Anticipated Impacts on Veterans'Healthcare This study is designed to inform the VA formulary treatment of thiazolidinediones (TZDs) and insulin analogues, directly affecting VA patients'access and exposure to these medications. Beyond formulary policy, local P&T committees establish procedures to manage access to non-formulary medications. If this project finds evidence that TZDs or LA insulin analogues are more or less effective than alternatives in particular subpopulations, this information could be used in the non-formulary request process. The results of this study will also be of interest to VA clinicians who need to assess the risks and benefits of these medications for each patient. Research of this kind can provide valuable context in general and useful guidance for specific subgroups like patients with depression or substance use disorders. Project Background There is an epidemic of Type 2 diabetes in the general US population and 20% of VA users are estimated to have Type 2 diabetes. Long-term control of blood sugar, at least to near-normal levels, significantly reduces the microvascular complications (e.g., kidney, eye, amputation) of diabetes, and may reduce cardiovascular complications as well. In the VA, metformin is commonly used as first line therapy and the usual recommended second-line agent is a member of the sulfonylurea class. Despite the availability and function of these two classes of medication, over half of VA patients with diabetes require further blood sugar lowering medication beyond metformin and sulfonylurea (MET-SU). Unfortunately, there are few data comparing the effectiveness of various medications available for use at this treatment transition. The two most common drug classes used for patients not achieving adequate glycemic control on MET-SU are TZDs and insulin. Insulin can be further categorized into short- and long-acting, and human and analogues of human insulin. Project Objectives Objective 1: Describe the variation in glucose lowering medication prescribing patterns throughout the VA in patients who transition from MET-SU to additional treatments. Objective 2: Develop a statistical model that predicts variation in glucose lowering medication prescribing patterns. Objective 3: Estimate the relationship between glucose lowering medication prescribing patterns and short and longer-term health outcomes. Objective 4: Estimate the relationship between glucose lowering medication prescribing patterns and short and longer-term health outcomes among sub-populations. Project Methods The proposed project is a retrospective observational study of secondary data from several sources. Primary data sources include patient-level administrative and claims data from three healthcare systems: VA, Medicare, and Medicaid. This study will capitalize on variation in practice patterns to examine which diabetes medications are most effective in patients as their glucose control deteriorates over time. Using a unique national sample of veterans and instrumental variables analysis we will examine the causal relationship between prescribing patterns and risk-adjusted health outcomes.
Comparative effectiveness research is a high priority for the Department of Veterans Affairs. This study helps to implement VA's commitment by investigating the comparative effectiveness of two classes of widely used and costly anti-diabetic medications (TZDs and long-acting insulin analogues) relative to human insulin and low-cost oral medications. Despite the fact that neither of these medication classes is included on the VA formulary, both are widely used to treat VA patients and the long-run consequences of these treatment decisions are unknown. The results of this study will be directly applicable to VA pharmacy management policy as implemented by the VA pharmacy benefits manager and local pharmacy and therapeutics (P&T) committees. Beyond formulary policy, local P&T committees establish procedures to manage access to non- formulary medications. If this project finds evidence that TZDs or LA insulin analogues are more or less effective than alternatives, this information could be used in the non-formulary request process.
| Frakt, Austin B; Pizer, Steven D (2016) The promise and perils of big data in healthcare. Am J Manag Care 22:98-9 |
