important risk information, variance measures such as standard deviation, may not be clinically intuitive. Thus, our goal is to develop a new quality measure of A1c variability – A1c time in range (TIR) – that helps clinicians and patients control A1c in a way that balances long-term benefits and risks. We will define A1c TIR as the percentage of days a patient's A1c levels are in a specific target range, based on their clinical characteristics and clinical practice guidelines. We will study A1c TIR in a generalizable nationwide sample of over 365,000 patients from the Department of Veterans Affairs and Kaiser Permanente. We will apply advanced statistical methods that stringently control for selection bias by using an instrumental variable design, including process quality controls. These methods allow us to draw causal inferences between TIR and risk of new diabetes complications. We will study the predictors of A1c TIR, which we hypothesize will be affected by provider practice patterns, individual patient-level characteristics and medications. We will test the hypothesis that higher A1c TIR confers lower risk of diabetes complications. We will also study the converse – A1c time out-of-range (TOR), with interest in deviations both above (TOR [high]) and below the range (TOR [low]) to determine if either is uniquely associated with micro- or macrovascular complications. Then we will investigate clinical factors, hypoglycemic events and rapid declines in A1c, as mediators of the relationship between TIR and adverse outcomes. Each is linked to mortality and early worsening of diabetes complications, respectively. This study will advance diabetes care by developing a novel quality measure that identifies patients in a risk-stratified way and helps clinicians tailor diabetes treatment based on a patient's unique goals of care. Such a new measure will be used by clinicians at the point-of-care and by healthcare systems for population health management.