Implementation and Evaluation of a Real-Time Informatics Alert to Prevent Insulin-Associated Hypoglycemia in the Hospital

项目总结/ ABSTRACT_I_TH _______________________

PROJECT SUMMARY/ABSTRACT_I_TH_______________________ Nestoras Mathioudakis, MD MHS, is an Assistant Professor in the Division of Endocrinology, Diabetes, and Metabolism at Johns Hopkins University. Dr. Mathioudakis seeks a Mentored Patient-Oriented Research Career Development Award in order to obtain the skills, knowledge, and mentored research experience that are essential for a career as a clinician scientist in the field of inpatient glycemic management. This proposal is aimed at reducing preventable harm due to hypoglycemia from insulin treatment among hospitalized patients. Insulin accounts for the vast majority of hypoglycemic events among hospitalized patients, prompting the Joint Commission and the Institute for Safe Medication Practices to designate it a “high alert” medication. In fact, glucose-lowering medications rank first among drugs with the highest rates of adverse events in the hospital. A common cause of insulin-associated hypoglycemia is “therapeutic inertia”—failure to reduce or modify insulin in patients with downward trending blood glucose. The objectives of this proposal are to apply the safety management principles from high reliability organizations, such as commercial aviation, to reduce preventable harm from insulin-associated hypoglycemia in non-critically ill hospitalized patients. The specific aims of this research proposal are to: 1) develop a prediction model for insulin-associated hypoglycemia, 2) develop a real-time informatics alert, using stakeholder engagement and based on clinical practice guidelines, for patients at risk of incident insulin-associated hypoglycemia, and 3) evaluate the effectiveness of a real-time informatics alert in prevention insulin-associated hypoglycemia. For Aim 1, we propose a retrospective study using a large hospital dataset containing clinical, pharmacological, and laboratory data to develop and validate a prediction model for insulin-associated hypoglycemia. For Aim 2, we propose a prospective observational study using qualitative research methods to develop the real-time informatics alerts based on key stakeholder input. For Aim 3, we propose an interrupted time-series (before/after) study design to evaluate whether a real-time informatics alert reduces the incidence of insulin-associated hypoglycemia over a 12-month time period. The goals during the award period include developing expertise in the design, performance, analysis and presentation of clinical research through mentored research, didactic coursework, and formal training in healthcare analytics. Long-term career goals include developing a career as an independent investigator examining health information technology (IT) strategies to improve glycemic control among hospitalized and ambulatory populations with diabetes. The proposed research aims to provide new insights into the development process of informatics alerts and their effectiveness in preventing insulin-associated hypoglycemia in the hospital, which is associated with poor clinical outcomes for patients. This work will lay the foundation for future research to evaluate the role of health IT in targeting glycemic outcomes in both hospitalized and ambulatory patients with diabetes.