Self-Monitoring of Blood Glucose and Hypoglycemia-Related Hospitalization in a Population-Based Cohort of Canadian Patients With Type 1 or Type 2 Diabetes

Abstract

ObjectivesThe aim of this work was to determine whether: 1) blood glucose test strip use in the population is associated with hypoglycemia hospitalization rates, and 2) blood glucose test strip use among individuals is associated with a reduced risk of hypoglycemia hospitalization. MethodsAdministrative databases from Saskatchewan, Canada, were used to ascertain population-level hypoglycemia hospitalizations and test strip utilization over the period from 1996 to 2014. For objective 1, a generalized linear model with generalized estimating equations was fit to provincial data stratified by age group, sex and year. For objective 2, a nested case-control study was conducted for a cohort of insulin users with diagnosed diabetes. Multivariable conditional logistic regression was used to test the association of test strip use with hospitalization, after adjusting for clinical and demographic factors and health services use. Odds ratios (ORs) and 95% confidence intervals (95% CIs) are reported. ResultsA total of 5,166 hospitalizations for hypoglycemia were identified in the observation period. Annual glucose test strip use increased by over 350%; however, no association was found with provincial hypoglycemia hospitalization rate during the same period, even after controlling for all-cause hospitalizations and population demographics. In the case-control analysis, test strip use was not associated with hospitalization for hypoglycemia among insulin users (n=10,617; adjusted OR, 1.08; 95% CI, 0.88 to 1.31). A sensitivity analysis in an independent cohort of noninsulin users produced a similar finding (n=47,501; adjusted OR, 1.04; 95% CI, 0.55 to 1.94). ConclusionOur findings add to the body of evidence against a protective effect of blood glucose test strip use for serious hypoglycemia.