Type 1 Diabetes Treatment Patterns and Glycemic Control in a Pediatric Cohort

Embargo until
Date
2015-03-27
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Type 1 Diabetes Mellitus (T1DM) is a chronic disease caused by an auto-immune response where the insulin producing cells in the pancreas are destroyed. Patents with T1DM do not produce any insulin; consequently, they must inject insulin and monitor blood glucose levels manually to attempt to maintain glycemic homeostasis. There is no cure for T1DM. Patients must focus on maintaining blood glucose levels to prevent morbidity from the disease. However, even tight control of blood glucose levels does not eliminate the risk of morbidity, and acute complications are common, especially in early childhood and adolescence. The high doses of injected insulin required to maintain glycemic homeostasis in T1DM patients put them at high risk of severe hypoglycemia (SH). Inadequate insulin, frequently caused by illness, is the primary cause of diabetic ketoacidosis (DKA). The purpose of this dissertation was to describe demographics, initial care, and treatment patterns among T1DM patients and to estimate the effects of different treatment modalities on glycemic control and utilization of emergency care. We used insurance claims and electronic medical records (EMR) from the Department of Defense (DOD) Military Health Systems (MHS) database to create a cohort of pediatric T1DM patients, though different subsets of patients were used for each Aim depending on the requirements of each analysis. For the purposes of these analyses, a switch was defined as a change in insulin delivery method from multiple daily injections (MDI) to continuous subcutaneous insulin infusion (CSII) and an augment was defined as addition of a continuous glucose monitor (CGM) to self-monitoring of blood glucose (SMBG). The first Aim focused on describing an incident cohort of T1DM patients. We found over 96% of patients in our cohort were using conventional therapy, consisting of MDI and SMBG, three months after diagnosis. During follow-up, more than half of patients switched insulin delivery to CSII, and about a third augmented with CGM. Nearly every patient who used a CGM had previously initiated CSII. Female gender, higher military rank of sponsor, and residing in the Central US were associated with a treatment change. We tested for trends in CSII and CGM use over time and found calendar year of T1DM diagnosis was not associated (p=0.78) with CSII use; however, CGM use significantly increased (p<0.0001) during the study period. In Aim 2, we estimated the effect of a treatment change on the next hemoglobin A1c measurement. We found there was no interaction between switch and augment (p=0.93); however, changes in hemoglobin A1c after a switch or augment were depended on the A1c level before the treatment change. Improvements in hemoglobin A1c after a treatment change were, on average, higher among patients with poorer glycemic control. We estimated that patients with an initial hemoglobin A1c of 6% experienced no effect from a switch or augment. However, a patient with an initial A1c of 9% would experience a reduction of 0.55% (p<0.0001) from a switch and 0.30% (p=0.002) from an augment, and a patient with a hemoglobin A1c of 12% would experience a reduction of 1.13% (p<0.0001) from a switch and 0.51% (p=0.007) from an augment. The purpose of the Aim 3 was to estimate differences in total emergency care utilization among patients initiating CSII and/or CGM compared to patients using MDI and/or SMBG. We observed 2.15 total emergency care days per patient year of follow-up. We estimated that CSII (RR = 0.81; p<0.0001) and CGM (RR = 0.88; p<0.0059) decreased total emergency care days among patients initially using conventional therapy. Patients using both CSII and CGM experienced no additional benefit compared to those using CSII (with SMBG) or CGM (with MDI). CSII and CGM therapy were associated with a reduction in ER/ambulance days as well as hospitalization days. A secondary analysis of the effects of CSII and CGM on diabetes-related emergency care showed treatment choices may play an important role in both diabetes-related and total emergency care utilization. Overall we found a substantial proportion of patients utilizing CSII and CGM. Changes in treatment were associated with reductions in hemoglobin A1c and decreased utilization of emergency care. Pediatric T1MD patients using conventional therapy who struggle to attain hemoglobin A1c levels less than 7.5%, the level of glycemic control recommended by the American Diabetes Association, may benefit from CSII and/or CGM.
Description
Keywords
Diabetes, Type 1 Diabetes, Juvenile Diabetes, Insulin Dependent Diabetes, Continuous Glucose Monitor, Continuous Subcutaneous Insulin Infusion
Citation