Self-Monitoring of Blood Glucose and Hypoglycemia-Related Healthcare Utilization in Saskatchewan

Abstract

Background Use of blood glucose test strips has dramatically increased over the past two decades in Saskatchewan and has emerged as a major health care expenditure. Although regular self-monitoring of blood glucose using test strips is imperative in certain patient groups, there is no strong evidence for clinical benefit in the majority of patients with diabetes. As a result, public and private payers have begun to limit the quantity of test strips eligible for reimbursement under drug benefit plans. The majority of literature supporting self-monitoring of blood glucose focuses on improvements in glycemic control. Self-monitoring of blood glucose may also be used to detect hypoglycemia, but the relationship between test strip use and hypoglycemic events has not been well studied. Thus, the purpose of this study was to determine the relationship between blood glucose test strip use and hypoglycemia-related healthcare utilization. Methods Hypoglycemia endpoints and test strip utilization among beneficiaries of the Saskatchewan Drug Plan were described using Saskatchewan’s administrative health databases from 1996 to 2014. A time-series analysis using generalized estimating equations was conducted to test the association between hypoglycemia hospitalizations and utilization of blood glucose test strips at a population level. A nested case-control study was conducted within a cohort of patients with diabetes using a conditional logistic regression model to determine if individual risk of hospitalization for hypoglycemia was lower among patients who used blood glucose test strip compared to those who do not. Results A total of 5,166 hospitalizations for hypoglycemia were recorded during the study period. The average crude rate of hospitalization for hypoglycemia was 26.2 admissions per 100,000. No consistent trend in hypoglycemia hospitalizations was evident. All measures of blood glucose test strip use increased over the study period. After controlling for health care utilization and changes in population size, the number of test strips dispensed was not associated with a significant change in the rate of hospitalization for hypoglycemia (p=0.41). Due to substantial clinical differences between cases and controls, modeling was not conducted in the overall cohort of diabetic patients. Instead, two subgroups were created to represent those at highest (i.e. patients using insulin) and lowest (i.e. patients using low risk oral hypoglycemic agents) risk of developing drug-induced hypoglycemia. After controlling for confounders, blood glucose test strip use was not associated with hospitalization for hypoglycemia in insulin users [adjusted OR 1.08; 95% CI (0.88,1.31); p=0.48], or in low risk oral hypoglycemic users [adjusted OR 1.04; 95% CI (0.55,1.94); p=0.91]. Discussion Blood glucose test strip use was not associated with hospitalization for hypoglycemia in both population and individual level analyses. These findings were consistent among those at high risk and low risk of developing drug-induced hypoglycemia. This research adds to the existing body of literature suggesting that policies limiting blood glucose test strip reimbursement in patients not on insulin are unlikely to be detrimental to patient safety.