Abstract WMP36: Reliability of ICD-10 codes for Stroke in a Representative US Population

• DOI: 10.1161/str.56.suppl_1.wmp36
• Publication Date: 2025-01-30T10:18:44Z
• Authors: Laura Gutierrez Q...; David Robinson; Simona Ferioli; Paul Wechsler; Robert Stanton; Yasmin Aziz; Joseph Laporta; Lauren Menzies; De los Rios La Ro...; Brian Stamm; Mary Haverbusch; Kyle Walsh; Sabreena Slavin; Lisa Nobel; Shea Wright; Lili Ding; Dawn Kleindorfer; Daniel Woo; Brett Kissela
• Keywords: Stroke

Introduction: ICD-10 codes are often used for stroke research in administrative databases. Few studies have assessed their reliability large, representative populations of the United States. We validated a large population-based study Greater Cincinnati/Northern Kentucky (GCNK) region. Methods: ascertained all acute strokes GCNK region during 2020 using methodology. All hospitalizations were screened comprehensive list ICD any diagnosis position (G45-46/H34/I60-69). Additional cases captured through cold pursuit. Each case was adjudicated by stroke-trained physician. Only ischemic (AIS) and hemorrhagic (HS, defined as intracerebral hemorrhage or subarachnoid hemorrhage) included this analysis. examined how many AIS identified standard (G46/I63), assessing sensitivity positive predictive value (PPV) primary secondary positions. A similar analysis conducted HS (I60/I61). Differences demographics with position, other methods then evaluated Chi-squared Kruskal-Wallis tests. Results: 3,522 AIS/HS events 2020. For (Table 1), we found that had 72.8% PPV 89.1%. When considered improved to 89.3%, but decreased 82.3%. HS, 80.0% 81.0%. 93.4% 54.4%. looking at baseline characteristics 2), patients code lower mRS (overall P<0.01); nursing home residence also more common (7.8%, overall P<0.01). Conclusion: AIS, showed only moderate reasonable PPV. Restricting analyses likely selects against disability baseline.