Introduction and Objective: Early type 1 diabetes (T1D) detection is vital to prevent DKA, which affects 60% of pediatric cases. However, traditional screening relies on invasive blood draws, limiting accessibility for young children. We launched the first “fully remote” T1D screening program designed for large-scale implementation using a novel genetic stratification tool alongside online monitoring and support. Methods: We developed a Genetic Risk Score (GRS) using 300 T1D cases and 10,000 controls. To ensure regulatory stability and scalability, the GRS uses a specific SNP panel requiring no imputation—a key requirement for FDA approval. The model achieved 0.9 AUC. The workflow utilizes at-home saliva collection and a fully automated central lab handling 1,000 samples/day. We enrolled 1,000+ first-degree relatives (FDRs) in a prospective cohort. All participants received at-home genetic testing, annual autoantibody (Ab) testing via dried blood spots, and virtual psychosocial support from certified counselors. Results: Initial data confirms the program's ability to identify multiple Ab-positive individuals remotely. Those stratified as “high-risk” by the GRS had 10 times higher odds of testing positive for multiple autoantibodies compared to the low-risk group—significant given that nearly all participants are FDRs. Longitudinal assessments showed that remote genetic counseling and support effectively reduced participant anxiety over time, confirming the safety of a decentralized model. Conclusion: This study demonstrates a scalable, fully remote T1D screening model. By overcoming imputation and logistical barriers, it provides a blueprint for population-level screening. A multi-tier, at-home approach driven by a high-performance GRS is a viable and cost-effective strategy to identify at-risk individuals early, optimizing resource allocation and improving clinical outcomes. Disclosure Y. Matsuda: Stock/Shareholder; Current; T1D Scout. Z. Yu: Employee; Current; Genomelink Inc.
MATSUDA et al. (Sun,) studied this question.