Diabetes is one of the most common chronic diseases among persons aged <20 years (1). Onset of diabetes in childhood and adolescence is associated with numerous complications, including diabetic kidney disease, retinopathy, and peripheral neuropathy, and has a substantial impact on public health resources (2,3). From 2002 to 2012, type 1 and type 2 diabetes incidence increased 1.4% and 7.1%, respectively, among U.S. youths (4). To assess recent trends in incidence of diabetes in youths (defined for this report as persons aged <20 years), researchers analyzed 2002–2015 data from the SEARCH for Diabetes in Youth Study (SEARCH), a U.S. population-based registry study with clinical sites located in five states. The incidence of both type 1 and type 2 diabetes in U.S. youths continued to rise at constant rates throughout this period. Among all youths, the incidence of type 1 diabetes increased from 19.5 per 100,000 in 2002–2003 to 22.3 in 2014–2015 (annual percent change [APC] = 1.9%). Among persons aged 10–19 years, type 2 diabetes incidence increased from 9.0 per 100,000 in 2002–2003 to 13.8 in 2014–2015 (APC = 4.8%). For both type 1 and type 2 diabetes, the rates of increase were generally higher among racial/ethnic minority populations than those among whites. These findings highlight the need for continued surveillance for diabetes among youths to monitor overall and group-specific trends, identify factors driving these trends, and inform health care planning.

SEARCH is a population-based registry of diabetes with surveillance of 69,457,475 youths (aged <20 years) covering geographically defined populations in Colorado (all 64 counties plus selected Indian reservations in Arizona and New Mexico under the direction of Colorado), Ohio (eight counties), South Carolina (all 46 counties), Washington (five counties), and Kaiser Permanente Southern California (KPSC) health plan enrollees in seven counties (3). Although the SEARCH population is similar demographically to the U.S. youth population (4), it is not designed to be nationally representative. Case reports were obtained from medical records and validated based on physician diagnosis of diabetes. Eligible participants included nonmilitary and noninstitutionalized persons with diabetes diagnosed at age <20 years, who resided in one of the study counties at the time of diagnosis; for persons in California eligibility required membership in KPSC and for American Indians, participation in Indian Health Services at the time of diagnosis (3,4). Race and ethnicity were based on self-report (82%), medical records (15%), or geocoding (3%). Diabetes type was noted as the physician-assigned type at 6 months after diagnosis. Incidence rates are reported for all type 1 diabetes in persons aged <20 years. Because the number of type 2 diabetes cases diagnosed in children aged <10 years were too few to report trends in this age group (181 total cases during 2002–2015), incident cases of type 2 diabetes are only included for persons aged 10–19 years at diagnosis.

For each incident year, the annual denominators included all civilian residents of the SEARCH sites in the same age ranges on December 31 of that year (3,4). Incidence rates and 95% confidence intervals (CIs) are presented as 2-year moving averages and expressed per 100,000 person-years (5). A change point [or joinpoint] was placed at the year 2011 based on an information criteria measure (6). Comparisons were made between the periods 2002–2010 and 2011–2015 to determine whether the annual percentage change (APC) was constant over the 2002–2015 period. Consistency of the incidence trends over time by age, sex, and race/ethnicity was evaluated by testing for interaction between each of these variables separately with the change point at year 2011 using segmented regression. Rates adjusted for age, sex, and race/ethnicity and estimation of the change in the annual incidence trends during 2002–2015 are reported. A statistically significant change in incidence trends is indicated when the 95% CI excluded zero. Incidence trends were modeled separately for type 1 and type 2 diabetes assuming a negative binomial distribution with a logarithmic link and using a generalized autoregressive moving average to account for serial correlation and presented by race/ethnicity (7). Completeness of case ascertainment for the four geographically based centers was assessed using capture/recapture, where the number of times an individual case was found, either in hospital or other clinical setting, was used to estimate the number of missed cases (8). SAS (version 9.4; SAS Institute) and R (version 3.5.2; The R Foundation) statistical software were used for analyses.

During 2002–2015, among 69,457,475 youths at risk for diabetes, SEARCH identified 14,638 youths with type 1 diabetes and 3,916 with type 2 diabetes. Based on the capture/recapture analysis, few cases were missed, with 98%–99% completeness of ascertainment of cases of type 1 and 92%–97% for type 2 diabetes.

Incidence of type 1 diabetes increased during 2002–2015 in all demographic groups except those who received a diagnosis at age <5 years and American Indians (Figure) (Table 1). Incidence of type 1 diabetes differed by age at diagnosis, sex, and race/ethnicity, with higher rates observed among persons aged 10–14 years, males, and whites. The overall APC adjusted for age, sex, and race/ethnicity in type 1 diabetes incidence was 1.9% per year over the entire period (2002–2015). The APC remained constant for children and adolescents aged 5–19 years, in males, and in females. Steeper increases in age-adjusted and sex adjusted incidence of type 1 diabetes were seen among blacks (2.7% per year), Hispanics (4.0% per year) and Asians and Pacific Islanders (4.4% per year) than among whites (0.7% per year). Incidence among Asians and Pacific Islanders did not change significantly during 2002–2010, then steeply increased during 2011–2015 (8.5% per year).

During 2002–2015, the incidence of type 2 diabetes increased among youths aged 10–19 years in all age, sex, and race/ethnicity groups except whites (Figure) (Table 2). During 2014–2015, type 2 diabetes incidence differed by race/ethnicity, with lowest rates observed among whites (0.77) and higher rates among American Indians (3.69), blacks (5.97), and Hispanics (6.45). In the analyses adjusted for age, sex, race/ethnicity, type 2 diabetes incidence increased at a constant rate from the period 2002–2010 to 2011–2015, with an overall APC of 4.8% per year. The steepest APC increase was among Asians and Pacific Islanders (7.7% per year) followed by Hispanics (6.5% per year), blacks (6.0% per year), and American Indians (3.7% per year).