Objective: To estimate total diabetes (DM) incidence rates using electronic health records (EHR) for youth <20 years of age from 11 health maintenance organizations (HMO) in 10 US states.
Methods: The denominator was all HMO members <20 years of age from 1/1/2005 to 12/31/2009. To be an incident DM case, youth required > 6 months of membership without DM prior to diagnosis, which was defined as > 2 lab results indicating DM; > 1 inpatient or > 2 outpatient DM diagnosis codes; or dispensing of a glucose-lowering medication other than metformin, outside of pregnancy. Follow-up was censored on the earliest date of DM diagnosis, death, disenrollment, or 12/31/2009. Results: We identified 4 118 DM cases in 9.8 million person years of follow-up for a 5-year DM incidence of 41.9/100 000 person-years. DM incidence increased with age and was higher for girls than boys in the 2 older age groups. The prevalence of insulin use only, oral agents ± insulin use, and no use of any glucose-lowering medication in the diagnosis year was 58%, 23%, and 19%, respectively, for those 10–14 years; and 26%, 48% and 25%, respectively, for those 15–19 years.
Conclusions: We report DM incidence rates for 10–14 and 15–19 year olds that are higher than other US studies, perhaps due to our inclusion of some undiagnosed type 2 DM cases in our estimate of total DM incidence. EHRs provide comprehensive data for case identification and may support low-cost, continuous surveillance of childhood DM in the US.