OBJECTIVES: Prediabetes is a condition where blood sugar levels are elevated, but not high enough to be diagnosed as type 2 diabetes (T2D). There is currently only one ICD-10-CM (ICD) to identify prediabetes. It is uncertain whether this ICD alone accurately identifies prediabetes patients in claims. This study aims to explore the identification of prediabetes patients using claims data linked with HbA1c lab results and to assess progression to T2D.

METHODS: The study cohort included de-identified patients with prediabetes selected from the Komodo Research Dataset linked to Komodo Lab Results and Quest lab data via MapEnhance from January 2017 to May 2023 (≥2 ICD R73.03 or ≥1 lab record HbA1C 5.7–6.4%). The proportion of patients identified using ICD and/or HbA1c was evaluated. The index date was the date of the first qualifying ICD or HbA1c result for prediabetes. Continuous enrollment (1 year pre- and 3 years post-index) and no evidence of prior diabetes was required to assess progression to T2D (≥2 ICD or ≥1 HbA1C>6.4%). Baseline characteristics for those who did not progress to T2D (NDM), progressed within 1 year (D1Y), and within 2 years (D2Y) were examined.

RESULTS: Among 29,566,421 prediabetes patients identified, 21% and 65% were identified using ICD and HbA1c alone, respectively. Among the study cohort, 6.0% and 4.3% progressed to T2D during the first and second year post-diagnosis, respectively. Among the cohorts NDM, D1Y, and D2Y, respective mean age (years) and female distribution were 50.3, 54.6, 53.9 and 56.5%, 57.5%, 55.0%.

CONCLUSIONS: The number of prediabetes patients identified was higher using lab data vs claims data alone, but highest when combined. Similar baseline characteristics among the cohorts (NDM, D1Y, D2Y) were observed except for older age distribution (>18 years) for patients who progressed to T2D (70.2%, 79.9%, 78.9%).