1286-P: Identification of Atypical Pediatric Diabetes Using Electronic Medical Records (EMR)

Abstract

We aimed to identify cases of rare or atypical pediatric diabetes who may be candidates for studies to understand their etiology and pathogenesis. We tested two strategies in a large pediatric hospital in Southwestern USA. STRATEGY 1: We designed a questionnaire that would allow a health-care provider to rule out typical diabetes types by manual EMR review. This questionnaire was first applied to the EMR of 50 youth (0-21 years old [y/o]) consecutively seen in the diabetes outpatient clinic, then modified to address problems identified, and tested in a second iteration, and on 50 additional cases (n=100). STRATEGY 2: In collaboration with local EMR analysts, we built three queries to generate periodic, automated reports of possible pediatric atypical diabetes cases: unknown type, type 2 diabetes (T2D) diagnosed <10 y/o, and autoantibody-negative type 1 diabetes (T1D). Diabetes type (i.e., unknown, T1D or T2D) was determined by the treating pediatric endocrinologist, through a customized flowchart embedded in the EMR. STRATEGY 1: We found 6 cases (6%) of atypical diabetes (mean age at diagnosis 11 y/o [SD 2.64], 16.6% male, 33% non-Hispanic White [NHW] and 66.6% Hispanic [Hisp]). STRATEGY 2: Unknown diabetes type: 67 cases (15, out of 6676 total diabetes cases), with mean age at diagnosis 12.6 y/o (SD 3.3), 32.8% male, 23.8% NHW, 47.6% Hisp, 25.4% African-American (AA), 3.2% other. Out of 1142 children with T2D, 64 (6.6%) cases were diagnosed <10 y/o, with mean age at diagnosis 8.6 y/o (SD 1.6), 20.3% male, 4.7% NHW, 65.6% Hisp, 28.1% AA, 1.6% other. Out of 680 cases of new onset T1D, 38 (5.6%) were negative for islet autoantibodies (IAA, GAD, IA-2), with mean age at diagnosis 11.3 y/o (SD3.8); 57.9% male, 50% NHW, 19.4% Hisp, 22.3% AA, 8.3% other. In sum, we designed and tested two strategies that use EMR to identify possible atypical pediatric diabetes cases for further studies that may shed light on the heterogeneity of pediatric diabetes, possibly opening new opportunities for diagnosis, treatment and prevention. Disclosure M. Astudillo: None. M. Tosur: None. A.F. Siller: None. A. Balasubramanyam: None. M.J. Redondo: None.