TYPE 1 AND TYPE 2 DIABETES MELLITUS IDENTIFICATION USING ICD-10-CM CODES AMONG NEW USERS OF ANTIDIABETICS: A REAL-WORLD STUDY FROM KOMODO RESEARCH DATASET
Author(s)
Ahmed Noman, BA, Pam Kumparatana, MPH, Catherine Park, MHI, Ting-ying jane HUANG, PhD.
Komodo Health, New York, NY, USA.
Komodo Health, New York, NY, USA.
Presentation Documents
OBJECTIVES: Differentiating patients with type 1 (T1D) and 2 (T2D) diabetes mellitus has been a challenge in observational research. ICD-10-CM has structurally enhanced the accuracy of T1D/T2D diagnosis codes in the US since late 2015. This study aimed to examine the advantage of this adoption in real-world practice and subsequent utility in improving diabetes classification in database analysis.
METHODS: Using administrative data and claims from the Komodo Research Dataset, we identified new users of common antidiabetics - SGLT-2 inhibitors, sulfonylureas, DPP-4 inhibitors, insulin, and GLP-1 receptor agonists from January-December 2024. Patients were required continuous health plan enrollment and no use of respective drug class or gestational diabetes codes (O24*) during 12 months before their first dispensing. The numbers of baseline T1D (E10*) and T2D (E11*) diagnoses were summarized. Mixed diagnosis patterns were benchmarked against T1D diagnosis proportion exceeding 50% (modified Klompas algorithm).
RESULTS: Respectively 94.8%, 96.9%, 97.2%, 67.2%, and 83.1% new users of SGLT-2, sulfonylurea, DPP-4, insulin, and GLP-1 with T1D diagnosis were also coded with T2D diagnosis at baseline, whereas only 3.3%, 2.2%, 3.0%, 5.9%, and 3.2% among initiators with T2D diagnosis experienced the reverse. The modified Klompas algorithm narrowed T1D assignment down to 20.8%, 12.3%, 11.2%, 82.2%, and 46.9% among the mixed-diagnosed initiators. Almost half (47.0%) of GLP-1 initiators had neither diabetes diagnosis at baseline, reflecting recent uptake for non-diabetes indications.
CONCLUSIONS: Compared with similar assessment of ICD-9-CM in the literature, mixed diagnosis patterns for diabetes patients remain in the US health care system after the introduction of ICD-10-CM, despite being observed significantly less often for patients with T2D diagnosis. Reasons other than code accuracy such as off-label treatment intention may be the driving factor to the limited practice change among patients with T1D diagnosis.
METHODS: Using administrative data and claims from the Komodo Research Dataset, we identified new users of common antidiabetics - SGLT-2 inhibitors, sulfonylureas, DPP-4 inhibitors, insulin, and GLP-1 receptor agonists from January-December 2024. Patients were required continuous health plan enrollment and no use of respective drug class or gestational diabetes codes (O24*) during 12 months before their first dispensing. The numbers of baseline T1D (E10*) and T2D (E11*) diagnoses were summarized. Mixed diagnosis patterns were benchmarked against T1D diagnosis proportion exceeding 50% (modified Klompas algorithm).
RESULTS: Respectively 94.8%, 96.9%, 97.2%, 67.2%, and 83.1% new users of SGLT-2, sulfonylurea, DPP-4, insulin, and GLP-1 with T1D diagnosis were also coded with T2D diagnosis at baseline, whereas only 3.3%, 2.2%, 3.0%, 5.9%, and 3.2% among initiators with T2D diagnosis experienced the reverse. The modified Klompas algorithm narrowed T1D assignment down to 20.8%, 12.3%, 11.2%, 82.2%, and 46.9% among the mixed-diagnosed initiators. Almost half (47.0%) of GLP-1 initiators had neither diabetes diagnosis at baseline, reflecting recent uptake for non-diabetes indications.
CONCLUSIONS: Compared with similar assessment of ICD-9-CM in the literature, mixed diagnosis patterns for diabetes patients remain in the US health care system after the introduction of ICD-10-CM, despite being observed significantly less often for patients with T2D diagnosis. Reasons other than code accuracy such as off-label treatment intention may be the driving factor to the limited practice change among patients with T1D diagnosis.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
EPH49
Topic
Epidemiology & Public Health
Topic Subcategory
Disease Classification & Coding
Disease
SDC: Diabetes/Endocrine/Metabolic Disorders (including obesity)