Background and aims: Despite the growing concern about the incidence and prevalence of diabetes mellitus (DM) among youth, few data sources allow for the tracking of U.S. cases on a population level. Electronic health records (EHR) provide an opportunity to identify persons with DM in large diverse populations. We estimated annual DM prevalence by age category and sex over 5 years, 2005-2009, among members of 11 HMOs from 10 U.S. States who were <20 years of age.
Materials and methods: For each year, the denominator was composed of health plan members < 20 years of age on the last day of the year with > 10 years, categorized as insulin only (probable type 1), oral agent + insulin (probable type 2), and no medications (probable type 2 or undiagnosed DM).
Results: The average annual denominator was 2,406,796 HMO members < 20 years of age. DM prevalence increased by 9.1%, from 2.70/1,000 in 2005 to 2.95/1,000 in 2009 (Table). Increases were observed among youth 5-9 years (15.6%), 10-14 years (8.2%) and 15-19 years (11.8%) with a 19.4% decline among youth < 5 years. Among 10-14 year olds, increases were seen in the oral agent + insulin group and among those not using medications. Among 15-19 year olds, increases were seen in all three treatment groups, with the largest absolute and relative increases among insulin users. DM prevalence was higher for girls than boys for all years (e.g., 3.00/1,000 for girls vs. 2.77/1,000 for boys in 2009). Among youth ages 10 and older, girls with DM were more likely than boys to be in the insulin only group.
Conclusion: Using EHR data from members of 11 geographically dispersed HMOs; we found that the prevalence of DM was higher than previously reported for U.S. youth. This may be due to differences in the populations under surveillance, approaches to case ascertainment and definition, the inclusion of undiagnosed cases, or an increase in DM prevalence in recent years. Prevalence was higher for girls than for boys, a finding observed in other studies. The rise in probable type 2 DM prevalence among 10-14 year-olds is of particular concern. EMRs provide a cost-effective opportunity to provide sustainable nation-wide childhood diabetes surveillance. > 6 months of health plan membership. The numerator included individuals with > 2 laboratory test results indicative of DM, > 1 inpatient or > 2 outpatient DM diagnoses, or dispensed any glucose-lowering medication other than metformin, outside of any pregnancy period. We estimated DM prevalence/1,000 members by age and sex, and calculated absolute rate change and % change from 2005 to 2009. We report glucose-lowering medication use for youth